UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The role of autonomic arousal and of perceived skill in return of fear Craske, Michelle Genevieve 1985

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1985_A1 C73.pdf [ 13.3MB ]
Metadata
JSON: 831-1.0096539.json
JSON-LD: 831-1.0096539-ld.json
RDF/XML (Pretty): 831-1.0096539-rdf.xml
RDF/JSON: 831-1.0096539-rdf.json
Turtle: 831-1.0096539-turtle.txt
N-Triples: 831-1.0096539-rdf-ntriples.txt
Original Record: 831-1.0096539-source.json
Full Text
831-1.0096539-fulltext.txt
Citation
831-1.0096539.ris

Full Text

THE ROLE OF AUTONOMIC AROUSAL AND OF PERCEIVED SKILL IN RETURN OF FEAR by MICHELLE GENEVIEVE CRASKE B.A., University Of Tasmania, 1980 B.A. (Hons), University Of Tasmania, 1981 M.A., University Of British Columbia, 1983 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTORATE in THE FACULTY OF GRADUATE STUDIES Psychology We accept this thesis as conforming to the recjii^'ed standard THE UNIVERSITY OF BRITISH COLUMBIA September 1985 © Michelle Genevieve Craske, 1985 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make i t freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It i s understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of PSYCHOLOGY The University of British Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 Date j&#to*™<bbL (> .  DE-6 (3/81) i i ABSTRACT The hypothesis that high heart rate and low perceived s k i l l would be associated with greater return of fear than low heart rate and high perceived s k i l l was investigated in a group of anxious musical performers (N=63) in response to an anxiety-reduction training program. Musicians were taught progressive muscle relaxation and attention-focusing s k i l l s over the course of four weekly meetings. Return of subjective fear was assessed between training program sessions and at a three-month followup assessment. The three major fear response systems were measured at pre, post and followup assessments and throughout the three training program sessions that included behavioural rehearsal. Subjects (pianists, v i o l i n i s t s and vocalists) performed a short piece before an audience (at assessments) or before their fellow group members. Performance quality ratings by independent musicians served as the behavioural measure, anticipatory heart rate as the physiological measure, and subjective units of distress scales as the subjective measure. In addition, subjects completed memory questionnaires, designed to assess their recall of the setting and their response to previous performances, and thought questionnaires, to provide data pertinent to processes hypothesised to underlie return of fear. Four classification groups (high heart rate, low perceived s k i l l ; high heart rate, high perceived s k i l l ; low heart rate, low perceived s k i l l ; and low heart rate, high perceived s k i l l ) were formed on the basis of median splits of heart rate and perceived s k i l l pre-assessment levels. At post-assessment, each group demonstrated fear reduction, heart rate reduced in high-heart-rate subjects, and performance quality improved overall. Followup return of fear was evident in high-heart-rate subjects regardless of i n i t i a l perceived s k i l l status, and was not dependent on i n i t i a l fear levels. Perceived s k i l l was not associated with return of fear. High-heart-rate subjects also overestimated their level of fear for previous performances, and reported more anxious thoughts and thought resensitization between performances. High non-performance heart rate was associated with greater return of fear only in extreme group analyses. Post-hoc analyses compared subjects who did (n=24) and did not (n=25) display followup return of fear. Return-of-fear subjects, in general, had higher heart rates and lower perceived s k i l l than no-return-of-fear subjects, and tended to report thought resensitization between post and followup assessment. In addition, return-of-fear subjects were generally less skilled and performed on fewer occasions over the followup interval. The assessment of between-session return of fear was limited by design faults. The results were consistent with a dishabituation model of return of fear. They also lent support to Wagner's consolidation model in which an alteration of stimulus representations between exposures is believed to produce i v dishabituation. The findings did not support the hypothesis that lack of consolidation may also arise from f a i l u r e to attend to contextual cues, and hence, impaired r e t r i e v a l of stimulus representations. S i m i l a r i t i e s of the data to Bower's description of mood-dependent cognitions were noted. It was suggested that salient internal autonomic cues during mood-congruent states f a c i l i t a t e d overestimation of previous fear, expectation of di s t r e s s and return of fear. F i n a l l y , research p o s s i b i l i t i e s and treatment implications were considered. V Table of Contents Abstract i i List of Tables v i i i List of Figures ix List of Appendices . ...x Acknowledgement x i i i INTRODUCTION 1 LITERATURE REVIEW 4 THREE SYSTEMS MODEL OF FEAR 4 PREVIOUS RETURN OF FEAR RESEARCH 6 AUTONOMIC AROUSAL 12 Habituation Model 16 Habituation and Arousal 20 Consolidation Model 23 Consolidation and Arousal 28 Thought Resensitization 30 PERCEIVED SKILL 38 Self-Efficacy Theory 39 Self-Efficacy and Arousal 43 EMOTIONAL PROCESSING 46 STATEMENT OF THE PROBLEM 48 Experimental Hypotheses 50 MUSICAL PERFORMANCE ANXIETY 52 BETWEEN-SESSION AND LONG TERM RETURN OF FEAR 54 FEAR REDUCTION PROGRAMS 56 Design 58 METHOD 59 Subjects 59 Experimental Setting 62 Assessments 62 Training Program 63 Self-Report Measures •••• 6 4 Behavioural Measures 67 Autonomic Measures 68 Apparatus and Equipment 68 Procedure 69 vi RESULTS 77 INTRODUCTION 77 PRELIMINARY ANALYSES 82 Instrumental Group Differences 82 Classification Group Differences on Subject Variables 82 Inter-rater Rel i a b i l i t i e s 83 Correlations Among Variables at Pre-assessment ....85 In i t i a l Classification Group Differences on Dependent Variables 85 MISSING DATA 87 RESPONSE TO INTERVENTION OF THE CLASSIFICATION GROUPS 88 Pre-performance SUDS analyses 88 Pre-performance Thoughts Analyses 95 S k i l l Analyses 98 Effectiveness of the Training Program 101 Minor Analses 102 FOLLOWUP RETURN OF FEAR 108 Heart rate, efficacy and pre-performance thoughts .110 S k i l l 120 Tonic Heart Rate 122 Multiple Regression 122 BETWEEN-SESSION RETURN OF FEAR 123 Classification Groups 128 ADDITIONAL ANALYSES 130 Correlations Among Dependent Variables 130 Correlations Among Minor Variables 130 External Validity of Training Program 132 Minor Comparisons between Followup Return-of-Fear Groups 135 Thought Item Analysis 137 Recall and Current Fear Levels ......137 I n i t i a l Discordance .139 SUMMARY OF FINDINGS ..140 DISCUSSION 145 TRAINING PROGRAM EFFECTIVENESS 147 FOLLOWUP RETURN OF FEAR 158 Classification Variables ........158 Autonomic Arousal ...... ..........165 Perceived S k i l l 177 BETWEEN-SESSION RETURN OF FEAR 182 v i i DISCORDANCE AND DESYNCHRONY 188 SELF-EFFICACY THEORY 193 GENERALISABILITY OF THE RESULTS 195 CONCLUSION 197 REFERENCES , 205 FOOTNOTES 218 APPENDICES 219 v i i i Table I Table II Table III Table IV Table V Table VI Table VII Table VIII Table IX Table X Table XI Table XII Table XIII Table XIV Table XV List of Tables Inter-rater R e l i a b i l i t i e s 84 Pearson Correlations among Variables at Pre-assessment 86 Classification Group Means for Major Dependent Variables 89 Pre-performance SUDS Means Adjusted for I n i t i a l Differences 94 Heart-Rate Group Means for Memory for Time B .104 Means for Tonic-Heart-Rate Groups 107 Pre-performance SUDS Means for Return-of-Fear Groups 111 Return-of-Fear Group Means for Major Dependent Variables 113 Prevalence of Between-Assessment Return of Fear 124 Between-Assessment Return-of-Fear Group Means 126 Between-Assessment Return of Fear in Classification Groups 129 Average Correlations Among Major Dependent Variables 131 Correlations between Confidence and Skil l s Application and S k i l l s Benefit 133 Followup Application of, and Benefit from, Training Program S k i l l s 134 Frequencies of Expected Comfort and Discomfort in Heart-Rate Groups 138 ix Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 List of Figures Pre-Performance SUDS Means across Assessments for High- and Low-Heart-Rate Groups 91 Pre-Performance SUDS Means across Assessments for High- and Low-Efficacy Groups 92 Pre-Performance Thoughts Means across Assessments for High- and Low-Heart-Rate Groups 96 Pre-Performance Thoughts Means across Assessments for High- and Low-Efficacy Groups 97 S k i l l Means across Assessments for Classification Groups 99 Pre-Performance SUDS Means across Assessments for Extreme High- and Low-Tonic-Heart-Rate Groups 109 Heart-Rate Means across Assessments for Return-of-Fear Groups 115 Efficacy Means across Assessments for Return-of-Fear Groups 117 Pre-Performance Thoughts Means across Assessments for Return-of-Fear Groups 118 Change in Confidence Means across Assessments for Return-of-Fear Groups 119 Change in Thoughts Means across Assessments for Return-of-Fear Groups 121 X L i s t of Appendices Appendix 1 Screening Q u e s t i o n n a i r e 219 Appendix 2 Subject Data 220 Appendix 3 T r a i n i n g Program Manual 222 Appendix 4 T r a i n i n g Program Procedures f o r Research A s s i s t a n t s 238 Appendix 5A P r o g r e s s i v e Muscle R e l a x a t i o n Handout 243 Appendix 5B R e l a x a t i o n P r a c t i s e Sheet 245 Appendix 6A Qu i c k - R e l a x a t i o n Handout 246 Appendix 6B Qu i c k - R e l a x a t i o n P r a c t i s e Sheet 247 Appendix 7A A t t e n t i o n a l T r a i n i n g Handout 248 Appendix 7B A t t e n t i o n a l T r a i n i n g P r a c t i s e Sheet 249 Appendix 8 S t a t e - A n x i e t y Scale 250 Appendix 9 Pe r c e i v e d S k i l l Q u e s t i o n n a i r e 251 Appendix 1 OA Pre-Performance SUDS 252 Appendix 1 OB During-Performance SUDS 253 Appendix 11A Pre-Performance Thoughts Q u e s t i o n n a i r e 254 Appendix 11B Post-Performance Thoughts Q u e s t i o n n a i r e 256 Appendix 12 General Information Q u e s t i o n n a i r e 258 Appendix 13A P r a c t i s e Information Questionnaire-Preassessment 260 Appendix 13B P r a c t i s e Information Questionnaire-Postassessment 261 Appendix 13C P r a c t i s e Information Q u e s t i o n n a i r e - F o l l o w u p Assessment 262 Appendix 14 Session P r a c t i s e Information Q u e s t i o n n a i r e ...264 Appendix 15 Memory Qu e s t i o n n a i r e 265 Appendix 16 Followup P r a c t i s e Information Q u e s t i o n n a i r e ..266 Appendix 17 Performance Q u a l i t y Dimensions 268 xi Appendix 18 Appendix 19 Appendix 20A Appendix 20B Appendix 21 Appendix 22 Appendix 23 Appendix 24 Appendix 25 Appendix 26 Appendix 27 Appendix 28 Appendix 29 Appendix 30 Appendix 31 Appendix 32 Appendix 33 Appendix 34 Appendix 35 Appendix 36 Performance Quality Rating Sheet 269 I n i t i a l Contact 270 Letter to Music Teachers 271 Letter to Music Students 273 Outline of Assessment Procedures for Subjects 274 Assessment Procedure Outline for Research Assistants 276 Consent Form 279 Practise Instructions 280 Memory Questionnaire Information Sheet 281 Debriefing Letter 282 Means, Standard Deviations and t values for I n i t i a l Differences between Instrumental Groups 284 Means, Standard Deviations and F values for I n i t i a l Differences between C l a s s i f i c a t i o n Groups on Subject Variables 285 Means, Standard Deviations and F values for I n i t i a l Differences between C l a s s i f i c a t i o n Groups on Dependent Variables 286 Heart Rate and Ef f i c a c y Means for C l a s s i f i c a t i o n Groups 287 Summary of HRXEFXTime Analyses of Pre-Performance SUDS 288 Summary of HRXEFXTime Analyses of Covariance of Pre-Performance SUDS 289 Summary of Simple Effects Analyses of Pre-Perf ormance SUDS 290 Summary of HRXEFXTime Analyses of Pre-Performance Thoughts 291 Summary of HRXEFXTime Analyses of Covariance of Pre-Perf ormance Thoughts 292 Means and Summary of HRXEFXTime Analysis of Practise ... . 293 x i i Appendix 37 Appendix 38 Appendix 3 9 Appendix 40 Appendix 41 Appendix 42 Appendix 43 Appendix 44 Appendix 45 Summary of HRXEFXTime Analyses of S k i l l 294 Summary of Pre, Post, and Followup Assessment of Pre-Perf ormance SUDS and Heart Rate 295 Summary of Analyses of Memory Variables for High-and Low-Heart-Rate Groups 296 Summary of Analysis of Pre-Performance SUDS for High- and Low-Tonic-Heart-Rate Groups 297 Summary of Analysis of Pre-Performance SUDS for Extreme High- and Low-Tonic-Heart-Rate Groups 298 Summary of Analyses for Return-of-Fear Groups 299 Summary of Multiple Regression Analyses 301 Summary of Analyses for Between-Session Return-of-Fear Groups 302 Correlations Among Dependent Variables 303 xi i i Acknowledgement I wish to thank my supervisor, Professor S. J . Rachman, whose advice, guidance and support have been invaluable. The helpful instruction and advice of my committee members, Drs. K. D. Craig, K. Dobson, J . Johnson and P. Suedfeld, were also greatly appreciated. I am also indebted to Professor D. Reubart for his consultations on musical performance and for the use of his studio, and to Dr. M. Kendrick for allowing me access to audio-visual materials. I also wish to acknowledge the help of the various raters and research assistants who participated in data collection, and Daryl Zacharko for typing this thesis. Finally, I thank my fellow students for their .encouragement and friendship, and my family for their support. A special thanks is extended to Charlie for his patience and understanding. 1 INTRODUCTION The general question addressed by this investigation was: can return of fear be reliably predicted from levels of autonomic arousal and from perceived s k i l l ? Specifically, the study examined whether high levels of autonomic arousal and low levels of perceivedtskill predicted greater return of fear than low levels of autonomic arousal and high levels of perceived s k i l l . Return of fear refers to a "reappearance of fear that was earlier present but had undergone a decline" (Rachman, 1979, p. 65), and has been observed during treatment of simple phobias (Connolly, 1979; Grey, Sartory & Rachman, 1979; Grey, Rachman & Sartory, 1981; Sartory, Rachman & Grey, 1982) and obsessive compulsive disorders (Foa, 1979; Likierman & Rachman, 1980). The phenomenon is described by Rachman (1981) when referring to Likierman and Rachman's ( 1980) investigation of unwa'nted intrusive cognitions: the regular, but slight, within-session decreases in discomfort were followed by regular returns of discomfort in the intersession intervals. The intrasession changes were transitory and in this respect resembled the phenomenon known as 'return of fear'. (p. 96) That i s , although fear levels diminished within each session, the i n i t i a l re-presentations of "contaminants" to obsessive-compulsive patients consistently el i c i t e d high levels of subjective fear on subsequent sessions (Likierman & Rachman, 1981). 2 Return of fear does not describe the development of new fears nor incubation of existing fears, but a usually partial return, in the quantitative sense, of the intensity of the subjective discomfort associated with target fears. It does not seem to reflect an expectancy effect, as return of fear has been observed in individuals characterised as highly aroused, whom Borkovec (1974) and Borkovec and O'Brien (1976) have shown to be relatively unresponsive to the influences of demand and suggestion. Relatively l i t t l e is known of this phenomenon despite its important treatment implications. Prevalence is as yet undetermined, but i t appears to be a robust phenomenon in experimental research (Grey et a l . , 1979; Grey et a l . , 1981; Sartory et a l . , 1982). In addition, the significance of return of fear in natural settings is apparent in Rachman's (1978) analysis of typical responses to air-raids during wartime: long, quiet intervals between air-raids seemed to result in recovery of fears that had formerly diminished over a series of repeated raids. Reasons for further investigation of the return of fear phenomenon are two-fold. F i r s t , i t may advance theoretical understanding of fear reduction, especially regarding Rachman's (1976) proposal of differential mechanisms operating for behavioural, physiological and cognitive response systems. The findings would also pertain to the accumulating evidence for desynchrony amongst those response systems during treatment interventions (Rachman & Hodgson, 1974),. Second, the findings 3 may provide clinicians with information enabling them to identify those individuals for whom return of fear is most likely and to suggest means of either assessing and/or curtailing such resurgence. The identification and control of return of fear is important because the subjective report of fear is generally considered to be essential to clients' satisfaction with treatment, regardless of improvements in autonomic and behavioural response systems (Lick & Katkin, 1976). Long term treatment outcomes for those experiencing return of fear is not yet well established, although in a preliminary investigation Philips (1985) found that return of fear in vomit phobics gradually diminished with repeated treatment sessions. The importance of subject characteristics in the specification of treatment parameters was noted by Gauthier and Marshall (1977) and is emphasised by return-of-fear research. Patients, characterised by a return-of-fear response pattern may benefit more from an increased number of treatment sessions than patients without return of fear. The research paradigm used to date to assess return of fear entails continuous measurement (from each response system) over the duration of relatively brief treatment intervention, and analysis of maintenance of within-session gains. The fate of these gains between sessions and over short (one-week) followup periods has been assessed by comparing the last measurement of the previous assessment with the f i r s t measurement of the subsequent assessment. Published research has been restricted to circumscribed fear populations and exposure based treatments. 4 LITERATURE REVIEW THREE SYSTEMS MODEL OF FEAR Return of fear reflects discordance and desynchrony between the major response systems measured in the assessment of fear and anxiety. Low correlations (or, discordance) found upon sampling in each response system were taken by Lang (1968) to indicate that fear is not a unitary phenomenon. He viewed the emotional response as a complex of three major measurable systems which are modulated by neural brain centres, and whose outputs have low intercorrelations, yet are also interactive (Lang, 1971). Independence of the three systems becomes especially evident when fear is attenuated; mild feeling states frequently involve merely a verbal report of fear, without signs of physiological arousal or behavioural avoidance/avoidance (Lang, 1971). Response system independence has been attributed to several factors: no system uniquely defines the emotional state, and each is influenced by varying other factors; differential sensitivity, with autonomic indices being the least sensitive (Agras & Jacob, 1981); individual differences in system-sensitivity (Lang, 1977); and, error in measurement (Lang, 1977). Lang's original formulation centred on notions of concordance and discordance, while Rachman and Hodgson (1974) coined the terms synchrony and desynchrony to refer to degree of covariance of changes among the response systems. Tfiey extended Lang's three system model on the basis of the observation that 5 not only may the systems be imperfectly correlated at any given point in time, but they may also respond with different rates of change. Desynchronous change refers to either independent or inverse relationships. For example, Rachman and Hodgson (1974) described a pattern typical of obsessional neuroses involving intense fear and active avoidance followed by fear reduction with continued avoidance. In response to treatment, i t is predicted that: Theoretically, these response systems (subjective, behavioural and physiological) should not be uniformly susceptible to the presence or absence and the quality, of exposure in a therapeutic (or other) event; they should instead show differential changes, (de Silva & Rachman, 1981, p. 230) The hypothesis that highly demanding treatments (such as flooding) would produce greater desynchrony, by enforcing behavioural approach before reducing fear and arousal, than less demanding treatments (such as desensitization) led to the i n i t i a l study in return of fear (Grey et a l . , 1979). However, while Grey et a l . expected to find within-session desynchrony, they instead found between-session desynchrony: synchronous within-session reductions in subjective fear and autonomic arousal were followed by a return in subjective fear while lowered arousal levels were maintained. A major implication of the three systems model is the likelihood that each response system is influenced by a different mechanism. Return of fear research adds to the accumulating evidence regarding this issue. Rachman (1978) and Borkovec (1976) emphasised the role of biological differences in 6 propensity such as the susceptibility of some individuals to conditioned fears. Borkovec (1976), in'addition, suggested that individual learning histories may account for the varying relative intensities, or functional importance, of each response system in emotional states. Of special relevance to the three systems model is Rachman's (1978) prediction that links his delineation of three pathways of fear acquisition (conditioning, vicarious and informational) to the three systems model. He speculated that fears acquired via conditioning are likely to involve mostly behavioural and autonomic responses, while socially transmitted fears ( i . e . vicariously and informationally acquired) are li k e l y to comprise predominantly subjective responses. Rachman (1978) also postulated that different decremental processes operate for different.response systems. He suggested that the autonomic system is particularly susceptible to habituation, the behavioural system to extinction, while the subjective system is responsive to both sets of processes. It is to this issue that return of fear research is most pertinent, as in attempting to understand the processes accounting for return of fear, the mechanisms affecting change in the subjective response system must be examined. PREVIOUS RETURN OF FEAR RESEARCH It had been suggested that return of fear is neither dependent upon the amount or speed of fear reduction, nor upon i n i t i a l fear levels (Rachman, 1979). However, recent research suggests that i n i t i a l fear levels are indeed positively 7 correlated with return of fear: the more anxiety present at i n i t i a l presentations of the feared stimulus, the greater the extent of return of fear between sessions (Grayson, Foa & Steketee, 1982; Philips, 1985). An i n i t i a l but untenable explanation of return of fear relates to the notion that permanent fear reduction does not occur under conditions of brief exposure duration, because the subjective response system - which typically responds at a slower rate than other response systems - requires lengthy exposure to attain optimal levels (Gauthier & Marshall, 1977). Therefore, it may be argued that return of fear occurs only in those for whom subjective fear is insufficiently reduced by a • given treatment procedure. However, research has shown that subjective fear reduces to the same level throughout the course of treatment sessions for both those who do and those who do not display return of fear (Grey et a l . , 1979; Grey et a l . , 1981; Sartory et a l . , 1982). Several hypotheses for the return of fear phenomenon have been tested, but have received l i t t l e supportive evidence. The findings of Grey et a l . (1979) that groups assigned to high demand (close approach) and increasing demand (gradual approach) treatment conditions experienced greater return of fear than low demand groups led Grey et a l . (1981) to hypothesise the role of reactive inhibition mechanisms. Grey et a l . (1981) postulated that return of fear represented dissipation of inhibition during rest periods that had built up during treatment. As inhibition is augmented by demanding treatments, greater dissipation (or, 8 return of fear) would be expected following such treatments than following nondemanding treatments. They tested their hypothesis by comparing groups receiving massed versus distributed presentations of phobic stimuli. Contrary to their hypothesis, however, the distributed group (who should have experienced the least build up of inhibition) demonstrated the greatest amount of return of fear. Closer examination of the data revealed i n i t i a l group differences which appeared to account for the results: four of the 10 members of the distributed group displayed very high heart rates (over 110 beats per minute). When these extreme subjects were excluded from analyses, the groups did not display different amounts of return of fear. Moreover, the fact that return of fear was not evident in either group following thirty minutes of relaxation further suggested that return of fear does not represent dissipation of inhibition during rest. Grey et al.'s (1981) study tested the theoretical model of dissipation of inhibition, but it did not adequately test the role of treatment demand per se, since massed as opposed to distributed stimulus presentations would not seem to be the most legitimate contrast for a comparison between highly demanding conditions that enforce exposure to feared stimuli and less demanding conditions that allow subjects to self-terminate exposure. Although distributed and massed practice entail different durations of exposure, they enforce the same level of approach to the feared object. Treatment demand may be an important determinant of return of fear for reasons other than 9 reactive inhibition, such as the desynchrony evoked by highly demanding conditions. Therefore, Grey et al.'s (1979) more appropriate test of treatment demand (flooding versus in vivo desensitization) requires further investigation. The four extreme subjects in Grey et al.'s (1981) study experienced substantial return of fear on subjective ratings. These subjects were characterised by a discordant response pattern of very high heart rates (indicating high autonomic arousal) and low toleration for stimulus proximity, when reporting no subjective discomfort in the presence of the stimulus. It was therefore hypothesised that i n i t i a l discordance between the subjective and the remaining response systems produces return of fear, though this has yet to be directly tested. Lazarus (cited in Pppler, 1977), Barlow, Mavissakalin and Scholfield (1980) and Watson, Gaind and Marks (1972) have suggested that discordance at completion of treatment is predictive of relapse, and that synchronous improvements within each response system are necessary for long lasting treatment effects. Similarly, Hodgson and Rachman (1974) suggested that "therapeutic interventions can be considered fully satisfactory only i f neurotic responses are eliminated across cognitive, behavioural and physiological systems" (p. 322). Several approaches for dealing with response desynchrony and its effects on treatment outcome have emerged (Barlow & Mavissakalin, 1981). Some suggest that a major change in one response system w i l l produce changes in the other response systems (Bandura, 1977). Evidence for sustained 10 desynchrony, however, is substantial (Andrasik, Turner & Ollendick, 1980; Leitenberg, Agras, Butz & Wincze, 1971; Marks & Huson, 1973). Borkovec (1976) suggested that such response system generalisation occurs only when the intervention technique is specific to that response system of greatest functional importance in the emotional state. Another approach is represented by Lazarus's multi-modal therapy in which several different techniques are implemented in order to affect their various target response systems. A more efficient approach is to "evaluate the multisystem efficiency of specific therapeutic regimes" (Lang, 1977, p. 182) and to optimise those effects through determination of most beneficial treatment parameters (such as length of exposure). The return-of-fear phenomenon has been observed following both synchronous and desynchronous changes amongst the response systems during treatment. Barlow et a l . (1980), in three single case study reports of agoraphobia, recorded relapse from a subject who experienced a desynchronous pattern of diminishing subjective fear and increasing autonomic arousal. In contrast, studies specifically examining return of fear by Grey and colleagues have demonstrated return of fear following synchronous reductions in the response systems. Discordance and desynchrony require further examination in relation to return of fear. In both instances, however, return of fear may occur due to high levels of autonomic arousal per se. 11 A third return-of-fear hypothesis tested by Sartory et a l . (1982) was that high intensity exposure treatments such as flooding provoke covert rehearsal of frightening events associated with the highly fear evoking treatment, which in turn elevates arousal and produces greater return of fear. However, contrary to Sartory et al.'s (1982) predictions that instructions to think about the exposure experience following its completion (rehearsal) would result in greater return of fear (especially with high intensity exposure) than distracting or nonrehearsal instructions, nonrehearsal was associated with greater return of fear regardless of treatment intensity. It was speculated that nonrehearsal conditions may enhance return of fear by interfering with the consolidation of fear reduction experienced within the exposure session. Lack of consolidation would also explain Grayson et als. (1982) finding that distraction during exposure sessions produced greater return of fear than attention to the exposure stimulus and context. It also is possible that rehearsal facilitated longer lasting fear reduction by simply providing lengthier exposure (Sartory et a l . , 1982). If so, this would suggest that longer exposure increases the likelihood of sustained fear reduction perhaps by f a c i l i t a t i n g consolidation, and that return of fear would be characteristic of brief exposure durations. That i s , exposure duration should be sufficiently long to 1) establish the precondition of reduction in the subjective report of fear; 12 and 2) to allow consolidation of that fear reduction. 1 An important qualifying factor to Sartory et al s . finding is that rehearsal occurred under conditions of low autonomic arousal directly following exposure treatment. Rehearsal under high arousal conditions might have produced greater return of fear. The connection between attention and arousal is central to the hypotheses presented herein. While Sartory et a l . (1982) discovered means of curtailing (via rehearsal) and/or fac i l i t a t i n g (via distraction) return of fear, an attempt was made in the current study to identify subjects most likely to experience return of fear. AUTONOMIC AROUSAL The failure to observe return of fear in response systems other than the subjective system may be due to one of at least two related p o s s i b i l i t i e s . F i r s t , the physiological system may simply require more time than the subjective system to recover from habituation. Partial evidence for this possibility arises from several sources reporting minor elevations in heart rate at final treatment sessions similar to those observed in reported fear levels at earlier treatment sessions (Barlow et a l . , 1980; Grey et a l . , 1979; Sartory et a l . , 1982). Leitenberg et a l . (1971) suggested that heart rate increases were often a function of increasing demand, and that heart rate either stabilised or The importance of exposure duration was partly demonstrated in a study (Craske, Rachman & Lopatka, 1985) in which subjects who stayed in the presence of a feared snake 10 minutes following fear elimination tended to display less return of fear than those whose exposure ended immediately upon fear elimination. 13 decreased when demand level was held constant. However, the resurgence of heart rate in Grey et al.'s (1979) final treatment session occurred in groups in which demand was either held constant or increased. They attributed their result to anticipation of treatment completion. Since the subjective system is more sensitive than the physiological system (Agras and Jacob, 1981), a second related possibility is that the subjective system is more susceptible to dishabituation, and that this susceptibility is enhanced by a state of high arousal. Although empirical evidence for the relationship between autonomic arousal and return of fear is not substantial, sound theoretical models that explain the association exist. Consideration of those models will follow discussion of the empirical evidence. While Grey et a l . (1979) examined differences in treatment outcome between high and low heart rate responders, they did not specifically assess group differences in return of fear (i. e . , between-session changes). The data reported by Barlow et a l . (1980) and Grey et a l . (1981) showed greater return of fear in subjects with high heart rates during exposure. Grey et a l . (1981) attributed their results to i n i t i a l discordance between heart rate and subjective fear. However, their results could be equally attributed to arousal per se, as their four extreme subjects displayed high heart rates relative to other subjects at a l l four subjective fear levels (0, 50, 75 and 100). Also, although heart rate decreased over the session for a l l subjects, i t remained considerably higher in the four extreme subjects in comparison to the rest of 1 4 the sample at treatment completion. In an analysis of i n i t i a l discordance,.Sartory et a l . (1982) failed to observe return-of-fear enhancement in subjects displaying high heart rates when reporting no subjective distress. However, in a related research area, Rachman (1981) observed that unwanted intrusive cognitions and associated discomfort often reappear when individuals enter an agitated state. Similarly, Parkinson and Rachman (1981) reported that physiological reactions were associated with d i f f i c u l t y in removing unwanted intrusive cognitions. The general relationship between autonomic arousal and fear reduction has been extensively studied. Some researchers suggest that the success of fear reduction techniques depends on experiencing arousal during treatment implementation. Lang et a l . (1970) (cited in Grayson et a l . , 1982) stated that complete processing of feared stimuli is inferred from strong i n i t i a l physical responsiveness to those stimuli. Wolpe (1978) claimed that systematic desensitization is helpful only for those who experience fear when imagining the objects they fear. Lang (1977) found, in accordance with Wolpe, that subjects displaying minimal heart rates during systematic desensitization showed l i t t l e overall improvement. Stern and Marks (1973) also found that physiological arousal during phobic imagery correlated well with improvement. Similarly, Borkovec and Sides (1979) stated that fear reduction techniques proceed more favourably when stimulus material provokes some reaction. However, Grey et a l . (1979) were unable to show that subjects displaying large heart 15 rate responses to feared stimuli achieved greater fear reduction than those whose heart rates were relatively unresponsive. Another viewpoint suggests that high arousal interferes with therapeutic progress. Hence, Lader and Mathews (1968) argued that relaxation is essential to the success of systematic desensitization because it reduces general autonomic arousal (rather than its effect on muscle tension per se). Lader, Gelder and Marks (1967) also reported that those who habituated readily to feared stimuli were less physiologically aroused than those who did not. They also referred to a study by Gelder, Marks and Wolff (1967) in which high levels of arousal interfered with treatment progress by encouraging partial relapse between sessions of systematic desensitization. Like Lader and Mathews (1968), Grayson et a l . (1982) noted the f a c i l i t a t i v e effects of relaxation in fear reduction techniques through the removal of high levels of arousal which were believed to disrupt attentional processes, and hence interfere with progress. Borkovec and Sides (1979) also suggested that relaxation enhances functional exposure to feared stimuli by fa c i l i t a t i n g attention. This second viewpoint accords with the hypothesis that highly aroused subjects are likely to experience return of fear. However, the two viewpoints are not incompatible, as the f i r s t refers to the facilitatory effects of some degree of arousal while the second refers to the detrimental effects of excessive degrees of arousal. 16 Theoretical models that explain the connection between arousal and return of fear will now be considered. Habituation Model Autonomic arousal is an important element in habituation models of fear reduction. These will be briefly described and followed by a statement of the hypothesised habituation mechanisms operating in the return-of-fear phenomenon and their association with autonomic arousal. Habituation refers to response decrement as a result of repeated stimulation. Sensitization, on the other hand, refers to increased responsiveness to repeated stimulation (Groves & Thompson, 1970). Dishabituation refers to increased amplitude of an habituated response following presentation of either strong, extra or different stimuli (Groves & Thompson, 1970). In their formulation of the dual-process theory of habituation, Groves and Thompson (1970), like Thompson and Spencer (1966), argued and presented empirical evidence to demonstrate that a l l instances of dishabituation are examples of sensitization. Sensitization is believed to reflect and to be largely influenced by the "state" of an organism. State is an inferred construct that refers to "general levels of excitation, arousal, activation, tendency to respond, etc." (p. 440). In contrast, the process of habituation is believed to operate within the stimulus-response pathway of the central nervous system: "the most direct route through the central nervous system from stimulus to discrete motor response " (p. 440). 17 Groves and Thompson (1970) proposed that habituation and dishabituation are independent to the extent that they can occur in relative isolation from each other, but interact to produce behavioural outcomes. They presented evidence for separate neuronal substrates underlying each procesSi Interneurons in the dorsal grey matter of the spinal cord that showed only progressive decrement to repeated stimulation were called type H, as opposed to type S interneurons, that showed increment followed by decrement. The strength of habituating and sensitizing processes for any given stimulus presentation depends on the stimulus parameters of intensity and frequency (Grayson, 1982). Sensitization is more pronounced at high stimulus intensities, especially with more frequent presentations, and therefore dishabituation occurs to a greater extent with intense stimuli and rarely occurs at low stimulus intensities. Habituation, on the other hand, is inversely related to stimulus intensity and directly related to stimulus frequency. The degree to which habituation generalises across stimuli depends upon their similarity. With sufficiently different stimuli, habituation will not generalise and the sensitizing process will exert a strong influence (Groves & Thompson, 1970). The magnitude of dishabituation diminishes over repeated instances (Groves & Thompson, 1970; Smith, Dickel & Deutsch, 1978; Thompson & Spencer, 1966). It is important to note that the habituation model is a descriptive model of changes in response strength associated with specified stimulus conditions. Explanatory mechanisms are 18 not implicit in the model, and it is unlikely that a single process underlies the phenomenon (Watts, 1971). The habituation model has been used as a descriptive framework to predict responses under varying stimulus conditions during systematic desensitization (Lader & Mathews, 1968; Watts, 1971). It is conceivable that return of fear reflects a process of resensitization or dishabituation , and that, as apparent from preliminary evidence showing that return of fear gradually diminishes with repeated treatment sessions (Philips, 1985), the dishabituation gradually subsides. Rachman's (1979) hypotheses regarding return of fear accord with an habituation model of dishabituation occurring between treatment sessions. He hypothesised that return of fear: occurs after prolonged rest following exposure to stress; is more likely after long than brief rest; is likely during the early stages of mastering a threatening task; is likely i f high levels of arousal are experienced during mastery training; and, f i n a l l y , is likely i f exposure t r i a l s are brief - and hence, by not allowing sufficient time for the habituation process to reach an asymptotic level, increase the probability of dishabituation (Groves & Thompson, 1970). The habituation model can be used to explain return of fear only i f the subjective response system displays patterns of change akin to those characteristic of habituation processes. Data tend to demonstrate patterns consistent with habituation, and Rachman (1978) has suggested that habituation mechanisms to some extent account for fear reduction within the subjective 19 response system. In his statement of the differential decremental mechanisms underlying the major response systems, Rachman suggested also that reduction within the physiological system is mainly attributable to habituation. Several researchers have plotted habituation curves for the physiological and subjective response systems during exposure. Habituation patterns of the physiological system generally indicate rapid reduction in response to exposure and frequently demonstrate a linear relationship with subjective fear levels. Marzillier, Carroll and Newland (1979) reported a contrasting pattern of continuing acceleration of the physiological system in response to feared objects. However, their stimulus presentations were very brief (10 seconds) and hence the increase may have merely represented the i n i t i a l bump of sensitization typically found in habituation curves. Connolly (1979) also failed to find steady habituation curves for subjective responses. Most data, however, reflect standard habituation curves for both physiological and subjective responses during exposure to feared stimuli (Boulougouris, Rabvalis & Stefanis, 1977; Grayson et a l . , 1982; Grey et a l . , 1979, 1981; Marks, Marset, Boulougouris & Huson, 1971; Nunes & Marks, 1975; Sartory, Rachman & Grey, 1977; Stern & Marks, 1973). Evidence for linear subjective habituation was provided by Parkinson and Rachman (1980) and Rachman and. deSilva (1978) who assessed habituation rates for subjective discomfort arising from noxious obsessive thoughts. The subjective response system usually habituates at a slower rate than the physiological 20 system (Foa & Chambless, 1978; Gauthier & Marshall, 1977; Lande, 1982; Nunes & Marks, 1975; Watson, Gaind & Marks, 1972). Foa and Chambless (1978) distinguished between subjective habituation curves for imaginal and in vivo exposure conditions; suggesting that imaginal exposure is typically characterised by a curvilinear pattern due to i n i t i a l time required to fully experience the imagery, while in vivo exposure is typically characterised by a linear pattern, as the stimulus is immediately present and potent. This distinction may explain Connolly's (1979) nonlinear pattern of subjective habituation, as the mode of stimulus presentation used in his study entailed tape recorded descriptions of feared objects. Habituation and Arousal The significance of autonomic arousal will now be incorporated into the previous brief description of the habituation model and i t s account of return of fear. Research has repeatedly shown that resensitization or dishabituation is dependent on state factors and is therefore much more like l y to occur when the system is highly aroused. High arousal has also been associated with slower or delayed habituation (Deitz, 1982; Lader, Gelder & Marks, 1967; Lader & Wing, 1966). Parkinson and Rachman (1980), for example, found that discomfort felt when instructed to form an intrusive thought habituated more quickly when preceded by relaxation training. However, the return-of-fear phenomenon has not yet been accompanied by slower or delayed within-session habituation: speed of within-session habituation has not 21 differentiated those who experienced return of fear from those who did not. Rachman (1971) proposed that high levels of arousal and very intense stimuli can result in sensitization as opposed to habituation during exposure. Similarly, Groves and Thompson (1970) claimed that tonic excitability influences the degree to which responses habituate. However, subjects experiencing return of fear also achieve the same degree of fear reduction within exposure as other subjects. Return of fear seems to relate more to the phenomenon of dishabituation than to the rate or the degree of habituation. Lader and Mathews (1968) suggested that there exists a c r i t i c a l level of arousal above which habituation does not occur, but which produces a positive feedback circuit resulting in sensitization. Therefore, they viewed response to feared stimuli as a function, of two factors that determine arousal level: prestimulus levels of arousal and stimulus intensity. Lader and Mathews further claimed that habituation or improvement gained from treatment by chronically highly aroused subjects would be short-lived. They proposed that feared stimuli encountered by those subjects in the natural environment would elevate their arousal above c r i t i c a l levels, which would in turn result in response increment. Lader and Mathews' hypothesis emphasised the importance of specific encounters with feared stimuli following within-treatment exposure. Rachman (1971), on the other hand, emphasised the importance of calm conditions during exposure for enduring habituation. Similarly, Watts (1971) found that under conditions of high arousal, longer 22 stimulus presentations were required to produce more permanent habituation e f f e c t s . Brief presentations either of highly intense stimuli or to highly aroused subjects are considered countertherapeutic because the high levels of arousal that are e l i c i t e d produce s e n s i t i z a t i o n without s u f f i c i e n t time for subsequent habituation or responses decrement to take place ( M a r z i l l i e r et a l . , 1979; Watts, 1971). Whether the conditions producing return of fear occur within exposure sessions (as suggested by the hypotheses emphasising treatment demand levels and discordance and desynchrony) or following exposure and within the i n t e r v a l prior to the next session, or are the result of a subject variable that mediates response during treatment and/or following treatment (as suggested by Lader and Mathews), is yet to be determined. In summary, high lev e l s of autonomic arousal and intense stimuli that e l i c i t high arousal levels are believed to increase the l i k e l i h o o d of dishabituation. That i s , given i d e n t i c a l stimulus presentations, response increment over the strength of the response at the completion of previous exposure i s l i k e l y to occur upon the f i r s t presentation i f subjects are highly aroused or i f the stimulus i s intense. Should return of fear r e f l e c t a process of dishabituation, t h i s would explain the greater occurrence of return of fear in high-arousal subjects. Dishabituation i s also f a c i l i t a t e d by stimulus a l t e r a t i o n or novelty. In contrast, low l e v e l s of arousal and weak stimuli are believed to decrease the l i k e l i h o o d of dishabituation, which would explain the infrequency with which return of fear i s 23 observed in less aroused subjects. It is notable that vomit phobics experiencing return of fear in Philips' (1985) study were not distinguishable until an extra and apparently more potent stimulus parameter (i . e . , sound) was introduced into stimulus presentations. In accordance with the habituation model, their return of fear represents dishabituation following an "extra, strong or different" stimulus. In addition, the apparent restriction of that effect to highly aroused subjects in Philips' study suggests an interaction between the two factors identified by Lader and Mathews (1968): arousal level and stimulus intensity. Although physiological indices were not measured by Philips, subjects displaying return of fear reported greater discomfort than subjects who did not in the presence of the feared stimulus: as previously mentioned, subjective fear often linearly relates to autonomic arousal. Consolidation Model The habituation model is a descriptive model. Several underlying processes have been proposed, including stimulus-model comparisons (Sokolov, 1963), afferent-neuronal inhibition (Hernandez-Peon, 1955), alterations in synaptic transmission (Thompson & Spencer, 1966) and classical conditioning (Stein, 1966) (cited in Groves & Thompson, 1970). Each has been c r i t i c i s e d but, as Watts (1971) stated, habituation is probably a multi-process phenomenon. The proposed mechanism that seems most appropriate to the understanding of return of fear as an example of dishabituation is the memory consolidation of stimulus representations which was formulated by Wagner (1974) 24 and described by Whitlow (1975). Consolidation models explain further the nature of dishabituation in highly aroused states. In Sartory et als. (1982) study, subjects in the nonrehearsal condition spent less time thinking about the phobic object than did subjects in the rehearsal condition, and yet the former experienced greater return of fear after one week. Thinking about the exposure experience when in a relatively relaxed state appeared to f a c i l i t a t e between-session habituation. This effect could be attributed to enhanced consolidation of habituation. Consolidation models elaborate upon the habituation model as described so far by differentiating long-term and short-term habituation. Watts (1971) suggested that the two kinds of habituation involve different mechanisms, since short-term within-session habituation is facilitated by short presentations of strong stimuli; while long-term between- session habituation is facilitated by long presentations of strong stimuli. In addition, Davis (1970) claimed that short-term habituation is faci l i t a t e d by short interstimulus intervals, while long-term habituation is facilitated by longer interstimulus intervals. Gogan (1970), Groves and Lynch (1972) and Groves and Thompson (1970) go so far as to suggest that different brain areas are involved: that long-term habituation involves elaboration of the frontal cortex and hippocampus, while short-term habituation involves the brainstem reticular formation. Similarly, Foa (1979) believes that long-term habituation is dependent on higher cognitive processes. Foa claimed that habituation within 25 a session is the result of reduced autonomic response to the phobic stimulus, while between-session habituation is due to change in expectancy, or, the realisation that the phobic stimulus can be experienced without "too much fear". Expectancy change is believed to consolidate short-term into long-term habituation. Foa considers within-session habituation to be a necessary but not sufficient condition for between-session habituation. Only when both mechanisms operate can permanent habituation be expected, since without changes in cognition, dishabituation and ^sensitization are believed to result. Foa's claims are based on research conducted with obsessive-compulsive patients whose fears tend to centre upon more specific and irrational beliefs than the simple phobics with whom return-of-fear research has been conducted. Foa labels such beliefs, which border on being delusional but are considered by the patient to be r e a l i s t i c , "overvalued ideation". Obsessive-compulsive individuals with "overvalued ideation" followed the return-of-fear pattern of substantial within-session but l i t t l e between-session habituation. Foa concluded that their rigid belief systems prevented changes in expectation, and hence prevented retention of within-session habituation. Wagner (1974, cited in Whitlow, 1975) hypothesised that stimuli are primed in short-term memory as a result of either recent stimulus presentations or by retrieval of stimulus representations from long-term memory. He further stated that response decrement or habituation occurs only when a stimulus 26 matches the stimulus representation primed in short-term memory. During continuous stimulation, response decrement occurs because the stimulus is present in short-term memory and further corresponding stimulation is assumed to be rendered less effective. In accordance with the habituation model, a mismatch between the representation and the actual stimulus would in effect render the presented stimulus different, resulting in dishabituation. Long lasting habituation therefore requires long-term memory storage of an accurate stimulus representation, and retrieval of that representation into short-term memory when the same stimulus is again presented. Retrieval,is believed to occur via the association of contextual cues (surrounding the i n i t i a l stimulus presentation) with the stimulus. Storage of the stimulus representation in long-term memory and retrieval together constitute the process of consolidation. Short interstimulus intervals f a i l to produce long-term habituation because the stimulus representation and habituation initiated by the f i r s t stimulus in a sequence may not have sufficient time to consolidate i f quickly followed by a second stimulus. If information in short-term storage is continually displaced by incoming information, l i t t l e reaches permanent storage and so stimulus configurations are not consolidated. Also, the consolidation model implies that poor attention to either the contextual cues or the stimulus may impair the retrieval of stimulus representations from long-term storage, or may result in an inaccurate stimulus representation that does not match the presented stimulus coihfiguration. Therefore, mismatch and 27 dishabituation would follow 1) stimulus alteration and 2) inaccurate or irretrievable stimulus representations. The finding that those who display return of fear experience i n i t i a l fear reduction to the same extent as other subjects suggests that they do not develop less accurate stimulus representations during stimulus presentation. Given that return of fear has been observed to occur following presentation of identical stimuli, mismatch between stimulus representations and actual stimuli must result from either failure to retrieve stimulus representations or a change in stimulus representations over the interval between exposures. To explain the debilitatory effects of their nonrehearsal condition, Sartory et a l . (1982) referred to Whitlow's (1975) discussion in suggesting that distracting activity unrelated to the phobic stimulus configuration may interfere with the consolidation of that configuration into long term memory. Similarly, the consolidtion model would explain Grayson et a l s . (1982) results which showed that gains made from sessions in which attention was focused on the feared stimulus were maintained, while distraction conditions produced return of fear. In a closely related formulation, Watts (1974) tested the hypothesis that enhanced stimulus familiarisation reduces the extent to which anxiety spontaneously recovers between sessions. He found greater maintenance of within-session anxiety reduction when items in imaginal desensitization were redescribed each time they were presented than when they were not. Stimulus redescription may f a c i l i t a t e consolidation processes. 28 The possibility that high levels of autonomic arousal similarly interfere with consolidation processes will be considered next. Consolidation and Arousal There are at least two ways in which arousal may disrupt consolidation of habituation and result in return of fear: by impeding the association between a stimulus representation and contextual cues needed for later retrieval; and by modifying stimulus representations between exposures. These will be considered respectively. Information processing models are relevant to the connection between arousal and consolidation. Mandler (1975) and Lindsay and Norman (1977) claimed that the Yerkes-Dodson law is mediated by the effects of arousal on attention. This law describes an inverted U-shaped function according to which performance i n i t i a l l y improves and then deteriorates as arousal increases. A moderate level of arousal produces "focused consciousness": attentiveness, alertness, and scanning of the environment or information seeking regarding sources of arousal or generally about the present state of the environment (Mandler, 1975). At this stage, attention is believed to focus away from peripheral and towards central tasks so that performance is improved because more task relevant information is obtained. However, as arousal becomes more intense, the amount of information "that the organism can recruit effectively either from the environment or from i t s own memory store" is decreased (Mandler, 1975, p. 123). Attention becomes 29 excessively focused, as shown by Easterbrook (1959) (cited in Norman, 1976) who found that the number of cues utilised decreases as "emotion" increases. Attention may be diverted onto irrelevant details, including the signs of heightened arousal which can be very attention demanding. Similarly, Wine (1971) reported that test anxious individuals not only direct their attention internally, in contrast to the increased task-focused attention of mildly test anxious individuals, but also attend to fewer cues. Such cue restriction has serious consequences regarding attention to and performance of tasks, resulting in decreased processing efficiency and impaired performance: In states of strong emotion there is a narrowing of the cognitive organization at the moment; the individual loses broader perspective, he is no longer, able to 'see' essential aspects of the situation, and his behaviour becomes, consequently, less adaptive. (Krech and Crutchfield, 1948, cited in Korchin, 1964, p. 63) Korchin (1964) claimed that at the most intense levels of arousal, attention is characterised by distractability and disarray. However, the evidence presented by Easterbrook (1959) suggests that narrowing of focus continues. Baddeley (1972) also reported that high anxiety associated with the danger of deep sea diving produces increased arousal and narrowed attention span. The effects of increasing arousal upon attention therefore seem to proceed from greater efficiency to excessive cue restriction. 30 The consolidation model implies that failure to attend fully to stimulus settings during exposure mitigates retrieval of corresponding stimulus representations at subsequent stimulus exposure. The absence of a matching stimulus representation primed in short-term memory at subsequent exposure would result in dishabituation. Information processing models suggest that heightened arousal excessively narrows the focus of attention. Consequently, highly aroused subjects may focus excessively on sections of the exposure setting and therefore f a i l to be fully aware of the contextual cues for which a wider range of attention is required. Retrieval of the stimulus representation at subsequent exposure may be impeded i f highly aroused subjects do not attend fully to contextual cues surrounding phobic stimuli. Therefore, according to the consolidation paradigm, return of fear can be viewed as dishabituation that is enhanced by a highly aroused state, since the latter interferes with attention to contextual cues during exposure that f a c i l i t a t e retrieval of stimulus configuration representations. This explanation of the role of autonomic arousal in return of fear focuses upon processes occurring within or immediately following exposure sessions rather than some time after. Thought Resensitization The second proposed way in which arousal impedes consolidation and enhances dishabituation is by altering stimulus representations in the interval between exposures. This involves examination of the effects of covert rehearsal of 31 previous exposures when subjects enter aroused states. It is speculated that heightened physiological arousal provokes thoughts of a sensitizing nature that effectively alter the cognitive representation of the previous exposure. Hence, upon re-presentation of the stimulus, the retrieved stimulus representation does not match the actual stimulus, and dishabituation occurs. Such dishabituation is reflected in a resurgence of subjective fear. The connection between heightened arousal and sensitizing thoughts is equivocal, but i t seems likely that high arousal levels promote more frequent thoughts about the fearful nature of both the phobic stimulus and one's responses during previous exposure than does the state of low arousal. Borkovec (1976) posited that one way in which fear states are maintained is through the production of fearful cognitions arising from a state of pronounced physiological arousal. In examining the hypothesis that phobic subjects, in general, engage in self verbalisations of a fear-inducing nature, Rimm, Landa, Lancaster, Nahl and Dittmar (1977) recorded thoughts stated aloud whilst subjects imagined approaching a phobic situation: very few thoughts were objective, and most involved escape or avoidance, catastrophy and an awareness of fear.. However, Rimm et a l . did not analyse the frequency of such thoughts in more as opposed to less physiologically aroused subjects. Meyer and Reich (1978) also noted that cognitive re-experiencing or rehearsal of "stressful situations" may account for anticipatory anxiety. Similarly, Stern and Marks (1973) suggested that 32 intervals between treatment sessions allowed agoraphobics time to think about their fear and that cognitive rehearsal between periods of exposure might affect incubation processes. Such thoughts may alter the representation of the previous exposure, and, according to Wagner's consolidation model, the resulting mismatch would produce dishabituation. It is important to note that the role of sensitizing thoughts in return of fear may be explained without incorporating a memory representation consolidation paradigm. Cognitive rehearsal of more fearful stimulus encounters than actually occurred, in i t s e l f , represents exposure (covert) to a different stimulus. The habituation model states that dishabituation follows the presentation of extra, strong or different stimuli (Groves & Thompson, 1970). Similarly, Rachman (1971) suggested that habituation to obsessional material may be disrupted by the occurrence of an intense or different stimulus in the form of a thought that causes dishabituation and hence a return in subjective disturbance when ruminative thoughts are re-experienced. Foa (1979) also observed that obsessional patients with "overvalued ideation" displayed within-session but not between-session habituation of subjective reports of fear. 2 As previously mentioned, Foa attributed this pattern to a strong and resistant belief that the object is contaminated 2 This was also observed in an obsessive-compulsive patient whose belief that she had cancer, despite contrary medical evidence, was very strong and whose fear associated with exposure to "harmful objects" showed within-session habituation and between-session dishabituation (Rachman & Craske, 1984). 33 and/or harmful; a type of belief that appears very similar to sensitizing thoughts of the fearful nature of an object or situation in phobic states. However, the two are not synonomous, since obsessive beliefs tend to be more specific and irrational than phobic cognitions. In addition, while evidence suggests that phobic return of fear gradually diminishes with repeated exposure (Philips, 1985), Foa's obsessive patients who had overvalued ideation continued to seek reassurance and to be disturbed by fears and doubts. Other evidence that suggests that sensitizing thoughts are elic i t e d during aroused states derives from studies already mentioned which indicate that anxiety, stress and agitated mood promote frequent intrusive cognitions (Parkinson & Rachman, 1980, 1981; Rachman, 1981). Arousal-induced sensitizing thoughts seem to resemble the emerging phenomenon in research analysing mood and memory: just as depressed mood appears to result in easier access to unpleasant or sad thoughts and memories (Bower, 1981), so may a physiologically aroused state result in a propensity to focus upon fearful qualities of encounters with feared stimuli as opposed to the less fearful qualities. Bower (1981) examined recall for pleasant and unpleasant events one week following their occurrence. He found that pleasant mood at time of recall resulted in recall of a greater percentage of pleasant than unpleasant events. Conversely, unpleasant events were recalled more readily than pleasant events when in an unpleasant mood. Moreover, unpleasant 34 incidents were rated as more unpleasant than at time of occurrence when recalled while in a depressed mood. Bower (1981) proposed that the tendency to recall more unpleasant events and to intensify their unpleasantness feeds back to intensify depressed mood. Similarly, i t is suggested that enhanced focus upon the fearful nature of encounters with feared situations in states of high arousal intensifies the fear state (reflected in the phenomenon of return of fear). It is suggested that highly aroused subjects are more likely to experience thought resensitization than are less aroused subjects due to the intensity of their physiological responses. Bower (1981) showed that more intense events are recalled more easily than less intense events. Similarly, Robinson (1980) showed that intense personal memories were recalled more easily because intensely arousing experiences evoke greater attention and more extensive encoding than weak experiences. Robinson (1980) also cited evidence to show that "more intensely affecting experiences are rehearsed more frequently than mildly arousing ones during the retention interval. The memory of an intensely arousing experience could be elaborated during rehearsal and thereby made more distinctive ... enhance the accessibility of those memories"(p. 171). The evidence therefore suggests that intense experiences are rehearsed more often, recalled more easily and judged to be more intense at time of recall than less intense experiences. Tversky and Kahneman's (1982) availability heuristic would seem to underlie such phenomena: salience increases the retrievability of 35 instances stored in memory, and instances that are retrieved more easily are judged as more numerous than instances that are retrieved less easily. Since highly aroused subjects have more intensely arousing confrontations with feared situations than low-arousal subjects, the former's recall of their previous experiences should be biassed to a greater extent, and they should demonstrate greater thought resensitization than less aroused subjects. The assumption that increased SUDS ratings ( i . e . , subjective units of discomfort scale) directly reflect the presence of more thoughts of a sensitizing nature is partly substantiated by Likierman and Rachman's (1980) observation that the somewhat variable nature of habituation curves for SUDS ratings is probably due to the influence of extraexperimental and internal factors. As previously mentioned, Sartory et a l . (1982) attempted to assess the frequency of sensitizing thoughts in subjects who were highly aroused, but succeeded only in assessing thoughts under low arousal conditions. They differentiated high and low heart rate responders on the basis of heart rate responses to the presence of the phobic stimulus at distances that el i c i t e d subjective ratings of zero. Despite overall reductions in arousal, the heart rates of high heart rate responders remained higher than those of low heart rate responders, and yet the content of their thoughts (categorised as mildly uncomfortable, uncomfortable or definitely uncomfortable moments of the exposure) did not d i f f e r . Their data suggest that individuals 36 with high heart rates when reporting no subjective fear in the presence of the phobic stimulus are not more l i k e l y than those with low heart rates to experience s e n s i t i z i n g thoughts immediately following exposure. However, their analysis i s very r e s t r i c t i v e , and does not exclude the following p o s s i b i l i t i e s : i ndividuals displaying high heart rates when reporting no subjective fear experienced more s e n s i t i z i n g thoughts than did low heart rate responders at a l a t e r time; individuals displaying both high phasic heart rates and moderate or higher subjective fear experience more s e n s i t i z i n g thoughts than low heart rate responders with the same fear l e v e l s ; and individuals with high tonic heart rates experience more s e n s i t i z i n g thoughts than individuals with low tonic heart rates. At what point s e n s i t i z i n g thoughts may arise within the i n t e r v a l between exposure sessions i s not known. Bower's (1981) discussion of mood-dependent memory emphasises processes occurring when the same state i s again entered: at subsequent exposure in the return-of-fear paradigm. It i s assumed that the assessment of thoughts occurring immediately after exposure may not y i e l d comparable results to the assessment of thoughts occurring midway through the i n t e r v a l or immediately preceding the next exposure session. Resensitization of the subjective system may occur when an aroused state i s again entered in a n t i c i p a t i o n of exposure, or, sometime between exposures when subjects cognitively rehearse the exposure experience whilst in an aroused s t a t e . Chronically highly aroused individuals may therefore tend to experience such r e s e n s i t i z a t i o n over the 37 interval. The relationship between return of fear and tonic arousal requires examination. According to the above, greater return of fear would be expected in highly tonically aroused subjects. It is assumed that either phasic arousal levels habituate but remain at an absolute level that is s t i l l sufficiently high to provoke sensitizing thoughts, or that tonic levels are high and can therefore also provoke sensitizing thoughts. Partial evidence for the role of phasic, stimulus specific arousal levels is found in previous return-of-fear research (Grey et a l . , 1981) in which the physiological system was observed to undergo and maintain the decremental effects of habituation, yet absolute heart rate levels following habituation were s t i l l higher in subjects classified as highly aroused in comparison to less aroused subjects. No data are yet available with respect to the role of tonic arousal, although, as previously mentioned, Lader and Mathews (1968) suggested that general arousal levels determine the probability of relapse. While Grey et al.'s (1981) examination of heart rate when subjective fear was recorded as zero may approximate tonic levels, heart rate was s t i l l measured in the presence of the stimulus and hence represented a stimulus-specific response. 38 PERCEIVED SKILL Self-perceived s k i l l i s a second factor proposed herein as an important predictor of return of fear in cases where fear reduction strategies e n t a i l s k i l l a c q u i s i t i o n as opposed to passive exposure. Rachman (1979) hypothesised that greatest return of fear occurs when s k i l l a c q u i s i t i o n i s involved in the fear reduction process, such as during the early stages of mastery or after long periods without pr a c t i s e : Examination of the n o n - c l i n i c a l l i t e r a t u r e suggests that the return of fear i s e s p e c i a l l y l i k e l y to occur i f the f e a r f u l person i s required to develop s k i l l e d or coping behaviour as part of the fear reduction process (eg., overcoming a fear of drowning by learning to swim, overcoming a fear of f l y i n g by learning how to control an a i r c r a f t e t c . ) . Furthermore, in these instances, the fear is prone to return i f the s k i l l e d behaviour i s not practised re g u l a r l y . (Rachman, 1983) Lack of practise usually results in at least p a r t i a l loss of s k i l l and hence reduced sense of competency, or perceived s k i l l , since the l a t t e r i s believed to be mainly dependent on s k i l l (Rachman, 1978). Consequently, lack of practise may increase the l i k e l i h o o d that fear w i l l return, given an inverse linear rel a t i o n s h i p between perceived s k i l l and fear r a t i n g s . Therefore, the hypothesis (Rachman, 1979) that return of fear should occur more often when a new s k i l l i s being mastered not only f i t s with an habituation model but also a perceived s k i l l theory of return of fear. The connection between s k i l l , p ractise and return of fear relates to data from studies with agoraphobics indicating that between-session "homework" practise 39 of coping s k i l l s taught during treatment sessions is c r i t i c a l for successful outcome (Marks, Grey, Cohen, H i l l , Mawson, Rimm & Stern, 1983). In contrast are fears for which s k i l l acquisition is not a necessary part of fear reduction. For example, Kleinknecht (1982) reported that most spontaneous remissions of spider fears were attributable to acquisition of knowledge alone. Self-Efficacy Theory The perceived-skill hypothesis arises out of Bandura's self-efficacy theory/ f i r s t proposed in 1977. Self-efficacy refers to one's conviction or confidence in one's own abil i t y to perform a certain task.. It is an inference about one's abili t y to cope based on the consideration of several sources of information (Norton, Dinardo & Barlow, 1983). .In accordance with a social learning theory framework, in which cognitive processes are attributed a prominent role in the acquisition and retention of behaviour, efficacy expectations are claimed to influence choice of both activities and environmental settings, and to determine degree of both effort expended and persistence in the face of obstacles or aversive experiences (Bandura, 1977, 1982). That i s , self-efficacy is believed to influence coping behaviour in threatening situations. In addition, judgements of capabilities also influence thought patterns and emotional reactions. Poor self-efficacy is believed to result in excessive dwelling upon personal deficiencies, that in turn produce stress and arousal (Bandura, 1982). 40 One of the four main sources of information on which self-efficacy judgements are made is. performance accomplishment. Performance provides evidence of personal achievement and develops s k i l l s for individuals to deal with threatening situations (Bandura, 1977). Therefore, lack of practise of s k i l l s and hence lack of performance accomplishment would be associated with uncertainty conerning one's own s k i l l s and, therefore, reduced self-efficacy. The remaining sources of information are vicarious experience, verbal persuasion and physiological state, a l l of which are claimed to be uniquely relevant to the different fear reduction techniques. Bandura (1977) therefore claimed that a l l therapeutic modes operate by raising and strengthening percepts of self-efficacy. Moreover, enactive mastery (or, performance accomplishment) is considered the most effective method of raising self-efficacy. Hence, performance accomplishment produces more improvement in approach behaviour, physiological arousal and reported fear than other fear reduction techniques (Bandura, 1977; Biran & Wilson, 1981). Research has repeatedly demonstrated a relatively strong relationship between self-efficacy and performance: increasing self-efficacy correlates with progressively higher performance accomplishment (Bandura, 1980, 1981, 1982; Bandura, Adams & Beyer, 1977). Bandura's claim that lowered self-efficacy is associated with heightened subjective fear is also empirically substantiated, but the relationship between self-efficacy and physiological arousal awaits further investigation (Lang, 1978; Wolpe, 1978; Craske & Craig, 1984). 41 While Bandura claimed that self-efficacy mediates each of the response systems, others argue that self-efficacy is significantly related only to the behavioural (Wolpe, 1978) or to the behavioural and subjective response systems. Kendrick, Craig, Lawson and Davidson (1982) found that self-efficacy correlated with only behavioural outcome measures in their treatment study. Craske and Craig (1984) found that self-efficacy correlated with only subjective measures in their assessment study. In fact, some have suggested that the relationship between efficacy and subsequent action arises solely from the influence of preceding action on self-efficacy judgements (Feltz, 1982). To refute this, Bandura (1982) has reported the accuracy of self-efficacy judgements which are based purely on vicarious information. However, Biran and Wilson (1982) found that the close congruence between self-efficacy and subsequent approach behaviour in exposure treatment was not replicated in a cognitive treatment group (which did not entail behavioural practise) where subjects overestimated their approach behaviour. Rachman (1978), interpreting self-efficacy as an elaborated version of the sense of controllability, also viewed self-efficacy as exerting predominantly behavioural effects. Controllability is seen as a major determinant of courageous behaviour; "courageous" is defined as the repeated approach and exposure to objectively dangerous situations in spite of high levels of autonomic arousal and subjective fear (Hallam & Rachman, 1980). Courageous behaviour, like self-efficacy, is 42 also believed to be largely dependent upon competence; training and practise lead to a growing sense of self-efficacy and reinforce courageous behaviour (Rachman, 1978). Rachman further claimed that the reduction in subjective and autonomic arousal accompanying repeated episodes of courageous behaviour is due to habituation processes rather than heightened self-efficacy. However, Rachman did not exclude the influence of self-efficacy upon subjective reports of fear. When postulating the mechanisms underlying fear reduction within each response system, Rachman (1978) stated that the physiological system is most susceptible to habituation processes, the behavioural system to extinction processes, and while the subjective system may be open to both habituation and extinction processes, i t is also modified by self-efficacy - which in turn is likely to be facilitated by s k i l l f u l instruction and training. An interesting connection arises between perceived s k i l l or sense of controllability and habituation. Ginsberg (1982) has demonstrated that the experience of unpredictability in stimulus presentations produces dishabituation. Similarly, Davis (1970) reported that habituation is facilitated by predictable fixed-ratio interstimulus intervals. Predictability would seem closely connected to a sense of con t r o l l a b i l i t y . Bandura (1982) also stated that predictability and controllability enhance self-efficacy, since predictability decreases stress and increases preparedness in coping with threats. 43 Therefore, it is suggested that not only heightened autonomic arousal but also low perceived s k i l l increase the likelihood of return of fear. Self-efficacy has been used as a predictor of relapse in addictive disorders, but not in the specific context of return of fear. Marlatt and Gordons' (1979) cognitive-behavioural model of addictions emphasises the relationship between self-efficacy and a b i l i t y to cope with high-risk situations following treatment. DiClemente (1981) examined the relationship between self-efficacy for smoking cessation reported at post treatment with maintenance at a five-month followup assessment. While the subjects who were successful abstainers (66%) did not differ from those who relapsed on either demographic measures or smoking history variables, they did report higher self-efficacy scores (degree of certainty that they would not smoke in 12 high-risk situations) at post treatment. However, even those who relapsed reported very high self-efficacy ratings at post treatment. Consequently, group mean self-efficacy ratings were similar (73.9/84 versus 66.7/84). Condiotte and Lichenstein (1981) reported a a moderately strong relationship (r=0.57) between treatment enhanced self-efficacy and five-week followup maintenance of smoking abstinence. Self-Efficacy and Arousal Bandura (1978) claimed that arousal level and self-efficacy are inversely related. He suggested that physiological state serves as a basis for judgements concerning one's level of anxiety and for inferences about performance capabilities: 44 Because high arousal usually debilitates performance, people are more inclined to expect success when they are not beset by aversive arousal than if they are tense and viscerally agitated. (Bandura, 1982, p. 129) In an apparently tautological argument, Bandura stated that self-efficacy in turn affects levels of physiological arousal. Self-reported fear has been shown to decline as the strength of self-efficacy increases (Bandura & Adams, 1977; Bandura, Adams & Howells, 1980; Bandura, Reese & Adams, 1982). Moreover, in a 1980 study described in his 1982 review, heart rate and blood pressure were measured while subjects performed tasks associated with different levels of self-efficacy. Self-efficacy for each task was then raised through enactive mastery to maximal levels and physiological indices again measured: Subjects were viscerally unperturbed by tasks which they regarded with utmost self efficaciousness. However, on tasks about which they were moderately insecure concerning their coping inefficacy their heart rate accelerated and their blood pressure rose during anticipation and performance of the a c t i v i t i e s . After self-percepts of efficacy were fully strengthened these same task demands were managed unperturbedly. (Bandura, 1982, p. 138-139) It seems unlikely, however, that physiological relaxation was due solely to the enhancement of self-efficacy: enactive mastery per se habituates physiological responses. Bandura, Taylor, Williams, Mefford and Barchas' (1985) study of the relationship between self-efficacy and catecholamine levels - a neurochemical substrate of physiological arousal - is subject to the same criticism, as the elevation of self-efficacy to maximal levels via participant modeling is likely to have resulted in 45 habituation of the subjective and physiological response systems. Reduced catecholamine secretion following such exposure can, therefore, be attributed directly to reduced magnitude of the autonomic response rather than to an effect mediated by enhanced self-efficacy. The causal connection between arousal and self-efficacy has yet to be empirically demonstrated. Whether these two factors interact to yield multiplicative effects is not known. It is proposed that they at least act separately to enhance the return of fear. Hence, greater return should be evident from the additive effects of both variables than from the singular effect of either variable, which, in turn, should produce greater return of fear than in the absence of both variables. 46 EMOTIONAL PROCESSING A final comment will be made on the connection between return of fear and Rachman's (1980) concept of emotional processing. Emotional processing refers to the process of absorption of emotional disturbances to the extent that other experiences and behaviour can proceed without interference. The concept of emotional processing was based i n i t i a l l y on the work of Freud in his description of Anna 0. in which he claimed that the state of discordance resulting from suppression of the cognitive and verbal expression of fear and grief led to later d i f f i c u l t i e s . A second, more recent basis for the concept derives from the evidence (as previously discussed) indicating that successful outcome is dependent on the experience of fear during treatment sessions: hence the notion that fear must be experienced before i t can be reduced, or, processed. Fear-reduction techniques based on exposure are viewed as providing a means of absorbing persistent and disruptive emotional reactions. Return of fear can therefore be viewed as a sign of incomplete emotional processing; a phenomenon whose presence is detected by examining reactions when relevant stimuli are re-presented. Gains made in fear levels from previous presentations of relevant stimuli would be maintained if emotional processing had been complete. Failure in emotional processing, on the other hand, would result in the return of fear. The return of fear after an interval during which the new learning of fear is unlikely to have taken place, 47 implies that at least some of the fear-reduction originally accomplished, must have been transient and in that sense, incompletely processed. (Rachman, 1980, p. 54) It is therefore not surprising to find that some of the same factors believed to mediate emotional processing are also believed to influence return of fear. Rachman (1980) hypothesised that high levels of arousal and low levels of perceived control are detrimental to emotional processing, while controlled autonomic reactions and a sense of perceived control, or, self-efficacy, have facilitatory effects. Moreover, Grayson et a l . (1982) claimed that distraction produces return of fear by impeding emotional processing, whereas attention focused onto the feared stimulus facil i t a t e s emotional processing. Grayson et a l s . comment relates to the usefulness of distraction as a coping strategy for exposure. Rachman (1976) suggested that the treatment of obsessive-compulsive patients is enhanced by distraction techniques. Mavissakalin, Michelson, Greenwald, Komblith and Greenwald (1983), however, in accordance with the consolidation model, suggested that distraction can block effective exposure. Finally, the "stimulus factors" hypothesised to result in poor emotional processing (Rachman, 1980) are in common with those believed to f a c i l i t a t e dishabituation: sudden, intense and unpredictable stimuli. 48 STATEMENT OF THE PROBLEM In summary, the present study was an investigation of the relationships between both autonomic arousal and perceived s k i l l and return of fear. Though the existing data are sparse and somewhat contradictory, they suggest that greater return of fear occurs in individuals who are highly aroused than in individuals with low levels of arousal. The return-of-fear phenomenon was hypothesised to reflect a process of dishabituation that occurs between exposure sessions and is enhanced by heightened autonomic arousal of either a tonic or phasic nature. It was further speculated that such dishabituation occurs as a result of lack of consolidation of within-session habituation. Lack of consolidation in highly aroused individuals may occur due to 1) failure to attend to contextual cues during exposure, and subsequent impairment of the retrieval of stimulus representations, or 2) thoughts of a sensitizing or fearful quality that alter stimulus representations. Sensitizing thoughts were assumed to develop in individuals who are highly aroused when those individuals cognitively rehearse their exposure to feared situations between actual exposures. Therefore, two sets of processes that produce dishabituation, or return of fear, were postulated: one associated with memory consolidation, and the other with thought content. Each process is f a c i l i t a t e d by high autonomic arousal, yet initiated at a different time. 49 Perceived s k i l l was hypothesised to be another factor that may be used to predict return of fear in individuals for whom fear reduction entails s k i l l acquisition. The significance of perceived s k i l l was propounded on the bases of theory and empirical evidence which suggested that s k i l l and practise are major determinants of self-perceived s k i l l , which, in turn, influences subjective ratings of fear. It was hypothesised that individuals with low levels of perceived s k i l l would experience greater return of fear than individuals with high levels of perceived s k i l l . Return of fear is examined by the assessment of individuals' fear ratings over the course of fear-reduction programs. Comparisons between individuals who i n i t i a l l y display high levels of autonomic arousal versus low levels of autonomic arousal and between individuals who i n i t i a l l y display low levels of.perceived s k i l l versus high levels of perceived s k i l l were necessary in order to test the foregoing hypotheses. Those comparisons were achieved by classifying each subject's i n i t i a l responses to a feared situation as above or below the median of the entire sample's responses. A combination of each subject's median classifications with regard to autonomic arousal and perceived s k i l l yielded four classification groups (high, arousal, low perceived s k i l l ; high arousal, high perceived s k i l l ; low arousal, low perceived s k i l l ; and low arousal, high perceived s k i l l ) . According to the hypotheses, the comparison between the f i r s t and last groups yields the strongest contrast, regardless of whether the effects of the classification 50 variables are additive or multiplicative. Assessment of the responses of each group to a fear-reduction program was used to indicate relative amounts of return of fear, given the same amount of fear reduction. The rationale for testing the hypotheses associated with the current study in a sample of anxious musical performers is presented in the subsequent section. The types of return of fear assessed and the specific fear-reduction techniques employed are also discussed in more deta i l . Return of fear is detected by measuring the three major fear response systems (behavioural, physiological and subjective) at repeated exposures to feared situations. Examination of the specific processes speculated in the current investigation involved the additional assessment of: subjects' memory for the setting and their responses to prior exposures; and, the nature of their thoughts with regard to the feared situation before and after exposures. These measures are described in the next chapter. Experimental Hypotheses 1. Greatest return of fear will be displayed by subjects demonstrating high autonomic arousal and low perceived s k i l l . 2. Least return of fear will be displayed by subjects demonstrating low autonomic arousal and high perceived s k i l l . 3. Greater return of fear will be displayed by subjects with high autonomic arousal than by subjects with low autonomic arousal. 51 4. Greater return of fear w i l l be displayed by subjects with low perceived s k i l l than by subjects with high perceived s k i l l . 5. Greater thought r e s e n s i t i z a t i o n w i l l be reported by subjects with high autonomic arousal than by subjects with low autonomic arousal. 6. Poorer memory for previous exposures w i l l be demonstrated by subjects with high autonomic arousal than by subjects with low autonomic arousal. 7. Greater return of fear w i l l be displayed by subjects with high tonic arousal than by subjects with low tonic arousal. 52 MUSICAL PERFORMANCE ANXIETY Testing of these ideas required a task of s k i l l which demands repeated practise by individuals who often report fear. The current investigation was conducted on musical performance anxiety. Anxious musicians constitute a population in whom return of fear would seem to be especially significant, since return of fear is more prone to occur when skilled behaviour is not practised regularly. S k i l l in this case refers not only to musical a b i l i t y , but also to the s k i l l acquired in learning anxiety-reduction strategies such as relaxation and concentration. The findings would generalise more to other types of c l i n i c a l disorders in which treatment involves the acquisition of specific coping s k i l l s (such as agoraphobia and social phobia).than to disorders for which treatment involves mainly passive exposure. The latter kind of treatment characterises previous return-of-fear research. Although musical performance anxiety may be classified as an analogue fear to the extent that the subjects are not drawn from c l i n i c a l samples, Borkovec and Rachman (1979) argued,that the criterion of c l i n i c a l status may be inappropriate and that if the target population is similar to patients on c r i t i c a l dimensions, then the validity of generalising is likely to be increased. The f i r s t c r i t i c a l dimension they delineate is type of e l i c i t i n g condition . While musical performance situations do not represent the same fear e l i c i t o r s typical of many fearful c l i n i c a l patients, they do resemble the el i c i t o r s of performance 53 fears held by those seeking treatment for social phobias such as public speaking, and test anxiety. The second c r i t i c a l dimension is intensity of response, which is especially important as "mild fears are subject to a wide range of unwanted influences" (Borkovec & Rachman, 1979, p. 256), and hence threaten both internal and external v a l i d i t y . Musical performance anxiety can involve quite intense and debilitating fear because performances are demanding situations from which individuals can not readily escape. Kendrick et a l . (1982) observed that musical performance anxiety is of a pervasive nature and generally entails worry about performance, preoccupation with feelings of inadequacy and heightened somatic and autonomic arousal. The characteristically strong component of physiological arousal reduces the susceptibility of emotional reactions to extratherapeutic or demand influences. The final c r i t i c a l dimension is motivation to participate in the therapeutic enterprise. Motivation levels may be ensured by selecting only very fearful subjects, for whom the behaviour in question presents a problem, and by the exclusion of incentives other than therapeutic benefit. Performance anxiety seems particularly suitable for this area of research since i t entails a population in which there exists a wide range of s k i l l and for whom anxiety is a concern. It would be useful to compare the obtained results with those from a population for whom s k i l l acquisition is not essential to fear reduction, because, i f the prediction concerning perceived s k i l l is correct, perceived s k i l l should be a less effective 54 predictor of return of fear in those populations than in populations in whom fear arises from the performance of skilled tasks. BETWEEN-SESSION AND LONG-TERM RETURN OF FEAR The objective was therefore to obtain a group of students experiencing maladaptive levels of musical-performance anxiety and to provide them with the best available training program for reducing anxiety. The extent to which return of fear occurred, both during intervals between sessions and over a followup period, in groups differentiated on the bases of autonomic arousal and perceived s k i l l could then be examined. Followup assessment of return of fear after standardised treatment procedures has not been examined. Philips (1985) found that when the number of treatment sessions was determined by each individual's report of fear and the sessions were continued until dishabituation disappeared, there was l i t t l e evidence of long-term return of fear. An assessment of long-term return of fear following a treatment paradigm standardised across a l l subjects is required, as is an examination of the relationship between short-term (between-session) and long-term (followup) return of fear under that treatment paradigm. The notion of long term-return of fear contradicts to some extent Hodgson and Rachman's (1974) hypothesis that concordance among the three major response systems increases during followup. They attributed increased concordance to either improvement in subjective measures or deterioration in behavioural measures. However, return in subjective fear may 55 have been masked in previous research by failure to classify subjects on important variables. Subject classification on the basis of perceived s k i l l and arousal may reveal a pattern of increasing subjective fear in some subjects. It is interesting to note the connection in this respect with depression relapse rates. Kovacs, Rush, Beck and Hollon (1981) examined the effectiveness of a 12-week therapy for depression at a one-year followup. Both cognitive and psychopharmacological treatments reduced symptomatology at postassessment, but one-year relapse rates were quite high. When the relapse criterion was a score of 16 or more on the Beck Depression Inventory, 39% of the cognitive group and 65% of the pharmacological group relapsed. When the c r i t e r i a were either scores on the Beck Depression Inventory or return to treatment, the rates were 56% and 82% respectively. In addition, Beck Depression Inventory scores at one year followup were positively correlated with both number of previous depressive episodes and age. In accordance with previous theorising with regard to return of fear, i t can be speculated that practise per se may also influence long term return of fear. According to Rachman (1979), those who obtain a high level of performing experience (or, practise of ski l l s ) over the followup interval should experience l i t t l e return of fear, as lengthy intervals between exposures should lead to return of fear. Lack of practise is a factor believed to impede emotional processing as well (Rachman, 1980). 56 FEAR REDUCTION PROGRAMS Several psychological treatment studies have been conducted in the area of musical-performance anxiety with some success. Psychopharmacological methods were not considered for the current study, although Neftel, Adler, Kapelli, Rossi, Dolder, Kaser, Bruggesser and Vokauf (1982) claimed that while beta blockades did not reduce anticipatory anxiety, they were effective following the onset of pronounced sympathetic activity in improving technical musical performance. However, Neftel et a l . (1982) did not assess overall musical a b i l i t y . There are several psychological treatment approaches in the area of musical-performance anxiety including musical/technical s k i l l s training, systematic desensitization and relaxation training, and cognitive/attentional s k i l l s training (Decker & Russell, 1981). Kendrick et a l . (1982) implemented a combined program of behavioural rehearsal and attentional training and achieved fear reduction in behavioural and self-report measures, although physiological arousal was not affected. Relaxation would therefore appear to be a useful technique to include, especially as performance anxieties are usually conceptualised as involving both worry and emotionality components (Goldfrled, Lineham & Smith, 1978; Wine, 1971). Morasky, Reynolds and Sowell (1983) found that reduced muscular tension (via EMG biofeedback) did not reduce performance quality and in some cases improved quality, indicating that inducement of physical relaxation can benefit musical performance. Kendrick et a l . 57 (1982) cited previous studies that included a relaxation component that successfully reduced autonomic indices of arousal: Wardle (1975) found that systematic desensitization and relaxation reduced heart rate and observable signs of anxiety; Appel (1976) also found that systematic desensitization improved arousal, self report and performance measures. However, systematic desensitization had no effect in Goldstein's (1975) study (cited in Kendrick et a l . , 1982). In addition, the component of attentional training, or task focusing self-instruction training, is considered to be very important in any performance anxiety program. Thyer, Papodorf, Himle, McCann, Caldwell and Wickpert (1981), for example, in their treatment of speech anxiety, considered procedures that reduced task irrelevant thoughts and provided s k i l l s to remain task focused to be the core treatment elements. Sweeney and Horan (1982) combined cue-controlled relaxation and attentional training in the reduction of musical-performance anxiety. They found that both techniques reduced pulse rate and two self-report measures of anxiety. Attentional techniques were uniquely effective in improving musical competence, while relaxation reduced overt signs of anxiety. Therefore, the current investigation included both relaxation and attention training as s k i l l s to be acquired and practised in a behavioural rehearsal format for reducing musical-performance anxiety. 58 Design The design consisted of repeated measures on four pre-selected groups differentiated on two factors: autonomic arousal (high versus low) and perceived s k i l l (high versus low). Assessment was continuous throughout the training program and also occurred at pre-assessment, post-assessment and followup assessment. 59 METHOD Subjects The sample consisted of pianists, v i o l i n i s t s and vocalists from the music department of the University of British Columbia and the Vancouver locale. Subject selection was based f i r s t l y on responses to a screening questionnaire developed by Sweeney and Horan (1982) for identifying musical performers who experience anxiety to a degree that interferes with the quality of their public performances (see Appendix 1). Musicians who experience anxiety only to an extent that facilitates performance quality or who experience only debilitating anxiety when they are insufficiently musically prepared are screened-out by the use of this questionnaire. Scoring c r i t e r i a were those used by Sweeney and Horan (1982): "yes" responses to items 1, 2 and 6; four "yes" responses to item 4; and a minimum of nine total "yes" responses. Sweeney and Horan's item 7 - "Does your anxiety stand in the way of your majoring in music?" - was excluded, as the population for the current investigation extended beyond college and university music students (Sweeney and Horan required a minimum of ten total "yes" responses). In addition, in accordance with Bernstein and Paul's (1971) recommendation, only performers who said "yes" to the question "Are you interested in acquiring strategies for dealing with performance anxiety?" were contacted. This question accompanied the screening questionnaire which is presented in Appendix 1. Subject selection was based secondly upon self-reported anxiety 60 level during the preassessment performance: only those reporting a score of 50 or more on the Subjective Units of Discomfort scale were included. Eighty-nine musicians expressed interest in participating in the program. Seven of these did not meet screening questionnaire c r i t e r i a and another seven withdrew prior to pre-assessment. Of the remaining 75 subjects, two failed to report SUDS of 50 or more during pre-assessment, and another withdrew following preassessment because of other time commitments. Seven participants withdrew during the training program: five due to other time commitments, one due to ill-health and one who found the behavioural rehearsal component of the training program too demanding. Finally, two subjects could not be assessed at followup as they had moved away from Vancouver. Twenty-six of the i n i t i a l 89 musicians were either not suitable or withdrew from the study. The final sample consisted of 63 musicians: 42 pianists, 16 vi o l i n i s t s and 5 vocalists. Ages of the subjects ranged from 15 to 53. The sample consisted of 51 females and 12 males - a distribution similar to the one constituting Kendrick et a l s . (1982) sample, which was described as being f a i r l y representative of the proportion of females to males studying music. Number of years of music experience ranged from three to 40. Grade levels (or, levels of ability) in the group of pianists ranged from grade 7 to A.R.C.T. (according to the Toronto Royal Conservatory of Music ratings, in which A.R.C.T. represents the highest level of accomplishment) to performance 61 majors in a university program. Violinists ranged from grade 9 (Toronto Conservatory) to performance majors. Vocalists ranged from one to three years of voice training in a college performance program. The four classification groups were formed on the basis of orthogonal median splits of i n i t i a l heart rate and i n i t i a l perceived s k i l l : high heart rate, low perceived s k i l l (group 1); high heart rate, high perceived s k i l l (group 2); low heart rate, low perceived s k i l l (group 3) and low heart rate, high perceived s k i l l (group 4). They contained 16, 16, 15 and 16 subjects respectively. Provided in Appendix 2 are the following data for each subject in each group; age, sex, instrument, years of experience, grade level, years in a college or university program, and screening questionnaire totals. Age ranges and means for groups one to four respectively were : 15 to 53, mean =25.4; 15 to 24, mean = 22.8; 18 to 47, mean = 25.5; and 15 to 33, mean = 21.4. Distributions for each group with regard to sex were as follows: 12 females and 4 males; 13 females and 3 males; 14 females and 1 male; and 12 females and 4 males. Years-of-experience ranges and means were: 5 to 30, mean = 12.8; 3 to 18, mean = 9.3; 5 to 40, mean = 13.3; and 7 to 23, mean = 12.9, respectively. The following methodological procedures were f i r s t tested in a pilot study involving five subjects and were then altered or refined accordingly. 62 Experimental Setting Assessments and training sessions were conducted in a studio located in the music department of the University of British Columbia. During assessments, subjects performed for an audience of five people in the studio that also contained an audio-recording system. A polygraph was situated in an adjacent room where the various physiological and self-report measures were recorded. Assessments Assessment of the effectiveness of the anxiety-reduction program was conducted at three occasions; one week prior to the program (pre-assessment), one week following completion of the program (post-assessment), and three months following completion of the program (followup-assessment). Subjects performed a single musical piece (the same piece at each assessment) in a studio before an audience of five people. These performances will be referred to as assessment-performances. Subjects were informed that their performances were to be audiotaped for the purposes of s k i l l ratings to be made by musical experts. The audience was comprised of non-music students who were instructed to maintain a "neutral to friendly" facial expression across performers and to applaud at the completion of each performance. One member from the audience controlled the audio-taping equipment. The same five audience members listened to eight to 10 performances at each assessment occasion. Subjects did not perform for the same audience on more than one occasion. 63 Training Program The training program is outlined in the Training Program Manual which is given in Appendix 3. It was based largely on a combination of Kendrick's (1979) behavioural rehearsal and cognitive behavioural rehearsal manuals for musical-performance anxiety reduction. However, the program used in the current study differed from Kendrick's in two ways. F i r s t , progressive muscle relaxation (Paul, 1966) was taught in the f i r s t half of the training program. Second, in accordance with Wine's (1971) suggestion, the attentional training component of the second half of the program placed greater emphasis upon task focusing as opposed to positive thinking than Kendrick's program. The program consisted of four weekly sessions of one and one-half to two hours duration. It followed a stress inoculation model of s k i l l rationale, s k i l l training and practise applying s k i l l s to the criterion situation of public performing. Sessions were conducted in groups of four to seven subjects. Each training program group included subjects from every classification group. The sessions were led by the experimenter. Undergraduate research assistants collected the various self-report and pulse-rate measures during those sessions entailing behavioural rehearsal (i.e., the second, third and fourth sessions). Performances during those training program sessions wi l l be referred to as session-performances. An outline of the research assistants' procedures is provided in Appendix 4. 64 The two s k i l l s taught were progressive muscle relaxation and attention focusing. Attention focusing consisted of training in task focusing and in reinforcing self-instructions. Subjects were asked to practise the s k i l l s between training program sessions and to combine those s k i l l s in performances both within and between training program sessions. They were given handouts summarising the major procedures for progressive muscle relaxation and task focusing/reinforcing self-instruction. These handouts are provided in Appendices 5(a), 6(a) and 7(a). Each handout was accompanied by a between-session s k i l l s practise sheet to encourage subjects to practise the s k i l l s between sessions. They are presented in Appendices 5(b), 6(b) and 7(b) respectively. Self-Report Measures Eight self-report instruments were administered at each assessment (pre-assessment, post-assessment and followup-assessment). The State scale of the State-Trait Anxiety Inventory (Speilberger, Gorsuch & Lushene, 1970) was administered before subjects entered the studio to perform, and is shown in Appendix 8. This scale focuses upon qualities of tension, nervousness, worry and apprehension eli c i t e d by specific situations. It is considered to be a reliable and a valid measure of response to treatment intervention (Borkovec, Weerts Se Bernstein, 1 9 7 7 ) . Subjects also rated their perceived s k i l l in reference to the current assessment-performance before they entered the studio. Appendix 9 provides the Perceived S k i l l questionnaire. Two Subjective Units of Discomfort (SUD) 65 scales were administered to guage the degree of discomfort subjects felt immediately prior'to performing and during the performance. The former was the final questionnaire that subjects completed before entering the studio to perform. The latter was reported retrospectively following the performance. The pre-performance SUD and during-performance SUD scales are presented in Appendices 10(a) and 10(b) respectively. A Thought Questionnaire was developed for the current study. It represented an elaborated version of the thought questionnaire used by Sartory et a l . (1982) with additions relevant to musical performance. Two versions of this questionnaire were administered; one prior to the performance and the other following the performance. The pre-performance Thought questionnaire and post-performance Thought questionnaire are presented in Appendices 11(a) and 11(b) respectively. This questionnaire was included to provide data relevant to the hypothesis that highly aroused individuals experience thought resensitization which contributes to return of fear. Following completion of the performance, subjects also completed a General Information Questionnaire which is shown in Appendix 12. This questionnaire was designed to assess performing experience and to assess discomfort fel t upon returning to performing following an interval without performing. The latter was intended as a retrospective assessment of the relationship between practise and return of fear. A Practise Information questionnaire - that is presented in Appendix 13(a) - also was administered at each assessment 66 occasion. The questionnaire was slightly modified at post-assessment and at followup-assessment, as shown in Appendices 13(b) and 13(c) respectively. During session-performances (at the second, third and fourth training program sessions) the following self-report instruments were administered: Perceived S k i l l , SUDS (pre-performance and during-performance), Thought questionnaire (pre-performance and post-performance) and Practise Information. Only the latter varied from the questionnaires administered at assessment-performances. The slightly modified Practise Information questionnaire is presented in Appendix 14. In addition, at the third and fourth training sessions and at post-assessment, subjects completed a Memory questionnaire in reference to their performance at the previous session. The memory questionnaire is shown in Appendix 15. It was designed to assess the extent to which subjects were aware of and recalled the stimulus setting, contextual cues, and the nature of their responses associated with their previous performance. Data from this questionnaire are relevant to the hypothesis that high arousal interferes with the memory retrieval process and therefore produces dishabituation. At the end of the f i r s t and second months of the three-month followup period, subjects were contacted by phone to answer a Practise Information questionnaire. It is shown in Appendix 16. This questionnaire was designed to additionally assess the generalisability of the s k i l l s acquired from the training program. It asked to what extent those s k i l l s were 67 beneficial in performances over the preceding month. Behavioural Measures One behavioural measure was recorded for each assessment-performance. This was a measure of performance quality. Each subject's s k i l l was rated on the following dimensions: touch (replaced with voice production for vocalists), phrasing, pitch and omission, rhythm, tempo, dynamics, memory and overall quality. A definition of each dimension is presented in Appendix 17. This measure is based on the performance error count procedure used by Kendrick (1979) and is identical to that used by Craske and Craig (1984). Performances were scored by two music experts (for each instrument) who worked independently using copies of the music scores and audiotaped recordings of the performances. They were blind with respect to the subjects' classification group membership and to the assessment occasion at which the performance was given. Subjects' pre-assessment, post-assessment and followup-assessment performances were presented in a counterbalanced order. Each subject's three performances were rated consecutively to allow direct comparisons of performance quality. The raters listened to each performance at least two times: f i r s t to rate each dimension and second to check those ratings. The rating sheet is presented in Appendix 18. 68 Autonomic Measures Anticipatory heart rate was recorded over a 5-minute interval prior to a subject's performing at each assessment-performance. Tachometre output from a polygraph was averaged over the f i r s t 10 beats of every 30-second interval, and the obtained 10 averages were averaged to provide a mean score for the total 5-minute period. Averages were not computer generated and therefore a r e l i a b i l i t y check was conducted. Averages were calculated by an independent rater from ten percent (randomly chosen) of the 30-second intervals. During training sessions, pulse rate was recorded before subjects performed (as a measure of anticipatory arousal) and at the end of each session. The end-of-session pulse rate was intended as a measure of tonic heart rate, although the recorded rates may have been elevated from resting levels to some extent by being in a group. Pulse rate was sampled every 10 seconds over a 2-minute period and averaged. Heart rate represented the single physiological channel to be recorded, as i t is considered to be a reliable and sufficient autonomic measure of anxiety (Gauthier & Marshall, 1977; Hodgson & Rachman, 1974; Hugdahl, 1981) and is the index that was used in earlier return-of-fear studies. Apparatus and Equipment A Singer Caramate SP II projector was used in training program sessions to produce and display audiovisual sequences that demonstrated attention focusing techniques. Assessment performances were recorded with a Sony Stereo cassette recorder 69 (model number TC 158SD). A Shure Mixer was used to maximise sound resolution. Performances were recorded on Maxell UD 90 cassette tapes (normal bias). A Grass Model 7 polygraph was used at each assessment to record EKG signals. Med Associates nonpolarising silver-silver chloride electrodes (half inch sensor diameter) were placed on the right forearm and left leg (below the knee) to record heart rate. This electrode placement represented Lead 2 which is believed to provide the best recording of heart rate (Hassett, 1978). A ground electrode was placed on the l e f t forearm. The skin was f i r s t abraded with Hewlett Packard Redux paste. Beckman electrode paste was used to conduct heart rate signals. Double sided adhesive electrode collars (92mm squared) attached the electrodes to the skin. A Sanyo (model HRM 97E) portable pulse metre recorded pulse rate during training program sessions. This was a photoelectric pulse sensor that attached to the earlobe to provide a digital display of pulse rate and time (in seconds) with an accuracy rating of plus/minus 3% plus/minus 1 count. Procedure Selection of Subjects. Classes of piano and v i o l i n students in the music department of the University of British Columbia were given a brief description of the training program, which is outlined in Appendix 19. They also completed the screening questionnaire. Those who met screening questionnaire cr i t e r i a for selection were contacted by phone and informed of the 70 requirements of participation. Arrangements for preassessment were made for those wishing to participate. Letters were also sent to music teachers around Vancouver with an accompanying letter for students describing the program and requesting that interested students contact the expermenter. The letters are presented in Appendices 20(a) and 20(b) respectively. These subjects completed the screening questionnaire at preassessment. Subjects were instructed to choose a piece of music from their own grade level of performance which they could play from memory and which was three to five minutes in length. Playing from memory served not only to control to some extent the effects of degree of preparation (and to remove anxiety due solely to lack of preparation) but also to increase the demand level of the performance. In order to control for changing patterns of practise of the piece over the duration of the study, subjects were asked to choose a piece that they did not intend to perform in any other situations over the subsequent four months. Subjects were also instructed to provide a copy of their music with any deliberate changes marked on the score for the purpose of the performance quality ratings. They were given a brief description of the assessment procedure (as outlined in Appendix 21). Dates and times for the assessment were arranged. Subjects were reminded of the musical piece requirements and the dates, times and location of their performances by either phone or mail. 71 Pre-assessment. Performances were scheduled every 10 minutes over a 2-day period and were conducted one week prior to commencement of the training program. The assessment procedure is summarised in the Assessment Procedure Outline designed for the research assistants that is shown in Appendix 22. Subjects were instructed to arrive 20 minutes prior to their performance. Following collection of their music score, they were read a set of instructions describing the procedures for the assessment: Before you perform we will be taking a measurement of your heart rate by attaching these three electrodes to your arms and leg. There is no danger nor discomfort involved in this procedure. You are just required to si t quietly for a few minutes. Then, while we remove the electrodes, you can begin f i l l i n g in some questionnaires. Then we will take you into the next room where the audience are seated and you can perform your piece. Remember that your performance is being audiotaped so that i t can be later rated by musical experts. Are there any questions? If subjects inquired about the audience they were told that the "audience is a group of people interested in listening to music - like any regular audience." Subjects read and signed the consent form while the electrodes were attached. The consent form in presented in Appendix 23. Subjects were reminded to "sit quietly, and try not to move" while heart rate was recorded. The electrodes were removed as subjects began completing the self-report instruments in the following order: State Anxiety scale, Perceived S k i l l , Pre-performance Thought questionnaire, and pre-performance SUDS. They were told that " f i r s t impressions are the most appropriate responses to give to these questionnaires". 72 When the questionnaires were completed, subjects were reminded to announce their names and the t i t l e s of their pieces for the audience and were taken into the adjacent studio where a member of the audience began the audio recording. If subjects stopped during their performance they were encouraged to continue by an audience member only once. Subjects returned to the adjacent room when they finished performing and completed more self-report questionnaires in the following order: during-performance SUDS, post-performance Thought questionnaire, General Information questionnaire and Practise Information questionnaire. Subjects were asked to indicate their preferred times for the training program sessions and told they would be contacted by phone to confirm those arrangements. They were also given instructions to refrain from practising their performance pieces more than 15 minutes per day before the post-assessment. These instructions are shown in Appendix 24. Training Program. This is described in detail in Appendix 3. Session 1 began with introductions of the experimenter and group members and a discussion of the nature of performance anxiety and the way i t was manifested in individual group members. This was followed by a brief discussion of the ways performance anxiety can develop. Following an overall rationale and description of the s k i l l s to be taught during the program, 73 subjects were trained in progressive muscle relaxation. They were given a handout summarising the relaxation procedures and asked to practise the exercises at least once per day over the next week, recording their anxiety levels before and after each set of exercises. In session 2, completion of the Practise Information questionnaire was followed by further training in progressive muscle relaxation and observation of group members to ensure that the exercises were performed correctly. Subjects were then instucted in the brief method of relaxation to be applied to performance situations. An imaginal performance scene was used for practise of the "quick" relaxation method before subjects applied that s k i l l to performing before the group (at which time the desensitizing effect of repeated public performances per se was described). The session-performances consisted of performing a short piece (two to five minutes), usually from memory, but without the demand imposed by s k i l l ratings. Subjects' pulse rates were recorded, and the Perceived S k i l l questionnaire, pre-performance Thought questionnaire and pre-perf ormance SUD scales were completed (in that order) before the subjects performed. The during-performance SUD scale and post-performance Thought questionnaire were completed after subjects performed. During each subject's performance the experimenter recorded information necessary to score the Memory questionnaire. The memory questionnaire information sheet is shown in Appendix 25. Group members were given a handout outlining the brief relaxation procedure and were asked to 74 rehearse that method whenever they p r a c t i s e d or performed over the next week. I t was a l s o suggested that they arrange a performance f o r two to three f a m i l y members or f r i e n d s to pro v i d e an o p p o r t u n i t y to p r a c t i s e the r e l a x a t i o n technique before performing at the next s e s s i o n . The s e s s i o n ended with another measurement of pulse r a t e being taken. Task f o c u s i n g and r e i n f o r c i n g s e l f - i n s t r u c t i o n was int r o d u c e d i n s e s s i o n 3 ( f o l l o w i n g completion of the P r a c t i s e Information and Memory q u e s t i o n n a i r e s ) . The d i s t r a c t i n g and an x i e t y e l e v a t i n g e f f e c t s of task i r r e l e v a n t thoughts and t h e i r d e l e t e r i o u s e f f e c t s upon performance were d e s c r i b e d . A u d i o v i s u a l sequences were used to demonstrate f u r t h e r the nature of task i r r e l e v a n t and task r e l e v a n t thoughts. Subjects were asked to d e s c r i b e t h e i r major d i s t r a c t i n g thoughts and to r e p l a c e those with e i t h e r c omforting and rewarding s e l f statements ( f o r use before and a f t e r performing) or task focused and technique o r i e n t e d s e l f statements ( f o r use mainly d u r i n g p e r f o r m i n g ) . S u b j e c t s performed using the same procedure as s e s s i o n 2, and were then questionned about t h e i r thoughts. They were given a handout summarising the major p o i n t s of a t t e n t i o n f o c u s i n g and were again encouraged to arrange a performance f o r a s l i g h t l y l a r g e r group of people i n order to p r a c t i s e r e l a x i n g and t o r e c o r d task i r r e l e v a n t and task r e l e v a n t thoughts. Non-performance r e l a t e d pulse r a t e was again measured. The f i n a l s e s s i o n c o n s i s t e d of f u r t h e r a u d i o v i s u a l sequences demonstrating s e l f - i n s t r u c t i o n , an imaginal sequence f o r f u r t h e r p r a c t i s e of task r e l e v a n t t h i n k i n g , and performances 75 in which subjects were instructed to practise relaxation and self-instructional s k i l l s . A l l measures taken in session 3 were repeated in session 4. Post-assessment arrangements were also made. Post-assessment and Followup. The post-assessment was almost identical to the pre-assessment: the General Information questionnaire was excluded and the Memory questionnaire, was included in the set of pre-performance questionnaires. Subjects performed in the post-assessment one week after they had completed the training program. They were given instructions to refrain from practising their pieces more than 15 minutes per day over the next three months. They were also told they would be contacted by phone towards the end of the f i r s t and second months to answer several questions about practising (the Followup Practise Information questionnaire). The followup-assessment was conducted three months following completion of the training program. It was identical to the post-assessment, excluding the Memory questionnaire. 76 Debrief ing. Subjects were mailed a debriefing letter describing the major objectives and findings of the study and reminding them of the availability of s k i l l ratings for each of their performances. The letter is presented in Appendix 26. 77 RESULTS INTRODUCTION Pre-assessment, post-assessment and followup-assessment occasions were labelled A, E and F. Assessments conducted throughout the training program were labelled B, C and D. A through F constituted the six levels of the repeated measures factor (time). The two classification variables - heart rate (high versus low) and efficacy (high versus low) - were combined to yield four classification groups: high heart rate, low efficacy (group 1 ) ; high heart rate, high efficacy (group 2 ) ; low heart rate, low efficacy (group 3 ) ; and low heart rate, high efficacy (group 4 ) . Heart rate and efficacy constituted the two between-groups factors. The major dependent variables measured at each of the six assessments (A to F) were: heart rate, efficacy, pre-performance SUDS, during-performance SUDS, pre-performance thoughts, and during-performance thoughts. S k i l l and state-anxiety measures were assessed only at A, E and F. Memory variables (accuracy of recall of pre-performance SUDS, accuracy of recall for during-performance SUDS, accuracy of recall for context, and recall for physical arousal) were assessed only at B, C and D. Practise, post-performance confidence and satisfaction were additional variables assessed at each assessment occasion (A to F). The data were analysed in four major sections. The f i r s t section entailed preliminary analyses which were performed to assess the suitability of the data for subsequent analyses. In the second section, the major analyses were performed to compare 78 the classification groups. The post-hoc analyses that constituted the third section compared subjects who displayed return of fear with subjects who did not. Finally, several additional analyses of interest were performed in the fourth section. Each section w i l l be briefly described. Several preliminary analyses were conducted. These included comparisons among instrumental groups (piano, violin and voice), and tests for i n i t i a l differences among classification groups on subject variables: screening questionnaire scores, years of experience, age, number of years in performance programs, sex and instrument. Inter-rater r e l i a b i l i t i e s and correlations among variables at pre-assessment are then presented. Reported next are i n i t i a l differences among classification groups on dependent variables (pre-performance SUDS, pre-performance thoughts and s k i l l ) . Missing data estimation is also described. The major sets of analyses were conducted using three-way (2x2x3 in the analysis of s k i l l , and 2x2x6 in the analysis of pre-performance SUDS and pre-performance thoughts) fixed factors univariate analyses of variance. A univariate approach was chosen in preference to a multivariate approach because of the relatively small sizes of classification groups, and hence the diminished power of multivariate procedures (Winer, 1971). In addition, i t was considered that type 1 error would not be; markedly increased by performing separate univariate analyses as intercorrelations among the dependent variables were generally low. Pre-assessment correlations were as follows: pre-79 performance SUDS and s k i l l , r=0.03; pre-performance SUDS and pre-performance thoughts, r=0.39; and pre-performance thoughts and s k i l l , r=-0.1u. Univariate analyses of variance over repeated measures involve the assumption of equality of the covariance matrices across the repeated factor (time): covariance matrices for each group were pooled and examined using Bartlett's test of sphericity to test for (a) equality of within-time variances (i.e . , equal variability of scores within each time condition) and (b) homogeneity of covariances between pairs of time conditions. Homogeneity of covariances can be assessed by either testing for equality of covariances or by testing for zero covariances (following transformation). Bartlett's test of sphericity compares the pooled covariance matrix to a spherical matrix (a matrix consisting of equal variances and zero covariances) and thereby tests the circularity assumption. That assumption is both a necessary and sufficient condition for the validity of the univariate repeated measures approach. Equality of covariances, as opposed to zero covariances, across pairs of comparisons is a sufficient but not a necessary condition (Kirk, 1982). Bartlett's test was performed for each repeated measures analysis, and B and p_ values are reported only when the test was significant. A Greenhouse-Geisser correction was applied to the degrees of freedom when the test of sphericity was significant. Greenhouse-Geisser procedures correct degrees of freedom by a magnitude proportional to the extent to which the pooled covariance matrix deviates from a pattern of equal variances and 80 covariances, and hence compensate for the resultant positive bias in the F value (Keppel, 1973). Error degrees of freedom were reduced also by the number of missing data estimates in a l l analyses. Summary statistics (sums of squares, F values, degreees of freedom and p_ values) are reported in respective Appendices. Significant three-way interactions were followed by analysis of simple interactions and simple, simple main effects analyses. Significant two-way interactions involving the repeated factor, and significant main effects for the repeated factor, were followed by planned comparisons. Planned orthogonal contrasts employed in the comparisons among the six levels of the repeated factor were as follows: pre-assessment versus the mean of the five remaining assessment occasions (A-BCDEF); the mean of the training program assessments versus the mean of post-assessment and followup-assessment (BCD-EF); post-assessment versus followup-assessment (E-F), which was the contrast of main interest; and two residual contrasts testing linear and quadratic trends across the three training program assessments (BCD). Covariance analyses were also performed on the pre-performance SUDS and pre-performance thoughts measures to remove the effects of i n i t i a l group differences with regard to pre-performance SUDS. Differences among the classification groups on memory variables at assessment B were analysed by the use of multivariate procedures. Significant multivariate effects were followed by univariate analyses. The final analysis in this section of the results involved an examination 81 of differences between high versus low tonic-heart-rate subjects. Post-hoc analyses f i r s t compared subjects who displayed followup return of fear to subjects who did not. Group differences on the measures of heart rate, efficacy and pre-performance thoughts were analysed through the use of multivariate procedures. Significant multivariate effects were followed by univariate analyses and planned comparisons as outlined above. Differences between the return-of-fear and no-return-of-fear groups with regard to post-performance minus pre-performance thoughts, and to post-performance confidence minus efficacy were also compared. Univariate analyses were also conducted to assess group differences with regard to s k i l l and tonic heart rate. Second, post-hoc analyses compared subjects who displayed between- session return of fear with subjects who did not. Post-performance minus pre-performance thoughts, and memory variables (accuracy of recall for pre-performance SUDS and accuracy of recall for during-performance SUDS) were compared between the groups for each assessment interval by the use of multivariate procedures. Significant multivariate effects were followed by univariate analyses. Group differences on heart rate and efficacy were similarly analysed. Several additional analyses are fin a l l y reported. These include correlations among the major dependent variables, minor comparisons between return-of-fear and no-return-of-fear groups, and analyses of the relationships between recall and current fear l e v e l , and between i n i t i a l discordance and return of fear. 82 PRELIMINARY ANALYSES Instrumental Group Differences Differences between singers (n=5) and both v i o l i n i s t s and pianists (n=58) on pre-assessment levels of pre-performance SUDS, during-performance SUDS, heart rate, state-anxiety and s k i l l were tested using independent-samples t tests. Means, standard deviations, t values and probabilities are presented in Appendix 27. Tests of homogeneity of variance were not significant for any variable (see Appendix 27), and therefore pooled variance estimates were used in the calculation of t. No significant differences were found between singers and remaining subjects. Also, singers did not differ from other subjects on either of the following subject variables: screening questionnaire scores and years of experience (see Appendix 27). Therefore, data from vocalists were combined with those of other subjects in subsequent analyses. Classification Group Differences on Subject Variables Subject classification based on median splits of pre-assessment levels of heart rate (median = 87.6 bpm) and efficacy (median in high-heart-rate group = 37.5; median in low-heart-rate group = 47.5) yielded four groups of the following sizes: Group 1 (high heart rate, low efficacy) - n=16; Group 2 (high heart rate, high efficacy) - n=16; Group 3 (low heart rate, low efficacy) - n=15; and Group 4 (low heart rate, high efficacy) -n=16. Differences among the four classification groups on screening questionnaire scores, years of experience, age, and 83 number of years in a college or university performance program were tested by the use of separate one-way analyses of variance. Means, standard deviations, F values and probabilities are presented in Appendix 28. The classification groups differed significantly on only number of years in a performance program [F(3,59)=3.02, p<0.05]. However, Scheffe tests of comparisons between pairs of group means failed to indicate that any two groups significantly differed at the 0.05 significance level. The means for Groups 1 to 4 were as follows: 0.75, 2.13, 2.00 and 1.94. In addition, Chi-square tests did not yield significant differences among the classification groups with regard to either sex or type of instrument. Chi-square values, degrees of freedom and p_ values are presented in Appendix 28. Inter-rater R e l i a b i l i t i e s Inter-rater levels of agreement were assessed by the use of Pearson product moment correlations for both heart rate and s k i l l ratings at each assessment occasion. A l l coefficients surpassed the £=0.80 level (or, approached that level in the case of coefficients for voice ratings) at which ratings may be considered sufficiently reliable (Kazdin, 1980). Correlation coefficients are presented in Table I. Averages of the ratings given by the two s k i l l raters were subsequently used in the analysis of the s k i l l variable. Table I Inter-rater R e l i a b i l i t i e s Variable r df Heart Rate Pre-assessment 0.97 61 <0.001 Post-assessment 0.95 61 <0.001 Followup-assessment 0.96 61 <0.001 Piano s k i l l Pre-assessment 0.92 40 <0.001 Post-assessment 0.90 40 <0.001 Followup-assessment 0.91 40 <0.001 V i o l i n s k i l l Pre-assessment 0.93 14 <0.001 Post-assessment 0.92 14 <0.001 Followup-assessment 0.91 14 <0.001 Voice s k i l l • Pre-assessment 0.76 3 >0.10 Post-assessment 0.75 3 >0.10 Followup-assessment 0.83 3 <0.10 85 Correlations Among Variables at Pre-assessment The high positive correlation between pre-assessment levels of pre-performance SUDS and state-anxiety [r=0.64, p_<0.00l] provided concurrent empirical validation for the pre-performance SUDS variable as a measure of current distress. Since previous return-of-fear researchers used SUDS measures of subjective fear, the SUDS measure was retained and the state- anxiety measure was excluded from subsequent analyses to avoid redundancy. 3 The relationships between number of years in a performance program and the major dependent variables (pre-performance SUDS, s k i l l and pre-performance thoughts) were tested through the use of Pearson product moment correlations. Coefficients and probabilities are presented in Table I I . None of the correlations was significant at the 0.05 significance l e v e l . Therefore, number of years in a performance program was not included as a covariate in subsequent analyses. I n i t i a l Classification Group Differences on Dependent Variables Differences among the classification groups on pre-assessment levels of pre-performance SUDS, pre-performance thoughts and of variance. s k i l l were tested using separate one-way analyses Means, standard deviations, F values and probabilities are presented in Appendix 29. Scores could range from zero to 100 on pre-performance SUDS and from zero to 10 on s k i l l . High scores reflected high levels of anxiety and high 3 The patterns of group means across time were very similar for pre-performance SUDS and state-anxiety variables. 86 Table II Pearson Correlations among Variables at Pre-assessment Subject Variable Dependent Variable r_ p Years in a Pre-performance SUDS -0.20 <0 - 10 performance program Pre-performance 0.17 <0.10 Thoughts S k i l l 0.05 >0.10 87 levels of s k i l l respectively. The pre-performance thoughts measure was based on responses to items 1, 3, 4 and 5 of the pre-performance Thoughts questionnaire; scores could range from four to 11, with high scores reflecting anxious thoughts. The four groups differed significantly on pre-performance SUDS [F(3,59)=11.39, p<0.0l]. Scheffe tests of differences between pairs of group means indicated that Group 1 differed from Groups 2, 3 and 4 at the 0.05 level of significance. Examination of the means suggested that pre-performance SUDS increased from Group 4 to Group 1. The means for Groups 1 to 4 were as follows: 72.2, 59.1, 56.3 and 53.1. The four groups also differed significantly on pre-performance thoughts [F(3,59) = 4.62, p_<0.0l]. Scheffe tests indicated that Group 1 reported more anxious thoughts than Group 4. In addition, the s k i l l variable differentiated the groups at pre-assessment [F(3,59)=3.66, 2 < 0 « 0 5 ] « Scheffe tests indicated that Group 4 demonstrated superior s k i l l to Group 2. MISSING DATA Eight subjects (two from Group 2, two from Group 3, and four from Group 4) missed one training program session each. Missing data for these subjects were estimated by taking the average of (a) the mean of the subject's score on a variable over other assessment occasions, and (b) the subject's group's mean score on the same variable at the assessment occasion from which the subject was absent (Kirk, 1982). Missing data estimates were incorporated in subsequent analyses, and error degrees of freedom were reduced accordingly. 88 RESPONSE TO INTERVENTION OF THE CLASSIFICATION GROUPS The three major dependent variables assessed were pre-perf ormance SUDS, pre-performance thoughts and s k i l l . Means and standard deviations are presented in Table III. Classification variable (heart rate and efficacy) means at each assessment for each group are presented in Appendix 30. Pre-performance SUDS analyses Although the classification groups significantly differed at pre-assessment on pre-performance SUDS, this major analysis was f i r s t conducted without using i n i t i a l pre-performance SUDS as a covariate, because group differences in patterns of change over the six measurement occasions were the major concern. Hence, planned orthogonal contrasts were employed in this and in subsequent analyses. A summary of the sums of squares, F values, degrees of freedom and £ values for the analysis of pre-performance SUDS is presented in Appendix 31. Bartlett1s test of sphericity was not significant. The significant main effect for heart rate [F(1,51)=7.28, p<0.025] indicated that the high-heart-rate group, in general, reported higher pre-performance SUDS than the low-heart-rate group. Similarly, the low-efficacy group, in general, reported lower pre-performance SUDS than the high-efficacy group [F( 1 ,51 )=4.76, p_<0.05]. The main effect for time was also significant [F(5,287) = 1 7.94, p_<0.00l], and the entire sample tended to display a pattern of i n i t i a l l y higher pre-performance SUDS [A-BCDEF, F ( 1 , 57 ) =77 . 79, p_<0.00l]. 89 Table III Classification Group Means for Major Dependent Variables Variable Group Time - Mean (Standard Deviation) A B < C Pre-peformance 1 72 . 1 9 (10 .95) 59.06 (16.04) 51 .88 (14.59) SUDS 2 59 .06 (10 .36) 49.38 (18.34) 44 . 1 3 (13.70) 3 56 .33 (11 .72) 52.53 (13.46) 36 .27 (23.92) 4 53 . 1 3 (5 .44) 42.81 (17.12) 45 .00 (19.92) Pre-performance 1 9 .63 (1 .20) 9.00 (1.75) 8 .44 (1.09) Thoughts 2 9 . 1 3 (1 .02) 7.69 (1.96) 7 .63 (2.00) 3 7 .93 (3 .35) 8.60 (1.40) 7 .80 (1.70) 4 7 .06 (2 .29) 7.00 (2.07) 6 .25 (2.15) S k i l l 1 6 .45 (1 .38) 2 5 .57 (1 .53) 3 5 .99 (1 .01 ) 4 7 .02 (1 .94) D E F Pre-performance 1 49 .69 (19 .53) 49.38 (15.69) 61 .88 (13.28) SUDS 2 40 .25 (10 .28) 38.75 (12.04) 47 .81 (16.63) 3 41 .33 (20 .74) 42.00 (21.86) 41 .33 (24.60) 4 45 .81 (16 .62) 36.88 (20.40) 32 .19 (19.15) Pre-performance 1 8. 00 (1 . 55) 8.00 (1.27) 8. 75 (1 .53) Thoughts 2 7. 44 (1 . 21 ) 7.63 (1.09) 7. 88 (1.67) 3 6. 67 (2. 49) 7.00 (1.73) 7. 67 (1.54) 4 7. 1 3 (1 . 41 ) 6.19 (1.97) 5. 75 (2.46) S k i l l 1 6.62 (1.08) 6. 44 ( 1 .38) 2 6.85 (0.93) 6. 97 (1.14) 3 6.08 (1.26) 6. 46 (1.36) 4 6.99 (1.22) 7. 18 (1.46) Time A = pre-assessment B = second training session C = third training session D = fourth training session E = post-assessment F = followup-assessment Group 1 = 3 = 4 = high heart rate, low efficacy 2 = high heart rate, high efficacy = low heart rate, low efficacy low heart rate, high efficacy 90 Heart rate significantly interacted with time [F(5,287)=3.36, p_<0.025], and the pattern of change from post-assessment to followup (E-F) differed between the high- and low-heart-rate groups [F(1,57)=10.70, p_<0.0l]. Examination of the means showed that the high-heart-rate group displayed an increase in pre-performance SUDS from post-assessment to followup-assessment, while the low-heart-rate group's level of fear tended to decrease slightly over that time interval. This interaction is graphed in Figure 1. The interaction [F(5,287) = 2.04, p_>0.05] was not significant. Efficacy scores across time are depicted in Figure 2. The analysis of pre-performance SUDS was repeated covarying out i n i t i a l differences in pre-performance SUDS to insure that the above effects were not dependent upon i n i t i a l levels. Bartlett's test was not significant. The following planned orthogonal contrasts were tested: the third versus the fourth training program sessions (C-D); post-assessment versus followup-assessment (E-F); the mean of the third and fourth sessions versus the mean of post-assessment and followup-assessment scores (CD-EF); and the second session versus the mean of the remaining assessment occasions (B-CDEF). Only the second contrast (E-F) was of interest. A summary of sums of squares, F values, degrees of freedom and 2 values for the analysis of covariance is presented in Appendix 32. A test of homogeneity of regressions (or, common slopes for the separate regressions of the covariate for heart rate, efficacy and Heart Rate x Efficacy) was not significant [F(1 5,1 51) = 1 .30, p_>0.05] o - © High-Heart-Rate Group O Low-Heart-Rate Group Assessment Occasion Pre-performance SUDS at each assessment occasion for high-heart-rate (n=32) and low-heart-rate (n=31) groups. A=pre-assessment, B=second program session, C=third program session, D=fourth program session, E=post-assessment, F=followup-assessment. 92 A B C D E Assessment Occasion Figure 2; Pre-performance SUDS at each assessment occasion for high-efficacy (n-32) and low-efficacy (n=31) groups. 93 and hence the analysis of covariance was considered valid . Heart rate no longer produced a significant main effect, although the main effect of time was s t i l l significant [F(4,228)=4.45, p_<0.0l]. The three-way interaction again failed to reach significance. .The Heart Rate x Time interaction was significant [F (4 , 228) =3. 90 , p_<0.0l], and the planned contrast of major interest (E-F) was significant [F(1,51)=10.70, p_<0.0l], as it was in the analysis without covariance. The Efficacy x Time interaction was also s t a t i s t i c a l l y significant [F(4,228)=2.43, p_<0.05] but, as before, was not characterised by group differences in post-assessment to followup-assessment change. Means adjusted for i n i t i a l group differences with regard to pre-performance SUDS are presented in Table IV. Finally, as a further examination of the patterns of group differences on pre-performance SUDS, simple effects analyses were performed to cl a r i f y at which assessment occasions both high-heart-rate and low-heart-rate groups, and high-efficacy and low-efficacy groups, differed. Appendix 33 contains sums of squares, F values, degrees of freedom and p_ values. High-heart-rate subjects reported higher pre-performance SUDS than low-heart-rate subjects at both pre-assessment and followup-assessment. Moreover, high-heart-rate subjects did not differ from low-heart-rate subjects on pre-performance SUDS at post-assessment. The same pattern of results emerged when high- and low-efficacy groups were compared. 94 Table IV Pre-performance SUDS Means Adjusted for I n i t i a l Differences Group Assessment B C D E F 1 52.65. 45.46 43.28 42.96 55.46 2 50.01 44.7.6 40.88 39.38 48.44 3 54.63 38.36 43.43 44.10 43.43 4 46.63 48.82 49.63 40.69 36.01 Time Group B = second tra i n i n g session 1 = C = t h i r d t r a i n i n g session 2 = D = fourth t r a i n i n g session 3 = E = post-assessment 4 = F = followup-assessment high heart rate, low efficacy high heart rate, high efficacy low heart rate, low efficacy low heart rate, high e f f i c a c y 95 Pre-performance Thoughts Analyses Sums of squares, F values, degrees of freedom and p values for the analysis of pre-performance thoughts are presented in Appendix 34. Bartlett's test was not significant. Main effects for heart rate [F(1,51)=20.06, p<0.00l], efficacy [F( 1 ,51 ) = 1 1 .15, p_<0.0l], and time [ F (5, 287) =5. 47 , p<0. 001] were a l l significant. Examination of the means (presented in Table III) indicated that the high-heart rate and low-efficacy groups, in general, reported more anxious thoughts than the low-heart-rate and high-efficacy groups respectively. These group effects are depicted in Figures 3 and 4. In addition, the sample as a whole reported more anxious thoughts at pre-assessment in comparison to the average of scores over the remaining assessment occasions [A-BCDEF, F(1,57)=11.65, 2< 0«0 0 1] ' None of the interactions obtained significance. The analysis of the pre-performance thoughts measure was repeated using pre-performance thoughts at pre-assessment as a covariate. A test of homogeneity of regression slopes was not significant [F( 1 5,1 59) = 1 .29, p_ >0.05]. Bartlett's test was not significant. The covariance analysis produced the same pattern of results as the analysis without covariance. The main effects of heart rate [F(1,50)=7.55, 2< 0-0 1] , efficacy [F(1,50)=7.55, 2<0.01] and time [F(4,228)=2.94, 2< 0-0 5^ were a l l significant and no interactions emerged from the analysis. Summary statistics of the covariance analysis are presented in Appendix 35. 96 A B C D E F Assessment Occasion Figure 3: Pre-performance Thoughts at each assessment occasion for high-heart-rate (n=32) and low-heart-rate (n=31) groups. 97 Figure 4: Pre-performance Thoughts at each assessment • occasion for high-efficacy (n=32) and low-efficacy (n=31) groups. 98 S k i l l Analyses Amount of practise was not included as a covariate in the analysis of s k i l l because practise did not significantly correlate with s k i l l at any assessment occasion. Pearson product moment correlations were as follows: pre-assessment -r(6l)=0.11, p_>0.1u; post-assessment - r(6l)=0.l8, p_<0.lO; and followup - r(6l)=0.12, p_>0.1u. In addition, the classification groups were found not to differ in the amount they practised at any assessment occasion. Means and other summary statistics for the analysis of practise are presented in Appendix 36. Bartlett's test was not significant. Appendix 37 contains summary sta t i s t i c s (sums of squares, F values, degrees of freedom and p_ values) for the analysis of s k i l l . Planned contrasts in this analysis entailed the following two comparisons: s k i l l levels at pre-assessment versus post-assessment (A-E); and s k i l l levels at post-assessment versus followup-assessment (E-F). The comparisons were analysed by the use of Dunn's test, and the per comparison significance level was set at 0.025 in order to control type 1 error. The main effect of time [F(2, 1 1 8) =7.29, p_<0.0l] was significant: as seen from Figure 5, the training program resulted in an overall; improvement in s k i l l from pre-assessment to post-assessment [A-E, t (2, 59) =2 .77 , p_<0.05] which remained relatively stable until followup-assessment. That i s , the E-F contrast was not significant [E-F, t(2,59) =0.88, p_>0.05]. The three-way interaction between heart rate, efficacy and time was also significant [F(2, i 18) = 5.21 , p_<0.025], indicating that the 99 @ ffi Group 1 O 0 G r o uP 2 Q U | Group 3 A E F Assessment Occasion Figure 5: Skill levels at each assessment occasion for each classification group (N=63). Group l=high heart rate,' low perceived s k i l l ; Group 2=high heart rate, high perceived s k i l l ; Group 3=low heart rate, low perceived s k i l l ; Group 4=low heart rate, high perceived s k i l l . 100 interaction between heart rate and efficacy was significantly influenced by the time of assessment. Tests of simple interactions showed that heart rate and efficacy interacted only at pre-assessment [F(1,59)=8.64, p<0.0l] and not at post-assessment [F(1,59)=1.29, p_>0.05] or at followup-assessment [F(1,59)<1]. The interaction at preassessment is consistent with the i n i t i a l differences observed among the classification groups with regard to s k i l l . Group means at each assessment are presented in Table III and depicted in Figure 5. However, simple, simple main effects analyses of the significant pre-assessment to post-assessment change (Keppel, 1973) did not yield significant effects of time for any group: Group 1, F(1,118)<1; Group 2, F(1,118)<1; Group 3, F(1,118)<1; and Group 4, F(1,118)<1. Hence, s k i l l did not improve significantly for any single group. In addition, when i n i t i a l differences in s k i l l were covaried - regression slopes were homogeneous [F(6,110)<1] -neither heart rate nor efficacy were associated with post-assessment to followup-assessment change in s k i l l . The only effect to emerge from the analysis of covariance was a main effect of efficacy [F(1,51)=5.85, 2<0.05]: high-efficacy subjects were generally more skilled than low-efficacy subjects at post- and followup-assessment. Summary statistics of the covariance analysis are provided in Appendix 37 also. 101 Effectiveness of the Training Program To complement the pre- to post-assessment comparison examined in the analysis of s k i l l , group comparisons were made at only pre-, post- and followup-assessment occasions with pre-perf ormance SUDS and heart rate as dependent variables. Pre-performance SUDS was analysed by using a 2x2x3 univariate analysis of variance with planned contrasts ( i . e . , Dunn's test) as specified in the description of the analysis of s k i l l . Summary statistics for the analysis of pre-performance SUDS at A, E and F are presented in Appendix 38. Bartlett's test was not significant. The analysis of pre-performance SUDS revealed significant main effects for heart rate [F(1,59)=13.30, p<0.005], efficacy [F(1,59)=8.96, p<0.0l] and time [F(2,118)=37.64, g<(h0l ] . The Heart Rate x Time interaction was also significant [F( 2 ,1 1 8) =4. 57 , p_<0.025]. Examination of the means (presented in Table III) pertaining to the significant planned contrasts suggested that, f i r s t , the entire sample reported significantly reduced pre-performance SUDS from pre-assessment to post-assessment [A-E, t(2,59)=8.26, 2<0.025] . Mean levels were as follows: pre-assessment=60.24; post-assessment=41.75; and followup=45.87. Mean pre-performance SUDS for the entire sample did not increase significantly from post-assessment to followup [E-F, t(2,59) = 1 .84, p_>0.05] but the generality of this effect was limited by the Heart Rate x Time interaction, as reported in previous analyses: the high-heart-rate group reported increased pre-performance SUDS at followup while the low-heart-rate group 102 remained quite stable. A one-way repeated measures analysis of variance was used to analyse heart rate data, testing mean differences across time for the entire sample. Bartlett's test was not significant. Sums of squares, F values, degrees of freedom and p_ values for the analysis of heart rate are contained in Appendix 38. The significant effect of time [F(2,124)=11.34, p_<0'001] w a s characterised by the following trends for the entire sample: lowered heart rates from pre-assessment to post-assessment [A-E, t(2,62)=4.04, 2 < 0 « 0 2 5 ] ' Heart rate did not elevate significantly from post-assessment to followup [t(2,62)=2.08, p_>0.025]. Mean levels were as follows: pre-assessment = 87.35; post-assessment = 82.71; and followup = 85.10. In previous discussion of the analysis of s k i l l i t was indicated that, in general, s k i l l improved from pre- to post-assessment and thereafter remained relatively stable. The analyses in this section indicated that the program also reduced pre-performance SUDS and heart rate, that heart rate levels subsequently slightly, but not significantly, increased in Groups 1 and 4, and that improvements in pre-performance SUDS were not maintained by high-heart-rate subjects. Minor Analyses 103 Memory. Recall in the third training program session for levels of pre-performance SUDS, during-performance SUDS and physical arousal reported at the second session, and for the context that characterised the second session was assessed through the use of a two-way multivariate analysis of variance. Significant effects were followed by univariate analyses. A multivariate approach was chosen as i t was believed that type 1 error may have been substantially increased by a separate univariate analysis on each of the potentially strongly intercorrelated memory variables. Type 1 error was reduced by restricting separate variable analyses to those effects that obtained significance in the case of multivariate procedures. Neither the main effect for efficacy [F (4 , 54) = 1 . 27 , p_>0.l0] nor the interaction effect [F(4,54)<1] was significant. The heart-rate main effect was significant [F(4,54)=2.66, p<0.05]. Statistics from a one-way univariate analyses of variance comparing high- and low-heart-rate groups are reported in Appendix 39. The results indicated that these groups differed significantly with regard to recall for pre-performance SUDS [F(1,58)=4.53, 2 < 0 - 0 2 5 J a n d recall of arousal [F(1,58)=8.59, 2<0.01]. Means and standard deviations are presented in Table V. Scores for the accuracy of recall for pre-performance SUDS and during-performance SUDS represented the difference between remembered and actual levels ( i . e . , a positive value reflected an overestimation of previous levels). Arousal recall scores represented recall of the extent to which subjects felt physically aroused at time B, and could range from zero to 100. Table V Heart Rate Group Means for Memory for Time B 1 04 Var i a b l e Accuracy of pre-performance SUDS r e c a l l Group Mean High heart rate 5.31 Low heart rate -3.94 Standard Deviat ion 14.13 18.76 Accuracy of High heart rate dur ing-performance SUDS r e c a l l Low heart r a t e 4.54 1 .23 19.13 18.90 A r o u s a l r e c a l l High heart rate 50.75 Low heart rate 34.81 21.73 21.17 Accuracy of s t i m u l i r e c a l l High heart rate Low heart rate 3.22 3.29 1 .30 1 .68 105 Accuracy of stimuli recall referred to the number of correct responses to six questions concerning the stimulus context at time B (scores could range from zero to s i x ) . Examination of the means indicated that the high-heart-rate group tended to overestimate while low-heart-rate subjects tended to underestimate their previous pre-performance SUDS level. High-heart-rate subjects also recalled feeling more physically aroused than low-heart-rate subjects. Memory was not analysed over the three occasions on which it was measured (B, C and D) for two reasons. F i r s t , the classification groups had too few subjects to permit a valid multivariate analysis of four dependent variables over three occasions. Second, recall for time B was considered a more precise measure of memory than recall for time C or D, because the latter were likely to be influenced by the subjects' expectation at times C and D that they would subsequently be asked to recall the context and their responses. Tonic Heart Rate. In a two-way analysis of variance of tonic heart rate, neither the main effect of efficacy [F(1,59)=2.31, P>0.10] nor the interaction of Efficacy x Heart Rate [F(1,59)<1] was significant. However, the main effect for heart rate was significant [F(1,59)=8.79, 2<0.005]. High-heart-rate subjects demonstrated significantly higher tonic heart rates (mean=75.94, standard deviation=7.50) than low-heart-rate subjects 106 (mean=70.32, standard deviation= 7.13). Tonic heart rate (measured at times B, C and D) was averaged for each subject and used as a classification variable: high (n=32) and low (n=3l) tonic-heart-rate groups were identified on the basis of a median split (median = 73.5 bpm). A two-way multivariate analysis with repeated measures was used to assess jointly pre-performance SUDS and pre-performance thoughts. Significant effects were followed by univariate analyses. Bartlett's test was significant [B(45)=75.50, £<0.001], hence a Greenhouse-Geisser correction was applied to the degrees of freedom. Summary statistics for the univariate analysis of high- and low-tonic-heart-rate groups are presented in Appendix 40. The only multivariate effect to obtain significance was the main effect of time [F(10,44)=8.41, P<0.001]. Subsequent univariate analyses revealed significant time effects for both pre-performance SUDS [F(4.3,255.7)=16.72, • P_<0.001] and pre-per f ormance thoughts [F(4.3,257.7)=5.28, 2<0.001]. Means and standard deviations are presented in Table VI. Neither the group nor the Group x Time effect was significant in the multivariate analysis. Examination of the means suggested a tendency for pre-performance SUDS to increase from post-assessment to followup in the high-tonic-heart-rate group, whereas stability was observed across that interval in the low-tonic-heart-rate group, group. However, the E-F contrast did not reach significance [F(1,51)=3.31, £<0.10] in the GroupxTime interaction for pre-performance SUDS. 107 Table VI Means for Tonic Heart Rate Groups Variable Group Time - Mean (Standard Deviation) A B C Pre-perfomance SUDS 1 2 62. 58. 27 00 (12. ( 1 1 . 44) 57) 52. 52 49. 1 7 (17.00) (17.23) 44. 44. 1 2 80 (14. (23. 00) 18) Pre-performance Thoughts 1 2 9. 7. 03 80 (1 . (2. 83) 67) 8.27 7.83 (1.79) (2.10) 7. 7. 78 33 (1 . (1 . 96) 88) Pre-performance SUDS 3 4 60. 59. 71 33 (12. (10. 07) 83) 46. 64 45.67 (21.66) (18.50) 39. 45. 00 27 (14. (25. 20) 88) Accuracy of Pre-performance SUDS Recall 1 2 1 .91 -0.50 (16.92) (18.31 ) Accuracy of Dur ing-performance SUDS Recall 1 2 3.52 2.17 (19.48) (18.69) Arousal Recall 1 2 45.61 39.83 (26.65) (19.41) Accuracy of St imuli Recall 1 2 D 3.36 3.13 (1.30) (1.68) E F Pre-performance SUDS 1 2 44 44 .36 .27 (17 (17 .55) .13) 40.46 43. 17 (16.79) (19.59) 48 43 .33 .17 (21 (21 .35) .35) Pre-performance Thoughts 1 2 7 6 .76 .83 (1 (1 .48) .90) 7.52 6.87 (1.15) (2.06) 7 7 .64 .37 (2 (2 .01 ) .25) Pre-performance SUDS 3 4 39 44 .93 .13 (17 (19 .78) .96) 34.00 42.67 (15.98) (17.20) 45 42 .71 .00 (22 (23 .43) .51 ) Group 1 = High Tonic 2 = Low Tonic '3 = Upper Quartile 4 = Lower Quartile 108 A more refined analysis was conducted by comparing subjects in the upper (n=15) versus lower (n=15) quartiles of the tonic-heart-rate distribution with respect to pre-performance SUDS. Subjects with average tonic heart rates of 78 bpm or more and those with average tonic heart rates of 67 bpm or less were compared. Bartlett's test was not significant. Summary statistics for this extreme-groups analysis are reported in Appendix 41. Significant group and Group x Time effects s t i l l failed to emerge. However, the planned orthogonal contrast of major interest involving post-assessment versus followup-assessment was significant [F (1 , 20) =4.49, p_<0.05]. Means are presented in Table VI and depicted in Figure 6. The extreme high-tonic group reported substantially greater pre-performance SUDS at followup-assessment than at post-assessment, while the level of pre-performance SUDS for the extreme low-tonic group remained relatively stable. Finally, a one-way multivariate analysis of variance was used to compare tonic-heart-rate groups with respect to memory for time B. Recall for pre-performance SUDS, during-performance SUDS, physical arousal and the stimulus context were tested. The multivariate group effect was not significant [F(4,56)=<1]. Group means are presented in Table VI. FOLLOWUP RETURN OF FEAR For this set of analyses, subjects displaying an increase of 10 or more points from post-assessment to followup in pre-performance SUDS were classified as the "return-of-fear group" (Group 1, n=27) and those displaying either a less substantial 109 Figure 6: Pre-performance SUDS at each assessment occasion for extreme-high-tonic (n=15) and extreme-low-tonic (n=l5) groups. 1 10 increase or further reduction in pre-performance SUDS as the "no-return-of-fear group" (Group 2, n=36). Furthermore, subjects who did not display a reduction in pre-performance SUDS from pre-assessment to post-assessment equal to or greater than the overall mean reduction (18.49) minus one-half standard deviation (8.97) - i.e., greater than or equal to a reduction of 9.52, rounded to 10 - were excluded from these analyses. Three subjects were excluded from Group 1 (n=24) and 11 from Group 2 (n=25). The pre-performance SUDS means for the return-of-fear and no-return of-fear groups are presented in Table VII. Group differences on heart rate, efficacy, pre-performance thoughts and s k i l l were examined. Error degrees of freedom in both multivariate and univariate F tests were reduced by five, as three of the eight subjects who missed one training session were excluded from this set of analyses because they did not display sufficient fear reduction from pre- to post-assessment. The groups did not differ in the amount they practised overall [F( 1 ,42) = 1 .50, p_>0.l0], nor did they vary in patterns of practise over time [F(5,230)<1]. Heart rate, efficacy and pre-performance thoughts Data related to these variables were analysed through the use of two-way (group x time) multivariate analyses of variance with repeated measures. Significant multivariate effects only were followed by univariate analyses of variance and planned orthogonal contrasts. 111 Table VII Pre-Performance SUDS Means for Return-of-Fear Groups Assessment - Mean (Standard Deviation) A B C Return-of- 65.83 (14.04) 55.21 (19.48) 44.38 (19.35) Fear-Group No-Return-of- 59.40 (10.44) 47.40 (15.55) 42.00 (21.60) Fear-Group D E F Return-of- 45.63 (19.92) 38.54 (18.62) 57.29 (16.81) Fear-Group No-Return-of- 40.28 (14.87) 35.80 (16.37) 30.60 (17.64) Fear-Group 1 1 2 Bartlett's test was significant [B(105)=136.45, 2< 0«0 2 53 « Hence, Greenhouse-Geisser adjustments were made to the degrees of freedom in subsequent analyses. The multivariate group effect was significant [F(3,40)=8.31 , p_<0.00l]. Subsequent univariate analyses indicated that the groups differed significantly on heart rate, efficacy and pre-performance thoughts. A summary of sums of squares, F values, degrees of freedom and p_ values is provided in Appendix 42. Group means are presented in Table VIII. The return-of-fear group, in general, displayed higher heart rate [F(1 ,42) = 1 5.19, p_<0.00l] than the no-return-of fear group. This difference is depicted in Figure 7. Return-of-fear subjects- also reported lower efficacy [F (1 ,42) =8. 91 , p_<0.0l] and more anxious thoughts [F(1,42)=6.00, 2 < 0 « 0 2 5 ] than no-return-of-fear subjects. The multivariate interaction of Group x Time was also significant [F (1 5,28) =2.1 8 , p_<0.05]. Subsequent univariate analyses similarily produced a significant Group x Time interaction for efficacy [F (4. 7 ,21 2 .8) =2 . 96, p_<0'053. T h e same interaction only approached significance in the univariate analysis of pre-performance thoughts [F(4.3,205.9)=2.05, 2<0.10], Presented in Appendix 42 are summary statis t i c s for the above univariate analyses. Efficacy and pre-performance thoughts patterns for each group are depicted in Figures 8 and 9 respectively. The contrast of major interest (between post-assessment and followup) for the Group x Time effect was significant for both pre-performance thoughts [F(1,41)=5.25, 2<0.05] and efficacy [F(1,42)=22.51, 2< u'u u l3 « A s shown in 1 1 3 Table VIII Return-of-Fear and No-Return-of-Fear Group Means Across Assessment Occasions Variable Group Time - Mean (Standard Deviation) A B C Heart Rate 1 9 3 . 6 7 ( 1 2 . 9 7 ) 8 5 . 9 9 ( 1 0 . 5 2 ) 8 3 . 36 ( 1 0 . 5 7 ) 2 8 3 . 5 5 ( 1 0 . 3 5 ) 7 6 . 1 7 ( 9 . 1 8 ) 7 6 . 08 ( 1 0 . 0 5 ) Ef f icacy 1 3 5 . 2 1 ( 2 0 . 6 7 ) 4 9 . 6 7 ( 1 9 . 61 ) 5 3 . 0 0 ( 1 7 . 7 3 ) 2 51 . 0 0 ( 2 3 . 1 4 ) 5 8 . 4 0 ( 1 7 . 2 4 ) 6 4 . 1 6 ( 1 9 . 2 8 ) Pre-performance 1 8 . 8 3 ( 2 . 41 ) 8 . 3 8 ( 2 . 0 4 ) 7 . 71 ( 1 . 8 3 ) Thoughts 2 8 . 4 4 ( 2 . 0 8 ) 7 . 7 6 ( 2 . 0 5 ) 7 . 32 ( 1 . 9 5 ) S k i l l 1 6 . 2 7 (1 . 1 5 ) 2 6 . 5 5 (1 . 5 8 ) D E F Heart Rate 1 8 7 . 9 8 ( 1 0 . 0 9 ) 8 7 . 3 8 ( 1 1 . 6 3 ) 91 . 5 5 ( 11 . 9 0 ) 2 7 8 . 6 3 ( 8 . 5 6 ) 8 0 . 3 4 ( 9 . 6 8 ) 8 0 . 1 4 ( 7 . 4 9 ) Ef f icacy 1 51 . 0 4 ( 1 7 . 0 7 ) 6 6 . 2 5 ( 1 6 . 1 7 ) 5 0 . 8 3 ( 2 0 . 7 3 ) 2 57 . 4 8 ( 1 5 . 9 8 ) 6 8 . 2 0 ( 1 6 . 1 9 ) 7 2 . 52 ( 1 3 . 5 1 ) Pre-performance 1 7 . 7 9 (1 . 7 9 ) 7 . 6 7 (1 . 61 ) 8 . 54 ( 1 . 7 9 ) Thoughts 2 6 . 8 8 (1 . 6 4 ) 6 . 6 8 (1 . 5 2 ) 6 . 48 (1 . 8 1 ) S k i l l 1 6 . 5 1 ( 0 . 9 9 ) 6 . 42 (1 . 1 9 ) 2 7 . 0 8 ( 0 . 9 7 ) 7 . 47 ( 1 . 0 5 ) D i f f 1 D i f f 2 D i f f 3 A-B B-C c-D Ef f icacy 1 - 1 4 . 0 8 ( 2 9 . 31 ) - 4 . 0 4 ( 1 8 . 4 0 ) 7 . 8 3 ( 1 8 . 7 2 ) Di fferences 2 - 1 0 . 4 0 ( 2 0 . 2 6 ) 0 . 8 0 ( 1 6 . 2 0 ) 1 0 . 36 ( 1 7 . 8 4 ) Thought 1 0 . 4 2 ( 2 . 5 5 ) 0 . 7 1 ( 2 . 3 7 ) - o . 4 2 ( 2 . 2 1 ) Differences 2 0 . 8 0 ( 2 . 7 1 ) 0 . 3 2 (1 . 91 ) 0 . 12 ( 2 . 3 7 ) D i f f 4 D i f f 5 D-E Ejl F E f f i c a c y 1 - 8 . 6 3 ( 1 7 . 2 6 ) 10 . 2 1 ( 1 7 . 4 1 ) Di f ferences 2 1 . 8 8 ( 1 4 . 2 7 ) 1 . 0 8 ( 1 4 . 6 2 ) 1 14 Table VIII (eont.) Thought 1 -0.83 (2.37) -1.38 (1.79) Differences 2 0.04 (2.61) 0.04 (1.93) Group 1 = return of fear Group 2 = no return of fear 1 1 5 A B C D E F Assessment Occasion Figure 7: Heart rate at each assessment occasion for return-of-fear (n=24) and no-return-of-fear (n=25) groups. 1 16 Figure 8, efficacy levels decreased in the return-of-fear group and slightly increased in the no-return-of-fear group, from post- to followup-assessment. Also, thoughts became more anxious in the return-of-fear group and less anxious in the no-return-of -fear group over that same interval (see Figure 9). The Group x Time interaction effect was not apparent in heart rate data. Additional analyses were conducted on the basis of difference scores reflecting changes in thoughts (post-performance thoughts minus pre-performance thoughts) and efficacy (post-performance confidence minus efficacy) occurring over the intervals between assessments. Difference scores for the following intervals were tested: post A - pre B (difference 1), post B - pre C (difference 2), post C - pre D (difference 3), post D - pre E (difference 4) and post E - pre F (difference 5). The planned orthogonal contrasts in these difference-score analyses were identical to those conducted in the analysis of covariance on the pre-performance SUDS measure. Appendix 42 provides sums of squares, F values, degrees of freedom and 2 values for the difference-score analyses. Tests of sphericity were not significant for either efficacy differences or thought differences. The Group x Time interaction for efficacy differences was not significant [F(4,183)=1.87, 2> u»l u3 « However, the contrast of major interest indicated significant group differences for the comparison of efficacy change from assessment D to post-assessment with efficacy change over the followup interval (diff4 - diff5) [F(1,45)=17.52, 2< 0- 0 °1] . T he 1 1 7 A B C D E F Assessment Occasion Figure 8: Efficacy at each assessment occasion for return-of-fear (n=24) and no-return-of-fear (n=25) groups. 1 18 A B C D E F Assessment Occasion Figure 9: Pre-performance Thoughts at each assessment occasion for return-of-fear (n=24) and no-return-of-fear (n=25) groups. 1 19 A-B B-C C-D D-E E-F Assessment I n t e r v a l Figure 10: Post-performance confidence minus pre-performance efficacy scores across assessment intervals for return-of-fear (n=24) and no-return-of-fear (n=25) groups. 120 means are presented in Table VIII and are depicted in Figure 10. Efficacy scores tended to remain relatively consistent across both intervals in the group not displaying return of fear, whereas the return-of-fear group reported greater efficacy prior to post-assessment than after assessment D, and lower efficacy prior to followup-assessment than after post-assessment. A similar trend of results emerged from the analysis of thought differences. That i s , the same contrast (diff4-diff5) approached significance [F(1,45) = 3.18, p_<0.l0]. Means are reported in Table VIII and are depicted in Figure 11. The return-of-fear group's thoughts prior to post-assessment tended to be more anxious than their thoughts after assessment D. Also, their thoughts prior to followup tended to be more anxious than their thoughts after post-assessment. In contrast, subjects without return of fear displayed marked stability in their scores on the thought questionnaire across both intervals. S k i l l Bartlett's test applied to s k i l l data was not significant. The group main effect was significant [F ( 1 ,42) =4. 95, p_<0.05], indicating that the return-of-fear group, in general, displayed less s k i l l than the no-return-of-fear group . The Group x Time interaction approached significance [F (2 ,89) =3 . 08 , p_<0.l0], but neither the contrast between pre-assessment and post-assessment nor the one between post-assessment and followup reached significance. Summary statistics for the analysis of the s k i l l measure are reported in Appendix 42. Group means are presented in Table VIII. 121 A-B B-C C-D D-E E-F Assessment Interval Figure 11; Post-performance thoughts minus pre-performance thoughts scores across assessment intervals for return-of-fear (n=24) and no-return-of-fear (n=25) groups. 122 Tonic Heart Rate An independent-samples t test was used to test group differences in average tonic heart rate. Variances of the two groups were homogenous (F=1.30, p>0.10), and therefore a pooled variance estimate was used in the calculation of the t value. The group effect only approached significance [t(47)=1.78,' P_<0.10]. Mean tonic heart rate (and standard deviation) for the return-of-fear group were 75.97 (8.89) and corresponding values for the no-return-of-fear group were 71.72 (7.81). Multiple Regression Multiple regression equations were calculated in order to examine the relative contributions of different dependent variables to the difference in pre-performance SUDS from post-assessment to followup-assessment. In each case, significance of beta values and of the regression equation was examined. In confirmation of the previous set of analyses, a combination of the following variables accounted for a significant proportion of the variance in the difference between pre-performance SUDS E and pre-performance SUDS F [F(3,59)=4.61, p<0.025]: heart rate and efficacy at pre-assessment and the difference between pre-perf ormance thoughts at post-assessment and followup. However, examination of beta values indicated that only heart rate contributed significantly to the variance in pre-performance SUDS, while the value for thought differences only approached significance. Efficacy did not contribute significantly to the variance. Summary statistics are presented in Appendix 43. 1 23 The effects of number of performances and amount of practise over the three-month followup interval, and number of years experience and grade level were also examined. In combination, these variables did not account for a significant proportion of the variance in pre-performance SUDS differences from post-assessment to followup-assessment (see Appendix 43 for a summary of the regression s t a t i s t i c s ) . Also, number of performances only was significantly related to the variance in pre-performance SUDS. BETWEEN-SESSION RETURN OF FEAR The c r i t e r i a for between-session return of fear included both a within-session reduction in fear of 10 or more points (pre-performance SUDS minus during-performance SUDS) and a between-assessment increase of 10 or more points (during-performance SUDS minus pre-performance SUDS). Unfortunately, level of reported fear frequently failed to reduce within a session. Also, satisfaction ratings seemed to confound fear level ratings. Pearson correlations between satisfaction ratings and during-performance SUDS measures at sessions B, C and D were as follows: r(60)=-0.40, p<0.0l; r(58)=-0.45, p<0.0l, and r(57)=-0.20, p_<0.l0. Therefore, these analyses were conducted with restricted sample sizes. The presence of return of fear in terms of the above c r i t e r i a was assessed between assessment occasions B and C, C and D, and D and E. Percentages of the sample displaying return of fear are reported in Table IX. Also presented in Table IX are the percentages of subjects in the followup return-of-fear and no-return-of-fear groups who Table IX Prevalence of Between-Assessment Return of Fear Percentages Percentage of s u b j e c t s with ( f r e q u e n c i e s ) (n=63) (a) No between-assessment 57.1 (36) ret u r n of fear (b) One between-assessment 30.2 (19) re t u r n of fear (c) Two between-assessment 12.7 (8) ret u r n s of fear Percentage of s u b j e c t s with followup r e t u r n of f e a r (n=24) p l u s (a) No between-assessment 50.0 (12) re t u r n of fear (b) One between-assessment 41.7 (10) re t u r n of f e a r (c) Two between-assessment 8.3 (2) ret u r n s of fear Percentage of s u b j e c t s without followup r e t u r n of fear (n=25) p l u s (a) No between-assessment 60.0 (15) r e t u r n of f e a r (b) One between-assessment 24.0 (6) r e t u r n of fear (c) Two between-assessment r e t u r n s of fear 16.0 (4) 1.25 displayed zero, one or two between-assessment returns of fear. One-way multivariate analyses of variance were conducted to compare subjects who demonstrated return of fear over each assessment interval with subjects who did not. Two sets of variables were analysed. Significant effects were followed by univariate analyses. Group differences were assessed f i r s t with respect to thought differences (post-performance thoughts minus pre-performance thoughts) and accuracy of recall for the previous levels of pre-performance SUDS and during-performance SUDS, and accuracy of recall for the previous stimulus context. Appendix 44 contains sums of squares, F values, degrees of freedom and p_ values for these analyses. Error variances of the two groups (return-of-fear, n=14; no-return-of-fear, n=49) at the B-C assessment interval did not significantly d i f f e r . The multivariate group effect was significant [F (4, 55) =5.56, p_<0.005] - error degrees of freedom were adjusted for four missing cases. Univariate analyses showed a significant group effect for the thought difference variable [F ( 1 , 57) = 1 5.67 , p_<0.00l] and for the accuracy of recall of during-perf ormance SUDS variable [F ( 1 , 57) =8.77 , p_<0'n°5]« A summary of the univariate statistics is presented in Appendix 44. Group means are reported in Table X. Subjects who displayed return of fear between training sessions B and C reported more anxious thoughts prior to session C than after session B, while those without return of fear tended to report fewer anxious thoughts prior to session C than after session B. Similarly, return-of-fear subjects tended to overestimate their 1 26 Table X Between-Assessment Return-of-Fear and No-Return-of-Fear Group Means Interval n Variable - Mean (Standard Deviation) Group Thought Pre-Performance During-performance Di fferences SUDS Accuracy SUDS Accuracy B-C 1 14 -1.07 (1.94) 2.14 (13.97) 15.36 (12.63) 2 49 1.23 (1.91) 0.37 (18.48) -0.69 (19.05) C-D 1 9 -0.22 (2.05) 3.33 (11.46) 11.67 (11.18) 2 54 0.07 (2.48) 2.56 (15.56) -0.43 (15.34) D-E 1 12 -1.50 (2.10) -1.43 (14.06) 9.57 (12.93) 2 51 -0.02 (2.52) 0.06 (16.40) -1.55 (15.30) Stimuli Heart Ef f icacy Accuracy Rate 1 1 4 3. 71 (1 .54) 77 .39 (9 .40) 56. 43 (16. 34) 2 49 3. 12 (1.45) 78 .30 (11 .00) 61 . 31 (18. 49) 1 9 3. 1 1 (1.69) 84 .93 (11 .94) 56. 1 1 (15. 57) 2 54 3. 20 (1.48) 81 .21 (9 .99) 54. 39 (19. 03) 1 12 3. 21 (1.97) 81 . 1 1 (9 .42) 63. 33 (14. 36) 2 51 3. 69 (1.46) 83 .09 (11 .24) 68. 43 ( 15. 22) Group 1 = Return-of-fear group Group 2 = Remaining subjects 1 27 previous level of during-performance SUDS, while other subjects even slightly underestimated that value. Subjects who displayed return of fear between assessment occasions C and D did not differ from other subjects according to the multivariate analysis [ F(4,51) = 1.50, 2>0.10] - error degrees of freedom were adjusted for seven missing cases. Nine subjects displayed return of fear between assessments C and D. Means for subjects who displayed return of fear versus means for subjects who did not are presented in Table XI. Error variances of the two groups identified at the D-E assessment interval (return-of-fear, n=12; no-return-of-fear, n=5l) did not significantly d i f f e r . The multivariate group effect was significant [F(4,54)=3.27, 2<0.025] - error degrees of freedom were adjusted for four missing cases. Univariate analyses revealed significant group differences with regard to recall of during-performance SUDS [F(1,57)=6.13, 2< 0'0 2 5^' T h e group effect only approached significance in the case of the thought differences variable [F(1,57)=4 .01, 2<0 «1 u] « A summary of the univariate statistics is provided in Appendix 44. Group means are reported in Table X. Subjects who displayed return of fear tended to report more anxious thoughts prior to post-assessment than after session D and to overestimate their previous level of performance fear. The second set of variables analysed in the same manner were heart rate and efficacy level at the time at which return of fear was evident. However, the multivariate effect was not significant at any assessment interval: B-C interval, F(2,56)<1; 1 28 C-D interval, F(2,53)<1; and D-E interval, F(2,56)<1. Heart rate and efficacy means for subjects who displayed return of fear are provided for each assessment interval in Table X. Finally, point-biserial correlations were calculated between the subjects' satisfaction with their performances and the absence or presence of return of fear over the subsequent interval. The correlations were significant at each interval: B-C,r (60) = 0.43, p<0.0l; C-D, r(58)=0.56, p_<0.0l; and D-E, r(57)=0.44, £<0.01. Classification Groups Percentages and frequencies of between-assessment returns of fear for each classification group are presented in Table XI. These data suggested that return of fear was more evident in Groups 1 and 2 (high heart rate) than in Groups 3 and 4 (low heart rate). A chi-square analysis that compared frequencies of at least one between-assessment return of fear and no return of fear in high- and low-heart-rate groups was significant [x.2 (1 )=4.09, p<0.05]. Eighteen high-heart-rate subjects and nine low-heart-rate subjects demonstrated at least one instance of between-session return of fear. In addition, 13 of the 26 subjects who displayed at least one between-assessment return of fear had tonic heart rates above the sample median, and 13 had tonic heart rates below the median (73.5 bpm). Table XI Between-Assessment Return of Fear in Classification Groups Frequency of Between-Assessment Return of Fear Group ' 0. 1 2 Percentages (frequencies) 1 42. 1 (8) 36.8 (7) 21.1 (4) 2 46.2 (6) 46.2 (6) 7.7 (1 ) 3 81 .8 (9) 18.2 (2) 0.0 (0) 4 65.0 (13) 20.0 (4) 15.0 (3) Group 1 = high heart rate, low efficacy 2 = high heart rate, high efficacy 3 = low heart rate, low efficacy 4 = low heart rate, high efficacy 130 ADDITIONAL ANALYSES Correlations Among Dependent Variables Pearson product moment correlations were calculated between the major dependent variables at each assessment occasion: pre-performance SUDS, pre-performance thoughts, heart rate, s k i l l and efficacy. The coefficients are presented in Appendix 45. Averages of those correlations calculated for the entire sample and for the followup return-of-fear and no-return-of-fear groups separately are presented in Table XII. The significance of the coefficient was tested using students t test with n-2 degrees of freedom. Fisher's Z-transformation test of differences between correlations was used to compare the strength of the correlations for return-of-fear and no-return-of-fear groups. An error rate of 0.40 was used as a Bonferonni significance level for testing differences between correlations: 0.40 was derived from the per comparison alpha level of 0.05 for ten comparisons, used to control for the enhanced probability of type 1 error accruing from multiple comparisons. However, coefficients for return-of-fear and no-return-of-fear groups did not significantly differ for any correlation. Correlations Among Minor Variables The Pearson product moment correlation between pre-assessment pre-performance SUDS and grade level was zero [r(61)=0.003, £ > 0 . 1 0 ] . Pearson correlations between confidence measured at the end of each month of the three-month followup period and number of performances in the preceding month were 131 Table XII Average Correlations Among Major Dependent Variables Pre-performance SUDS Heart Rate Ef f icacy Prethoughts S k i l l Total Sample 0.22 •0.39* 0.46* •0.12 Return-of-Fear Group 0.30 -0.50* 0.59* -0.11 No-Return-of • Fear Group 0.08 -0.36* -0.40* 0.02 Heart Rate Ef f icacy -0.15 Prethoughts 0.24 S k i l l -0.09 •0.08 0.38* -0.05 0.06 •0.03 0.08 Ef f icacy Prethoughts -0.32* S k i l l 0.13 -0.31 0.12 0.45* 0.01 Pre-performance Thoughts S k i l l -0.10 -0.06 -0.12 *r s i g n i f i c a n t at a < 0.05 P.C. 132 a l l significant: f i r s t month, r(60)=0.20, p_<0.06; second month, r(6l)=0.38, 2<.025; and third month (followup-assessment), r(60)=0.34, p_<0.005. Data were missing for one subject from the fi r s t and the third month. Point-biserial correlations between confidence at each month of the followup interval and (a) the application of s k i l l s learned from the training program in the preceding month (yes/no) and (b) benefit received from application of those s k i l l s (yes/no), were significant and positive at each time of assessment. Correlation coefficients are presented in Table XIII. External Validity of Training Program The extent to which subjects were reportedly able to apply and receive benefit from s k i l l s taught in the training program over the followup interval was analysed. Percentages and frequencies are presented in Table XIV for the following categories: no performances (and therefore no opportunity to apply s k i l l s ) ; no = s k i l l s not applied during' performances/skills not helpful when applied; 1/2 = s k i l l s not fully applied/ s k i l l s only partly helpful; and yes = s k i l l s fully applied/ s k i l l s beneficial. In addition, an average (over the three months) of 85.8% of those who had the opportunity to apply the s k i l l s reported that they had done so, and an average of 84.8% of those who applied the s k i l l s reported that they found the s k i l l s to be beneficial. 1 33 Table XIII Correlations between Confidence and S k i l l s Application and Benefit to Performances over the Followup Interval Interval S k i l l s Application S k i l l s Benefit r_ df_ p r df p F i r s t month 0.30 57 <0.05 0.39 53 <0.01 Second month 0.43 57 <0.01 0.44 58 <0.01 Third month 0.31 57 <0.05 0.31 55 <0.05 (followup assessment) 1 34 Table XIV Followup Application of & Benefit from Training Program S k i l l s Percentages (frequencies) F i r s t Month No Perf . No 1/2 Yes Apply S k i l l s 45.2 1 .6 3.2 50.0 (28) ( 1 ) (2) (31 ) S k i l l s Helpful 45.2 1 .6 9.7 43.5 (28) (1 ) (6) (27) Second I Month No Perf . No 1/2 Yes Apply S k i l l s 63.5 0.0 6.3 30.0 (40) (0) (4) (19) S k i l l s Helpful 63.5 0.0 3.2 33.3 (40) (0) (2) (21 ) Third Month No Perf . No 1/2 Yes Apply S k i l l s 41 .3 6.3 3.2. 49.2 (26) (4) (2) (31 ) S k i l l s Helpful 41 .3 7.9 1 .6 49.2 (26) (5) (1 ) (31 ) 1 35 Minor Comparisons between Followup Return-of-Fear Groups Several comparisons were made between followup return-of-fear and no-return-of-fear groups. Previous return of fear. Twelve of the return-of-fear group reported earlier experiences of return of fear (i.e., an interval without performing followed by increased anxiety when performing in public) compared to eight of the no-return-of-fear group. This difference was not s t a t i s t i c a l l y significant _X2(1) = 1.01, p>0.05 . Performing Experience. Eleven of the return-of-fear group reported many (50 or more) public performances prior to the training program, compared to 16 of the no-return-of-fear group. Again, this difference was not s t a t i s t i c a l l y significant XZ(1)=1.01, £>0.05 . Musical Experience. The mean number of years (and standard deviation) of musical experience for the return-of-fear group was 10.79 (7.68) as compared to 13.72 (5.09) for no-return-of-fear subjects. The groups did not significantly differ on this variable [t(39.7) = 1 .57, p_>0«05] - group variances were unequal (F=2.28, 2<0.05) and hence separate variance estimates with 136 adjusted degrees of freedom were used. Level of Competence. The mean grade level for return-of-fear subjects was 9.4 (range = 7 to 11/A.R.C.T.). For no-return-of-fear subjects, the corresponding values were 10.1 (range = 7 to 11/A.R.C.T.). Followup Practise. Average amounts (and standard deviations) of reported practise per month over the three-month followup for the two groups were 68.53% (36.02) and 80.56% (30.18) of normal amounts of practise respectively. The groups did not differ significantly on this variable [t(47)=-1.27, p_>0.05]. Followup Performances. Average number of performances (and standard deviations) per month over the three-month followup interval were 3.96 (4.73) and 7.16 (6.30) for the return-of-fear and no-return-of-fear group respectively. The groups significantly differed on this variable despite the large degree of within-group variability [t(47) =-2.01 , p_<n«053« o f t h e fifteen subjects who reported zero performances over the entire followup interval, 12 displayed followup return of fear. In 1 37 addition, six subjects did not practise or perform as frequently as usual over the followup owing to i l l - h e a l t h . Five of those subjects displayed followup return of fear. Thought Item Analysis To provide a direct comparison with Sartory et a l s . (1982) data, item 3 of the pre-performance thought questionnaire was examined. This item referred to the degree of comfort subjects predicted they would experience. Scores from high- and low-heart-rate groups were compared through the use of chi-square analyses. Frequencies of the different categories of response to item 3 are presented in Table XV. The analysis was significant at pre-assessment [ x_2 (1 ) = 7.78, p<0.0l], post-assessment [_xf (1 ) = 1 0 .53 , p<0.0l] and at followup-assessment [_X_2 (1 ) = 3 .83 , p<0.05]. High-heart-rate subjects expected more discomfort than low-heart-rate subjects at each occasion. Recall and Current Fear Levels The relationship between accuracy of recall of level of fear for a previous session and current level of fear was examined by correlating the accuracy of recall for pre-perf ormance SUDS measure and the measure of expected distress. The correlations were significant at the second [r(57)=0.24, 2<0.05] and third training program sessions [i: (54) =0. 25, p_<0.05]. The correlation was not significant at the fourth session [r(57)=0.16, E>0.05]. A comparison of subjects in the upper (n=l5) and lower (n=12) quartiles of the expected-distress distribution with regard to their recall of pre-performance SUDS 1 38 Table XV Frequencies of Expected Comfort & Discomfort i n Heart Rate Groups Preassessment High heart r a t e group Low heart r a t e group Frequenc i e s (percentages) Not t h i n k i n g about the performance or t h i n k i n g of comfort-able moments Thi n k i n g of m i l d l y or very uncomfort-able moments 2 (6.3) 12 (38.7) 30 (93.7) 19 (61.3) Postassessment High heart r a t e group 5 (15.6) Low heart r a t e group 18 (58.1) Followup Assessment High heart r a t e group 7 (21.9) Low heart ra t e group 15 (48.4) 27 (84.4) 13 (41.9) 25 (78.1) 16 (51.6) 139 for the second session only approached significance [t(25)=1.92, P_<0.10]. Mean accuracy levels (and standard deviations) were 3.33 (11.13) and -8.75 (21.01) for the upper- and lower-quartile groups respectively. The relationship between accuracy of recall of level of fear for a previous session and level of fear at that previous session was examined by correlating accuracy of recall for pre-performance SUDS and during-performance SUDS with actual pre-performance SUDS and during-performance SUDS levels. Accuracy correlated with actual previous fear level at each assessment occasion. Correlations between actual pre-performance SUDS and and accuracy of recall for pre-performance SUDS, and between actual during-performance SUDS and accuracy of recall for during-performance SUDS respectively were as follows: second session, r(57)=-0.26, 2< 0'°25, and r(57)=-0.45, p<0.0l; third session, r(54)=-0.27, p_< 0«0 2 5»- and r(54)=-0.27, 2< 0« 0 2 5 ; and 'fourth session, r(57)=-0.23, p_<0.05, and r (57 ) =-0 .45, p<0.0l. I n i t i a l Discordance Subjects were considered to demonstrate a discordant response-system pattern when they either (a) belonged to both the upper-quartile of the heart rate distribution and the lower-quartile of the pre-performance SUDS distribution, or (b) belonged to both the lower-quartile of the heart rate distribution and the upper-quartile of the pre-performance SUDS distribution.. Concordance was considered to be demonstrated when subjects, belonged to either both the upper-quartiles or both the lower-quartiles of each distribution. Only eight 140 subjects were identified as discordant and 13 as concordant at pre-assessment. Of those, three and nine subjects respectively displayed return of fear at followup. A chi-square analysis did not reveal significant group differences with regard to the frequency of discordance and concordance [ X2( 1 ) = 1 . 9 4 , p>0.l0] SUMMARY OF FINDINGS It was proposed that high heart rate and low efficacy at i n i t i a l assessment would be associated with greater return of fear than would low heart rate and high efficacy. Return of fear was assessed at followup and between weekly performances over the course of the training program. Preliminary analyses indicated that the five vocalists did not significantly differ from other subjects on any relevant variables and their data were therefore included in subsequent analyses. Inter-rater agreements on heart rate and s k i l l ratings were sufficiently high for these variable to be considered reliable. The four classification groups that were identified on the basis of median splits of i n i t i a l heart rate and efficacy differed overall in the number of years they had attended a college or university performance program. However, pairwise comparisons failed to yield significant differences between any two groups. Number of years in a performance program did not significantly correlate with any major dependent variable, hence its use as a covariate in subsequent analyses was not warranted. The classification groups did not differ on any other subject variables: screening questionnaire scores, 141 years of musical experience, age, sex or instrument (i . e . , piano, vio l i n or voice). However, the high-heart-rate, low-efficacy group reported significantly higher subjective fear than a l l other groups at pre-assessment. The groups did not differ i n i t i a l l y on s k i l l or pre-performance thoughts. Finally, state-anxiety correlated sufficiently with pre-performance SUDS to be excluded from further analyses. In the analysis of immediate response to the training program measured at post-assessment, the sample in general demonstrated significantly reduced report of fear, significantly reduced heart rate and significantly improved s k i l l . Each classification group demonstrated significantly reduced pre-performance SUDS, but none of the groups displayed significant s k i l l improvements. In the analysis of long-term effects, i t was found that of those subjects who performed over the three-month followup interval, approximately 85% applied the s k i l l s taught in the training program, and that approximately 85% of those subjects found the s k i l l s to be at least partially helpful. Analysis of changes from post-assessment to followup showed that the sample, in general, demonstrated significantly increased subjective fear, while heart rate, and s k i l l remained stable. In addition, only high-heart-rate subjects, regardless of efficacy level, reported increased fear at followup. The same pattern of results emerged when i n i t i a l differences in pre-performance SUDS were covaried. Heart rate was also significantly associated with recall of previous fear: high-heart-rate subjects 142 overestimated and low-heart-rate subjects slightly underestimated previous fear. Recall of fear was also significantly positively related to current expected distress. Both heart rate and, to an almost significant extent, efficacy were associated with thoughts: low-efficacy and high-heart-rate subjects reported more anxious thoughts than high-efficacy and low-heart-rate subjects. In addition, high-heart-rate subjects expected more discomfort than low-heart-rate subjects at pre-assessment, post-assessment and followup. Tonic heart rate was associated with the subjective report of fear at followup only when extreme groups were compared. Followup return of fear was also examined by comparing subjects who did versus subjects who did not display followup return of fear. The return-of-fear group was found to have higher heart rates and lower s k i l l in general than the no-return-of-fear group. They also reported reduced efficacy and a stronger tendency for thoughts to become more anxious from post-assessment to followup-assessment than the no-return-of-fear group. Return-of-fear subjects also tended to have slightly higher levels of tonic heart rate than other subjects, although this difference was not s t a t i s t i c a l l y significant. Multiple regression analyses indicated the significance of heart rate and number of performances over the followup interval for return of fear at followup. The return-of-fear group reported significantly fewer performances than the no-return-of-fear group. These groups did not differ with regard to previous experience of return of fear, previous performing experience, 143 number of years of musical experience, or amount of practise over the followup interval. The analysis of between-session return of fear was very restricted by the limited number of subjects reporting within-session fear reduction. In two of the three between-session return of fear intervals, subjects who displayed return of fear overestimated their fear level for the previous session and reported more anxious thoughts in comparison to their thoughts at completion of the previous session than other subjects. While between-session return of fear was more evident in the high-heart-rate than in the low-heart-rate classification group, neither heart rate nor efficacy significantly differentiated those experiencing between-session return of fear from those who did not. Nor were these groups differentiated in terms of tonic heart rate. In addition, the presence of between-session return of fear significantly correlated in a positive direction with satisfaction ratings for the previous session's performance. Several additional analyses were also performed. Correlations among the major dependent variables, in general, were strongest amongst measures within the subjective response system (pre-performance SUDS, pre-performance thoughts and efficacy). Heart rate correlated quite strongly overall with pre-performance SUDS and pre-performance thoughts, but other cross-response system correlations were weak (heart rate and efficacy, heart rate and s k i l l , efficacy and s k i l l , and thoughts and s k i l l ) . Correlations tended to be somewhat stronger in the return-of-fear than the no-return-of-fear group, but no 1 44 significant differences emerged. Over the followup period, confidence correlated significantly with number of performances and with application of and benefit from training program s k i l l s . 145 DISCUSSION The results of the current study showed f i r s t that a behavioural rehearsal program including component s k i l l s of progressive muscle relaxation and attention focusing was, in general, quite successful in reducing musical performance anxiety. Second, long-term return of fear at a three-month followup assessment was evident in almost half of the subjects for whom the training program had successfully reduced reported fear. In addition, between-session return of fear (the assessment of which, however, was limited by design factors) was also apparent in some subjects. High heart rate and low efficacy at i n i t i a l assessment were hypothesised to be associated with greater return of fear than low heart rate and high efficacy. Long term return of fear was more evident in subjects displaying high heart rates than in subjects displaying low heart rates. Long-term return of fear was no more evident in subjects who reported low i n i t i a l efficacy than in subjects who reported high i n i t i a l efficacy, although trends in that direction were apparent. A reduction in efficacy from post-assessment to followup, however, was characteristic of subjects who displayed long-term return of fear. The relationship between followup and between-assessment return of fear remains somewhat elusive. The results were consistent with an habituation model of return of fear, as subjects with high levels of autonomic arousal are susceptible to dishabituation. Memory data did not fully support Wagner's (1974) model of consolidation, but rather 146 suggested that the state of arousal may f a c i l i t a t e overestimation of previous fear and the expectation of fear upon re-exposure to the performance situation. The data were consistent with Bower's (1981) model of mood-dependent cognitions. That i s , overestimation of previous fear seemed to be facilitated by current fear states. The failure of i n i t i a l efficacy, or perceived s k i l l , to relate to return of fear contrasts with Bandura's (1977) self-efficacy theory that emphasises the mediating role of efficacy in fear responses. The reduction in efficacy from post-assessment to followup that was characteristic of subjects who displayed return of fear was attributed to lack of relevant performing practise. The effectiveness of the procedures used in the current study is f i r s t described in this chapter. Considered next are long-term and between-session return of fear, the degree to which they provide instances of habituation-consolidation mechanisms, and the significance of fear-recall and fear-expectation. The influence of efficacy, s k i l l and practise is then considered. Design faults and methodological p i t f a l l s that limited the findings will be mentioned throughout these discussions. The chapter is concluded with a summary of theoretical and practical implications and directions for future research in the area of return of fear. 147 TRAINING PROGRAM EFFECTIVENESS The major hypotheses of this investigation were that high heart rate and low perceived s k i l l would be associated with greater return of fear than low heart rate and high perceived s k i l l respectively. The training program must have reduced musical-performance anxiety by a significant amount in subjects who were i n i t i a l l y anxious in order to adequately test these hypotheses. Several methodological steps were taken to ensure that mildly fearful performers were not included in the sample: the program was offered as a service for individuals whose performance anxiety interfered with their musical competence when they were performing, and no other incentive for participation was offered; the Sweeney and Horan (1982) screening questionnaire was used to exclude subjects whose anxiety was not sufficiently high or was due solely to lack of musical preparation; and subjects had to report at least a moderate level of fear at i n i t i a l assessment. I n i t i a l classification group mean ratings on the subjective units of discomfort scale ranged from 54 to 71, indicating substantial fear levels. Moreover, mean scores on the State scale of Spielberger et a l s . (1970) State-Trait Anxiety inventory ranged from 40 to 57, with an overall mean of 49. The latter exceeded the mean of 42.4 obtained with general medical and surgical patients (Spielberger et a l . , 1970), and was comparable to the means obtained with anxious pianists in Kendrick et a l s . (1982) study (51.4), and in Craske and Craigs' (1984) study (48.0). 1 48 Therefore, i t seems reasonable to conclude that subjects were i n i t i a l l y anxious in the public performance analogue setting employed in this study. That i s , the condition of performing a short piece in a studio before an audience of five people and an audiotape was sufficiently demanding to e l i c i t levels of anxiety similar to those experienced by these subjects in real performance settings. It is possible, however, that extremely anxious musicians were discouraged from participating by the task of performing for their colleagues: one subject withdrew for that reason following the f i r s t program session that included performance. Consequently, the sample may have been somewhat less anxious than the subset of most anxious performers in the musical population. However, this was not viewed as a major limitation upon the generalisability of these findings when the substantial anxiety recorded at i n i t i a l assessment is considered. Before evidence for the effectiveness of the training program is discussed, it should be noted that the inclusion of five vocalists did not appear to invalidate the analyses, as the vocalists did not differ from other subjects at i n i t i a l assessment on any relevant variables. The effectiveness of the training program was demonstrated by changes from pre-assessment to post-assessment in each of the three major response-systems of anxiety (cognitive-verbal, behavioural and physiological). The verbal response system was measured by using a subjective units of discomfort scale (SUDS) and the State scale of the State-Trait Anxiety inventory. The 1 49 behavioural response system was measured by obtaining ratings of performance quality. Heart rate served as the measure of the physiological response system. In general, subjects displayed improvements in each response system. Each classification group reported lower levels of SUDS at post-assessment than at pre-assessment; the mean reduction was 18 units. Post-assessment levels of SUDS for each group were below the i n i t i a l screening criterion of 50, and ranged from 37 to 47. Similarly, post-assessment State anxiety levels were substantially reduced from pre-assessment. Post-assessment levels ranged from 36 to 44, with an overall mean of 40. Only 14 of the 63 subjects did not display SUDS reduction of at least one-half standard deviation from the mean (i.e., 9.5 units). The classification groups did not differ in the amounts by which their subjective fear levels were reduced: that i s , neither heart rate nor efficacy was significantly related to the extent to which reported fear was affected by the training program. However, patterns of fear reduction across the training program differed between high- and low-heart-rate groups. Subjects who displayed i n i t i a l l y high heart rate demonstrated a linear decrease in fear from pre-assessment to the fourth training program session. Low-heart-rate subjects reported a slight increase in fear at the fourth program session. The absence of further reduction in fear for either group, between session four and post-assessment, was probably due to the more anxiety-eliciting or more demanding nature of the assessment performances in comparison to the training 150 program performances. Fear reduction seemed to occur quite rapidly for a l l groups except the low heart" rate, low efficacy group (group 3), whose major fear reduction was not apparent until the third training program session. Overall, heart rate reduced significantly from pre-assessment to post-assessment by a mean of approximately six beats per minute. However, there was very l i t t l e evidence of heart rate reduction in low-heart-rate subjects: high-heart-rate groups reduced their heart rate by an average of 8 beats per minute in comparison to a reduction of one beat per minute in low heart rate groups. A floor effect caused by low i n i t i a l values would explain the lack of change in low-heart-rate subjects. Thus, the trend of overall heart rate reduction from pre- to post-assessment was attributable to high-heart-rate subjects. The desynchrony between heart rate and subjective fear observed in low-heart-rate subjects would therefore seem to be a function of response-system sensitivity: the subjective system is considered to be more sensitive than other systems (Agras & Jacob, 1981), and in addition, reductions in physiological arousal were apparently limited by the restricted l a b i l i t y or nonresponsivity of that response in low-heart-rate subjects. The average i n i t i a l heart rate of 97 beats per minute for high-heart-rate subjects attested to the .strong autonomic component of fear for some performers. Thirteeen subjects displayed heart rates of 100 beats per minute or more, which is especially notable considering that recordings were taken in: 151 anticipation of, and not during, performance. Moreover, anticipatory averages for the high-heart-rate groups were comparable to those recorded during exposure from anxious subjects in other studies on fear: e.g., 82 b.p.m. (Gauthier & Marshall, 1977), 88.5 b.p.m. (Grayson et a l . , 1982) and 95 b.p.m. (Grey et a l . , 1979). Craske and Craig (1984) reported an average anticipatory heart rate of 109 beats per minute from anxious pianists seated before an audience. Neftel et a l . (1982) found that heart rate in performers accelerates most in the final 90 seconds before they perform. Accordingly, several subjects in the current study stated that they felt their heart rates rise considerably as they entered the studio where the audience was seated and as they prepared to perform. Performance heart rate was not recorded because polygraph measurement is intrusive, and hence, may not only be detrimental to performance technique, but may also reduce the external validity of the performance assessment. In contrast, reliable telemetric devices and portable EKG monitors with data storage capabilities can provide valid measurement of physiological responses during performances. Low-heart-rate subjects showed l i t t l e autonomic reactivity in anticipation of performances, as their mean i n i t i a l rate of approximately 77 beats per minute lay within the range of average resting heart rate (Hassett, 1978). Their lack of responsivity was shown further by the differences between "tonic" and anticipatory heart rates: high-heart-rate subjects displayed a mean increase from tonic levels of 21.07 beats per minute in comparison to an increase of 6.28 beats per 1 52 minute for low-heart-rate subjects. Despite the reduction in arousal experienced by high-heart-rate subjects, their heart rates remained at least 11 beats per minute, and on average 14 beats per minute, above those of low-heart-rate subjects. The wide range in heart -rate response is ill u s t r a t i v e of large individual differences in the extent to which the three major fear response systems are concordant or discordant, and of their potential for at least partial independence. It is this kind of discordance that resulted in Lang's (1968) reconceptualisation of the anxiety state. The unitary model of fear was replaced by one in which the three major response systems are viewed as loosely coupled components. The evidence for greater discordance in low-heart-rate subjects than in high-heart-rate subjects lends support to Hodgson and Rachmans' (1974) hypothesis that discordance is likely in the absence of strong emotional arousal. On the other hand, i t is also possible 'that heart rate may not have been the most appropriate measure of physiological arousal for a l l subjects. Physiological measurement is complicated by the lack of correspondence between different indices of autonomic arousal. Autonomic indices may be differentially affected by e l i c i t i n g stimuli and by individual differences (Lang, 1971). The latter refers to response specificity, or, strongly preferred individual response channels (Borkovec, Weerts & Bernstein, 1977). Several researchers recommend that autonomic arousal be sampled in various channels, such as cardiac, electrodermal and respiratory (Hassett, 1978). However, the analysis of 153 electrodermal responding in musically anxious performers has not yielded systematic results in the past (Craske & Craig, 1984). The demonstrated usefulness of heart rate measurement in musical performers (Craske & Craig, 1984; Kendrick et a l . , 1982), the restriction to heart rate indices of autonomic arousal in previous studies examining return of fear, and the argument that heart rate is a sufficient and reliable measure of anxiety (Gauthier & Marshall, 1977) led to i t s sole measurement in the present study. Patterns of change in heart, rate across the training program differed between high- and low-heart-rate subjects. High-heart-rate subjects displayed linear reductions until the fourth training program session, at which time their heart rate increased. Several factors may have contributed to this curvilinear pattern. Heart rate elevations at final exposure sessions have been previously reported (Barlow et a l . , 1980; Grey et a l . , 1979). Grey et a l . (1979) attributed the elevation to the subjects' anticipation of treatment completion. It is also possible that subjects chose to perform demanding pieces at that time, as elevated heart rate has been associated with increased demand (Leitenberg et a l . , 1971). Similarly, the more demanding nature of post-assessment in comparison to the training program assessments may account, to some extent, for the further elevations in heart rate experienced by a l l subjects at post-assessment. 1 54 The two low-heart-rate groups displayed different patterns of heart rate. The low heart rate, low efficacy group (group 3) displayed a pattern similar to high-heart-rate subjects. In contrast, low-heart rate-subjects with high efficacy (group 4) demonstrated a pattern of heart rate acceleration from the third training program session, following an i n i t i a l pattern of deceleration. The performance of challenging pieces is particularly likely to account for the pattern of acceleration in this group, as such "courageous" behaviour is considered to be characteristic of individuals with high efficacy (Bandura, 1982; Rachman, 1978). Subjects with low efficacy may not have played very demanding pieces in order to reduce their likelihood of becoming excessively fearful, in the same way that some phobic subjects in Sartory et al s . (1982) study maintained safe distances from feared objects. The absence of a similar pattern of heart rate acceleration in high-efficacy subjects with high heart rate suggests that heightened arousal may deter "courageous" behaviour. However, the influence of demand level can be only speculated, as the demand levels of the pieces performed within training program sessions were not assessed. The opportunity for such variability in demand would have been eliminated i f training program assessments had been conducted with the same performance requirements as the assessment-performances. Subjects performed the same piece at each assessment-performance and the d i f f i c u l t y or demand levels of those pieces were equated across subjects. Subjects could choose different pieces in 1 55 training program performances. Further standardisation of the performances in the training program sessions would have yielded more reliable estimates of habituation processes. Performance quality (i. e . , s k i l l ) was rated only in regard to pre-assessment, post-assessment and followup-assessment performances. Approach behaviour (the measure usually employed in studies of phobic fears) was not appropriate in this setting: the high level of demand of musical performances renders escape very d i f f i c u l t . Musical performances tend to represent "all-or-none" situations for which graded hierarchies of proximity are not easily generated. S k i l l was rated with a sufficient degree of agreement among raters to be considered a reliable measure. Similarly, raters in Craske and Craig's (1984) study achieved an agreement level of 0.86 using the same scales as used in the current study. It is notable that average s k i l l levels were moderate to quite high at i n i t i a l assessment. They ranged from 5 to 7 on a ten point scale on which 10 referred to "excellence". The analyses showed an overall significant improvement in s k i l l from pre- to post-assessment. However, the importance of this improvement is mitigated by two factors. F i r s t , unlike the clearly valuable improvements shown by heart rate and subjective fear data, the improvement in s k i l l was relatively minor and of questionnable external v a l i d i t y . Second, and as a consequence of the f i r s t , separate analyses of changes within each classification group showed that none of the groups experienced s t a t i s t i c a l l y significant s k i l l improvement. 1 56 Kendrick et al.'s (1982) anxiety-reduction program also failed to achieve immediate improvements in s k i l l . However, they recorded significant s k i l l improvement at a five-week followup. Sweeney and Horan's (1982) program, that entailed both relaxation and attentional training, also improved musical a b i l i t y . Both studies used experimenters who were proficient in music and who could, therefore, serve as more credible models than the experimenter in the current study for the demonstration of the combination of anxiety-reduction techniques with musical ab i l i t y s k i l l s . Their musical proficiency may have enhanced their programs' effects upon s k i l l . It has been suggested by other researchers that a skills-training approach is necessary to effect gains in actual performance. Decker and Russell (1981), for example, compared cognitive restructuring and relaxation with study-skills training for test anxiety: only the latter improved academic performance. A similar finding was reported by Allen (1972) and by Lent and Russell (1978) (cited in Decker & Russell, 1981). Fremouw and Zitter (1978) also cited studies in which skills-training for performance anxieties resulted in both self-reported and behavioural improvements, while systematic desensitization influenced only subjective discomfort levels. In contrast, Fremouw and Zitter (1978) found that a combined program of relaxation and cognitive training and a skills-training approach both improved behavioural measures in speech anxiety. However, given the more specialised s k i l l s involved in musical performance than in public speaking, improvement in musical a b i l i t y may be more dependent upon the 157 presence of competent models or skills-training than improvement in public speaking. An additional measure of the behavioural response system that has been used in the examination of performance anxiety is Paul's (1966) Timed Checklist of observable signs of anxiety. However, previous research in the area of musical-performance anxiety has questionned the usefulness of this measure. Kendrick et a l . (1982), for example, found that some of the more subtle behaviours observed during actual performances were not adequately recorded by the use of the Timed Checklist. Also, in contrast to the high inter-rater agreement level of 0.94 on a measure similar to the Timed Checklist reported by Sweeney and Horan (1982), Craske and Craig (1984) and Borkovec, Stone, O'Brien and Kaloupek (1974) both obtained relatively low inter-rater r e l i a b i l i t i e s using the Timed Checklist. Therefore, the global pattern of synchronous changes amongst the anxiety response systems from pre- to post-assessment masked the desynchrony between subjective fear and autonomic arousal characteristic of the low-heart-rate group, and the limited change in the behavioural response measure characteristic of a l l groups. Return of fear, however, has been documented following both synchronous (Grey et a l . , 1979; Grey et a l . , 1981; Sartory et a l . , 1982) and desynchronous (Barlow et a l . , 1980) response system changes. From the above summary, i t can be reasonably concluded that the training program was generally effective in reducing musical-performance anxiety. However, the effectiveness was understandably limited for the 158 autonomic component in subjects who demonstrated minor i n i t i a l autonomic responsivity, and relatively small in terms of actual musical s k i l l . The external validity of the training program was also addressed by questionning the applicability of, and benefit from, s k i l l s acquired in the program in performances other than those included in the study. Approximately 85 percent of the subjects who had an opportunity to apply the s k i l l s of relaxation and attention focusing to performances over the three-month followup interval reportedly did. so, and 85 percent of those reported that the s k i l l s were beneficial. It should be noted, however, that these percentages may be positively biassed by the influence of social desirability responding: information for the f i r s t two months of followup was collected by the experimenter during phone contact. Use of an independent assessor or provision for private responding may have yielded less favourable percentages than those obtained. FOLLOWUP RETURN OF FEAR Classification Variables High heart rate and low perceived s k i l l were predicted to be associated with greater return of fear than low heart rate and high perceived s k i l l . Heart rate was signficantly associated with followup return of fear in the manner predicted. Perceived s k i l l ( i . e . , efficacy) was associated with trends in the predicted direction that were not significant. Low-efficacy subjects tended to experience some elevation in fear level at 159 followup, whereas high-efficacy subjects' fear levels at followup were similar to their post-assessment levels. The two high-heart-rate groups (i.e., high heart rate/low efficacy and high heart rate/high efficacy) experienced significant mean increases in fear from post- to followup-assessment. An interaction between heart rate and perceived s k i l l was not apparent. That i s , the amount of return of fear that occurred in high-heart-rate subjects was not associated with efficacy level. Low-heart-rate groups did not display fear elevations at followup, and the low heart rate/high efficacy group tended to report continued fear reduction. The extent to which fear returned in high-heart-rate subjects was substantial (a mean of 10 to 12 units), but, as found in other return-of-fear studies, pre-assessment levels of fear were not attained. The current data agreed with the description of return of fear as a partial return in fear intensity, which, at the same time, is of sufficient quantity to warrant concern. The investigation of i n i t i a l rather than post-assessment or followup levels of efficacy and heart rate in association with long-term return of fear was based f i r s t on the u t i l i t y of i n i t i a l subject screening for treatment planning. Second, i t was expected, on the basis of previous research (Grey et a l . , 1979), that despite reductions in heart rate and increases in efficacy over the course of the training program, group differences would be maintained on these variables. Subjects who demonstrated high i n i t i a l heart rates in comparison to the total sample were expected to also demonstrate high heart rates 160 at every other assessment in comparison to the total sample. The results were consistent with these expectations. The significance of i n i t i a l heart rate for return of fear in the current study replicates the findings of Grey et a l . (1981) and Barlow et a l . (1980). The'independence of heart rate from return of fear in Sartory et al,'s (1982) study was confounded by individual differences in approach behaviour. Some high-heart-rate subjects in their study- avoided fear by maintaining large distances from the phobic object. Sartory et a l . subsequently predicted a return in the fear levels of their avoidant subjects, given closer approach to the stimulus. A less confounded examination of return of fear under varying arousal conditions was performed in the current study than by Sartory et a l . , because the level of approach was held constant. On average, low-heart-rate subjects were as responsive to the training program in terms of immediate subjective fear reduction as high-heart-rate subjects. Post-assessment levels of pre-performance SUDS we're very similar in high- and low-heart-rate groups, even when i n i t i a l differences in reported fear were covaried. Similarly, Grey et a l . (1979) did not observe differences between high- and low-heart-rate responders with regard to their exposure outcomes. However, close examination of the 14 subjects who did not report at least 10 points reduction in pre-performance SUDS from pre- to post-assessment revealed that ten displayed i n i t i a l heart rates below the sample median. In addition, eight of those 14 belonged to the low-tonic-heart-rate c l a s s i f i c a t i o n . These distributions 161 tend to support the claim that outcome is more successful following strong versus weak physical responsiveness to exposure stimuli (Borkovec & Sides, 1979; Lang, 1970; Stern & Marks, 1973; Wolpe, 1978). The use of median splits for the classification of subjects does not produce distinct groups. Strong group effects may be expected when subjects whose values l i e close to the median are excluded from subsequent analyses. The significance of the effect of heart rate arising from the use of the median split approach to classification serves to emphasise the importance of that variable. Efficacy effects may have been significant in the analysis of extreme groups, but such extreme-group analysis has limited generalisability and restricts the usefulness of the variable as a predictor in the total population of anxious performers. Results from the comparison of the classification groups were very similar to results from post-hoc analyses that compared subjects who displayed long-term return of fear with subjects who did not. The f i r s t criterion for return of fear was a pre- to post-assessment fear reduction of at least one-half standard deviation from the mean, or, 10 units on the SUD scale. The second criterion was a fear increase of at least ten units from post-assessment to followup. The use of the same value for a l l subjects f a i l s to consider the effect of i n i t i a l values: an increase from 20 to 30 may not equate with an increase from 70 to 80. However, group differences were not confounded by the above c r i t e r i a , as levels of pre-performance 1 62 SUDS for the return-of-fear and no-return-of-fear groups did not differ substantially at post-assessment. Twenty-four of the 49 subjects who displayed fear reduction reported return of fear. The prevalence rate of approximately 50 percent of subjects who displayed short-term fear reduction indicates the robustness of the return-of-fear phenomenon. Obviously, statistics of incidence are dependent upon the chosen c r i t e r i a . Fear levels (at the highest intensity) in subjects with heart rates over 100 beats per minute increased from 20 to 70 in Grey et a l s . (1981) study over one week. Their 50 point increase is much greater than the mean increase of approximately 20 points (38 to 58) in the return-of-fear group in the present study. However, Grey et a l . (1981) assessed the occurrence of return of fear following a single exposure session when it was likely to be at its greatest, since the magnitude of dishabituation gradually diminishes with repeated exposures (Groves & Thompson, 1970; Smith et a l . , 1978; Thompson & Spencer, 1966). The four-week "exposure" based training program used in the present study may have reduced the magnitude of dishabituation at a three-month followup in comparison to dishabituation one week following a single exposure. However, lengthy intervals between assessments may either increase or decrease the magnitude of dishabituation, depending on the number and quality of exposures over that period (Rachman, 1978). The importance of both heart rate and efficacy was indicated in the analyses that compared subjects who demonstrated return of fear with subjects who did not. Those 1 63 with return of fear demonstrated, in general, higher heart rates and lower efficacy than those without return of fear. Heart rate remained at a consistently higher level across time in the return-of-fear group than in the no-return-of-fear group. Efficacy levels for the return-of-fear group increased to the extent that they equalled the efficacy level for the no-return-of-fear group at post-assessment. The former group's efficacy level decreased substantially between post-assessment and followup, whereas i t remained stable in the no-return-of-fear group. The changes that subjects' efficacy underwent were revealed further by the analysis of differences between post-performance confidence and (pre-performance) efficacy. In the return-of-fear group, efficacy remained relatively stable from completion of the fin a l training program session to the start of the post-assessment performance, but i t reduced from completion of post-assessment to the start of the followup performance. Overall, these results seem to suggest that high i n i t i a l heart rate is an important characteristic of individuals who show long-term return of fear, as are changes in efficacy following completion of an anxiety-reduction program. The greater u t i l i t y of heart rate than efficacy as a predictor of return of fear is again suggested. The effects of heart rate and efficacy could not be attributed to varying amounts of practise, as neither the classification groups nor the return-of-fear groups differed in the amount they practised in general or in the amount they practised over the followup interval. However, number of 1 64 performances over the followup interval significantly differentiated the return-of-fear groups. The differences between the heart-rate groups were not attributable to i n i t i a l fear levels. Philips (1985) and Grayson et a l . (1982) reported that subjects who displayed return of fear were more fearful at i n i t i a l assessment than subjects who did not. The classification groups that displayed the most return of fear (i.e . , high-heart-rate groups) in the current study also reported higher i n i t i a l fear levels than other groups. However, controlling for i n i t i a l fear levels by the use of covariance analysis did not remove classification-group differences with regard to followup fear level. In addition, return-of-fear and no-return-of-fear groups reported similar mean i n i t i a l fear ratings. Followup elevation in fear levels observed in some subjects therefore seemed to represent a true return-of-fear phenomenon as opposed to general relapse, since subjects who displayed return of fear did not simultaneously demonstrate increased heart rate and decreased s k i l l . Moreover, those displaying return of fear did not appear to differ from those who did not in terms of the rate or degree of fear reduction during the training program: a finding which is also consistent with other studies that examined, return of fear. Return of fear can be viewed as an example of dishabituation that occurred frequently in subjects with high levels of arousal because dishabituation is most likely to occur in highly aroused states. Similarly, Parkinson and Rachman 165 (1980) and Rachman (1981) found that the discomfort associated with unwanted intrusive cognitions is likely to reappear when individuals enter anxious states. Also, Rachman (1979) postulated that the likelihood of return of fear is increased when high levels of arousal are experienced during training. Autonomic Arousal The current study also attempted to examine processes underlying the effects of heightened arousal upon dishabituation, or return of fear. It was suggested that high arousal may disrupt the long-term consolidation of short-term habituation by either impeding the retrieval of a matching stimulus representation, or by promoting thought resensitization. High levels of arousal tend to narrow excessively the focus of attention. It was suggested on the basis of Wagner's (1974) model that excessive focus would impair the association of stimulus representations stored in memory with contextual cues surrounding feared stimuli during exposure. Consequently, the retrieval of a stimulus representation upon re-exposure to a feared situation would be impeded. The absence of a matching stimulus representation that is primed in short-term memory upon re-exposure is believed to result in dishabituation. Unfortunately, recall for the post-assessment performance was not assessed at the followup-assessment, as floor effects were expected to mask group differences with regard to the accuracy of memory for contextual cues. In retrospect, an assessment of subjects' recall for their fear levels at the post-assessment 166 performance would have provided very useful information. Responses to the Memory questionnaire that was administered throughout the training program were used to examine differences in recall between the classification groups. Responses to this questionnaire were assumed to reflect a persisting tendency that would have been also apparent at followup. The stability of responses to the questionnaire was evident, at least across the three occasions on which i t was administered. Similarly, Eich, Reeves, Jaeger and Graff-Radford (1985) found that memory for pain was unrelated to treatment changes in pain levels, and occurred to the same extent in the second as compared to the f i r s t half of a pain management program. It was also assumed that anticipatory heart rate related closely to heart rate during actual performance, and that subjects with, high anticipatory heart rate would also have high performance heart rate. In support of this assumption, Neftel et a l . (1982), who monitored heart rate continuously during musical performance, found that heart rate elevations noted prior to performing were maintained throughout the entire performance. It was hypothesised, in accordance with information-processing models, that aroused subjects would have relatively poor recall of the context of their performance due to the disruptive influence of arousal upon attention. In contrast to predictions, high- and low-heart-rate groups did not differ in their recall of the stimulus setting. It is unlikely that the Memory questionnaire was not sufficiently sensitive to detect arousal effects because the obtained data covered the f u l l range 167 of possible scores (0 to 6), and there was no evidence to suggest the presence of ceiling effects. The obtained means (2.7 to 3.6) seem low for a l l groups in consideration of the relative ease of the questions. The possibility that narrow attentional focus accompanies not only high levels of physiological arousal, but is characteristic of heightened emotional states in general, is consistent with Wine (1971) and Sarasons' (1978) claim that anxious performers are typically inwardly oriented. Wine (1971) described test anxious individuals as internally focused, self-evaluative and strongly aware of their autonomic responses. Similarly, Bower (1981) claimed that attention is directed towards emotionally relevant information in states of "high-emotion", to the exclusion of incidental material. However, the finding that the fear responses of at least some subjects in the current study exemplified maintenance of the effects of short-term habituation suggests, according to Wagner's model, that mediocre recall of context is sufficient for later retrieval. The data pertaining to the recall of contextual cues do not seem to account for differences between more and less aroused subjects with respect to return of fear. High- and low-heart-rate groups did .differ in their recall of fear levels at the second training program session. High-heart-rate subjects tended to overestimate their previous level of fear by approximately 6 units, whereas low-heart-rate subjects even slightly underestimated that value. High-heart-rate subjects also recalled feeling more physically aroused at 168 the second session than low-heart-rate subjects: an expected finding given that heart rate data showed that they were indeed more aroused. Arousal or the memory for level of arousal at a previous exposure therefore seemed to be related to memory for fear. In turn, memory for fear may have mediated the occurrence of followup return of fear. In addition, the significant correlation obtained between accuracy of recall for previous fear and current distress (assessed in terms of the amount of distress expected at the current exposure) suggested that distortion of recall may occur just prior to re-exposure when an anxious state is re-entered. That i s , the processes underlying distortion in recall seem related to: the amount of arousal at the time of previous exposure; memory for the amount of arousal assessed one week later; and the amount of distress and arousal at the time the recall judgement is made. Moreover, in juxtaposition to Wagner's model, memory for response was apparently unrelated to memory for contextual cues. Increased arousal narrows the focus of attention (Easterbrook, 1959; Baddeley, 1972), and in states of high arousal, the arousal cues themselves can become the focus of attention. As previously suggested, the generally low level of accuracy for the recall of previous contexts suggested that the attention of a l l subjects was relatively internally focused. Given that highly aroused subjects have intense internal cues (such as a racing heart and rapid breathing), i t is conceivable that high-heart-rate subjects remembered being more fearful than they actually were because salient arousal cues promote 169 overestimation of prior fear during re-elicitation of arousal. Supportive evidence for the proposed relationship between intensity of response and distortion of recall derives from research in mood-dependent memory (Bower, 1981) and judgement heuristics (Tversky & Kahneman, 1982). The judgement bias labelled by Tversky and Kahneman (1982) as an a v a i l a b i l i t y heuristic occurs when judgements of event frequency are biassed by "the ease with which instances or occurrences can be brought to mind" (p. 11). Events whose instances are more easily retrieved are judged as more numerous than events that actually occurred with equal frequency but whose instances are less retrievable. Salience significantly influences retrievability: salient events are retrieved more easily than non-salient events. Moreover, these biases and heuristics persist despite corrective feedback. Given that arousal cues are salient in highly aroused subjects, and assuming a generally strong relationship between the strength of arousal cues and their perception (Borkovec, 1976), the memory for the state of fear and i t s associated salient arousal cues may predominate when highly aroused subjects recall their prior responses to feared situations. Hence, prior experiences may be remembered as being more intense than they actually were. Similarly, Robinson (1981) found that recall was much greater for intense personal memories than for weak ones, and that the estimate of the intensity of those memories undergoes change. 1 70 In his description of mood-dependent memory, Bower (1981) claims'that distortion of recall for an emotional response is particularly likely to occur when the same emotional state is evoked at r e c a l l . In agreement with Tversky and Kahnemans' (1982) and Robinson's (1981) findings, Bower (1981) found that intense emotional experiences were recalled more easily than weak emotional experiences. In addition, an emotional state is believed to enhance the salience of mood-congruent material; Bower predicted that people actively attend to material that is consistent with their feelings. Moreover, he found that unpleasant events were recalled as more unpleasant than they actually were when depressed at time of r e c a l l . He attributed that distortion to the mood intensifying effect of the repetition of associative networks in memory. Eich et a l . (1985) noted an analogous phenomenon in memory for pain research in which overestimation of previous pain was observed to occur when present pain was rated as high. When present pain was rated as low, previous pain was recalled as being less severe than i t actually was. Therefore, Bower's data regarding memory for mood intensity, and Eich et al.'s data regarding memory for pain intensity seemed to be paralleled by the current data regarding memory for fear intensity. However, Eich et a l . (1985) also reported that recall for pain did not relate to prior pain intensity: a finding that was not consistent with the availa b i l i t y heuristic, nor with the evidence presented by Robinson, Bower and by the current study. 171 The data seem to suggest that the processes underlying return of fear occur immediately prior to re-exposure rather than at previous exposure. In contrast, Grayson et a l . (1982) and Sartory et a l . (1982) suggested that the processes which underlie return of fear arising from the presence of distracting activities occur either within or immediately following exposure. Habituation is a multi-process phenomenon (Watts, 1971), and different processes may be influential under different conditions. Ideally, memory judgements should be made at regular intervals to determine the precise time at which changes occur. However, continuous judgement-calls and rehearsal would f a c i l i t a t e r e c a l l , and may consequently mask differences that appear spontaneously at a later time. Bower (1980) believes that intense experiences are likely to be cognitively rehearsed between exposures more often than less intense experiences. Therefore, performers with high arousal levels may think about their performances more often than performers with low arousal levels. Furthermore, distortions of the cognitive representations of their emotional state whilst performing are particularly likely to occur i f the individual is physiologically aroused at the time of cognitive rehearsal ( i . e . , i f the individual is in a mood-congruent state). Hence, chronically highly aroused individuals should be particularly susceptible to cognitive distortion, as they are likely to rehearse previous exposure experiences whilst in an aroused state. Consequently, subjects with high levels of tonic heart rate were also expected to display overestimation of 172 previous fear and greater return of fear than subjects with lower tonic levels. However, high- and low-tonic-heart rate groups did not differ with regard to either pre-performance SUDS or the accuracy of their r e c a l l . Only in the analysis of extreme groups on the tonic-heart-rate distribution were differences found regarding level of fear at followup. Moreover, in post-hoc analyses, differences between the tonic heart rate levels of subjects with return of fear and subjects without return of fear were small and only approached significance. Therefore, data regarding tonic heart rate did not support the hypotheses. High-phasic-heart-rate subjects had significantly higher mean levels of tonic heart rate than low-phasic-heart-rate subjects, but phasic and tonic arousal were also partially independent. Approximately one third of the high-heart-rate subjects did not have tonic levels above the sample median. Given that intense, arousing experiences (that were characteristic of the exposures for highly phasically aroused subjects) generate frequent cognitive rehearsal, and given that such rehearsal could potentially e l i c i t concurrent physiological arousal and thereby promote overestimation of fear, the substantial number of high-phasic-heart-rate subjects in the low-tonic-heart-rate group may have masked differences between more and less tonically aroused subjects. In addition, the non-performance pulse rate measure was unlikely to be a true measure of resting heart rate as i t was recorded in the group setting. Continuous measurement of arousal over intervals between 1 73 exposures is required. Therefore, the possibility s t i l l remains that individuals with high levels of tonic arousal are likely to be in an aroused state when cognitively rehearsing previous or future exposures, and are therefore likely to experience return of fear. Distortion at recall reflects a change in the stimulus representation stored in memory in terms of the consolidation model. Dishabituaton occurs because the stimulus representation does not match the actual stimulus upon re-exposure. Memory data, therefore, did not support the hypothesis that arousal interferes with the processes of retrieval, but did support the hypothesis that arousal alters stimulus representations. Grayson et al.'s (1982) data seem to contradict the proposed effects of the salience of arousal cues. They found that distraction during expousre, which probably included distraction from arousal cues, facilitated return of .fear. However, their extensive distraction technique may have prevented an adequate association between stimulus representations and contextual cues, and impeded later retrieval. Joint consideration of data from Grayson et al s . study and the current study suggests that general attentiveness may be necessary for consolidation, but selective attention to arousal cues may fa c i l i t a t e subsequent distortion of cognitive representations and impede consolidation. It was originally hypothesised that by thinking about the fearful nature of performing, aroused performers would experience thought resensitization. As seen from the above 174 discussion, data from the Memory questionnarie tended to support this hypothesis. Data from the Thought questionnaire were also supportive, as high-heart-rate subjects, in general, reported more anxious thoughts prior to performing than low-heart-rate subjects. Similarly, Galassi, Frierson and Sharer (1981) found that highly test anxious individuals had more anxious thoughts and fewer positive thoughts than low test anxious individuals, that their thoughts were more related to their level of anxiety than to the task, and that they had more bodily sensations of arousal. Similar results were also obtained by Hollandsworth, Glazeski, Kirkland, Jones and Van Norman (1979). In addition, thoughts following post-assessment tended to be less fearful than those preceding followup-assessment performances in the return-of-fear group (though this effect did not quite reach st a t i s t i c a l significance). In contrast, scores on the Thought questionnaire remained relatively stable in the no-return-of-fear group. This data suggests that subjects with high arousal and who display return of fear tend to experience thought resensitization; or, focus upon the fearful quality of the exposure experience to a greater extent in anticipation of re-exposure than at completion of exposure. The direction of causality, however, can only be speculated. Most researchers have suggested that anxious thoughts e l i c i t fear, and that cognitive rehearsal produces anticipatory fear (Borkovec, 1976; Meyer & Reich, 1978). This study suggests only that fearful thoughts accompany return of fear, as either a cause or a consequence. The usefulness of the 175 kind of classification design employed herein rests not upon specifying causal relationships as do designs involving independent variable manipulations, but in identifying i n i t i a l subject characteristics for ,use in prediction of certain kinds of response to intervention. Sartory et a l . (1982) attempted to show that conditions which e l i c i t arousal produce fear-provoking thoughts, and hence, also produce return of fear. However, they succeeded only in assessing thought content when subjects were habituated, and therefore were no longer physiologically aroused. Consequently, differences were not found between their high- and low-heart-rate groups in the frequency with which "comfortable", "mildly uncomfortable" or "very uncomfortable" moments of the exposure were thought about. High-heart-rate subjects in the present study thought mostly about "uncomfortable moments" when they were in an anxious state (i.e . , prior to performance), while low-heart-rate subjects thought more about "comfortable moments" or did not think about their level of discomfort at a l l . Thinking about performing when in an aroused state seems likely to lead to resensitization of thought content. Highly aroused subjects have a different kind of experience during exposure than those who are less aroused. Consequently, they seem to have different recollections of their previous exposure, as suggested by Memory questionnaire data, and different expectations of the nature of their response to future encounters with the feared stimulus, as suggested by Thought questionnaire data. 1 76 The tendency of highly aroused subjects to expect discomfort is also consistent with Bower's (1981) associative" model of mood-dependent cognitions. He hypothesised that depressed mood not only enhances the recall of unpleasant events but also produces bleak interpretations of future events, and that analogous effects occur in anxiety mood states. Bower (1981) explains mood-dependent cognitions by the use of a network theory of memory. Emotions are allocated specific units in memory that are associated with other units representing the various aspects of that emotion (autonomic, verbal and so on). A unit is defined as a cluster of descriptive propositions that describe events and concepts. Emotion units are also believed to be associated with other units that represent events during which that emotion was aroused. Emotion units can be activated by many stimuli. The spread of activation throughout, other memory structures to which the emotion unit is connected is facilitated when the emotional state is e l i c i t e d . Consequently, the whole network is strengthened. Judgements regarding prior or future events are biassed by the avail a b i l i t y of positive and negative features that come to mind. Salient arousal cues are easily available, and their association with the emotional state of fear is strengthened during aroused states. It is speculated that during cognitive rehearsal, highly aroused individuals easily recall or attend to their strong arousal cues. Those cues have been associated with the state of fear through the associative networks. Hence, greater fear is expected, just as greater fear is recalled, by highly aroused individuals than by 177 individuals without salient arousal cues. In general, the above account for the patterns in reported fear differs from the cognitive-attentional models of test and performance anxiety proposed by Wine (1971) and Sarason (1978) in at least two ways: Wine and Sarason do not specify a change in thought content over intervals between exposures; and they did not make explicit the implications of attentional focus for the recall and prediction of fear. However, the current speculations are consistent with Foa's (1979) claim that between-session habituation requires cognitive changes that are permanent. Foa suggested that r i g i d , delusional belief systems account for the reappearance of fear in obsessive-compulsive patients. The present data suggest that states of high arousal similarly promote the reappearance of fearful cognitions. Perceived S k i l l Efficacy was considered a potentially important predictor of return of fear in fears involving a large s k i l l component. As previously mentioned, i n i t i a l efficacy (i . e . , perceived s k i l l ) neither related significantly to the occurrence of long-term return of fear, nor contributed significantly to the variance in the difference between fear levels at post-assessment and at followup-assessment. However, efficacy significantly decreased from post- to followup-assessment in subjects who experienced return of fear. Practise and s k i l l are important determinants of efficacy judgements (Bandura, 1977). The return-of-fear group was, in general, less skilled than the no-return-of-fear group. 178 Although s k i l l differences were quite small, the superioritiy of those without return of fear is consistent with Rachman's (1979) hypothesis that return of fear is likely when a s k i l l is being mastered. S k i l l f u l task performance has been associated with a strong sense of perceived control, and a strong-sense of predictability of the situation (Rachman, 1979). S k i l l differences may therefore account for the overall group differences with regard to efficacy. Ginsberg (1982) and others have shown that unpredictability facilitates dishabituation, and hence may enhance the likelihood of return of fear. However, reduced predictability associated with inferior s k i l l is unlikely to account fully for return of fear, as neither the classification groups, nor the return-of-fear groups, differed in terms of efficacy levels at post-assessment. Unlike smoking cessation research (DiClemente, 1981; Condiotte & Lichenstein, 1981), post-treatment levels of efficacy did not predict long-term outcome. The finding that the groups differed neither in the amount they practised overall nor in their patterns of practise over time is surprising, as efficacy levels are believed to be highly dependent on practise (Bandura, 1978; Rachman, 1978). However, an important distinction was revealed between exposure-relevant and exposure-irrelevant practise. Although general practise may influence one's confidence to perform for an audience without experiencing undue distress, practise in performance situations ( i . e . , exposure-relevant practise) appears to be the more crucial determinant of confidence for that situation. Levels of 179 confidence recorded over the followup interval correlated significantly with the number of performances over that period, as they did with the use of, and benefit achieved from, applying the training program s k i l l s to those performances. Confidence measures and other information gathered at completion of the f i r s t and second months of the followup interval were collected during phone contact. They are therefore susceptible to the effects of social desirability responding, especially as the assessor was also the experimenter. Tyron (1981) has commented upon the way in which social demand influences bias efficacy judgements. In reply, Bandura (1982) recommended that such influences are eliminated by private recording and the use of independent assessors. However, spurious elevation of confidence levels should not have affected the correlation with number of performances in the current study. The significant difference in number of performances over the followup interval between subjects who experienced return of fear and subjects who did not may therefore explain the former groups' reduction in efficacy over that same time period. Conversely, i t could be argued that the low efficacy which was characteristic of the return of fear group at pre-assessment resulted in avoidance behaviour, few public performances and, therefore, reduced perceived s k i l l . Bandura (1982) claimed that low self-efficacy results in slackening of efforts while high self-efficacy motivates effort to master challenges. However, since efficacy did not differ between the return-of-fear and no-return-of-fear groups at post-assessment, i t cannot account for 180 the difference in number of performances over the followup interval. The data illustrate that return of fear is enhanced when s k i l l acquisition is involved in the fear reduction process and when that s k i l l is not practised regularly (Rachman, 1983). The importance of the behavioural rehearsal of s k i l l s in criterion situations (i. e . , relevant practise) reinforces Meichenbaum's (1978) stress inoculation model of anxiety reduction, and is consistent with the evidence that advocates the importance of in vivo exposure for the treatment of fears such as agoraphobia (Marks et a l . , 1983). Loss of efficacy would not only predict increased fear, but also loss of s k i l l and increased physiological arousal according to self-efficacy theory. However, heart rate did not significantly increase and s k i l l did not significantly decrease over the followup interval in the return-of-fear group. Given that amounts of general practise did not differ either between those groups or over time, i t is not surprising that performance s k i l l remained stable, since actual competence may relate more to amounts of general practise than to performance practise. A variable that should have been assessed is the frequency of aversive performance experiences over the followup interval: Rachman (1978,1979) postulated that, in addition to long intervals and lack of practise, aversive experiences also lead to return of fear. Similarly, Watts (1974) stated that the extent to which anxiety spontaneously recovers between exposure sessions is likely to be related to phobic experiences between sessions. The relationship between number of performances, 181 return of fear and confidence may have been more precisely described had the subjects' satisfaction with their performances over the followup interval been assessed. In summary, despite the use of median splits on classification variables (a weaker, although more generalisable procedure than the comparison of extreme groups), group differences were found in the frequency of followup return of fear. Levels of phasic arousal, efficacy and s k i l l , and amount of relevant practise were a l l distinguishing characteristics. The evidence did not support the use of efficacy as a predictor of return of fear, but rather suggested that efficacy was a function of relevant practise. The combination of low arousal, s k i l l , and performance-experience apparently contributed to the maintenance of the improvements in subjective fear that accrued from the training program. Moreover, given the -relatively small differences in s k i l l between those who displayed return of fear and those who did not, and the absence of s k i l l differences between classification groups, heart rate seemed to be the subject-characteristic of most importance for those who experienced return of fear. Relevant practise is a random variable that can be manipulated. The evidence from this and from other studies supports the description of the return-of-fear phenomenon as an example of dishabituation: dishabituation is most likely in states of high arousal, as was return of fear in the current study; dishabituation diminishes over time, as does return of fear (Philips, 1985); dishabituation is likely to occur over a period 182 without exposure, as illustrated by the appearance of followup return of fear in the current study; and dishabituation occurs in response to potent and different stimuli, as does return of fear in response to different and fearful cognitive representations. That i s , return of fear was believed to occur via distortion of r e c a l l , or overestimation of previous fear levels, and thought resensitization, or enhanced expectation of distress. These processes were attributed to the effects of the cognitive rehearsal of exposure when in an aroused state. The salience of arousal cues in highly aroused subjects may enhance their availability in memory, and, in combination with a mood-congruent state, result in fearful interpretations of previous encounters and expectations of fear in future encounters. The result of these processes could be viewed within the consolidation model to be an alteration of the cognitive representation of the stimulus configuration. BETWEEN-SESSION RETURN OF FEAR The analysis and discussion of between-session return of fear was severely restricted by design limitations. Performances within any given session were quite short (3 to 5 minutes), and therefore did not allow sufficient time for within-session habituation for the majority of subjects. Research has shown that short stimulus presentations are not conducive to the maintenance of within-session habituation. Watts (1974) found that repeated 45-second presentations produced more long-term habituation and less spontaneous recovery between sessions than 15-second presentations. 183 Rabavalis, Boulougouris and Stefanis (1976) found that long practise with feared stimuli produced better results for obsessional patients, while short practise produced i r r i t a b i l i t y and anxiety in the majority of patients. The failure to allow habituation processes time to reach asymptotic levels not only restricted sample sizes in between-session analyses to the numbers of subjects who experienced within-session fear reduction, but may have also facilitated dishabituation, or the return of fear, at subsequent sessions, and impeded between-session habituation. As previously mentioned, performances in the training program were less standardised than the assessment-performances. The reduced predictability of training program sessions may have facilitated further the likelihood of between-session return of fear (Davis, 1970; Ginsberg, 1982). Similarly, Grey et a l . (1979) suggested that stimulus presentations which follow a prearranged order, regardless of previous fear levels, are less effective than constant stimulus presentations in producing habituation because they are less predictable. The slight changes in the d i f f i c u l t y level of the piece chosen to perform in training program sessions, and in the amount of practise of the piece, in addition to the short length of exposure, were likely to faci l i t a t e between-session dishabituation and impede within-session habituation. However, despite these factors, between-session habituation was apparent - shown by the linear reduction in anticipatory fear over the course of the program. Processes operating against the probability of between-session return of fear were also present. 184 The less demanding nature of training program performances compared to assessment-performances may have mitigated between-session return of fear, since dishabituation occurs in response to strong, extra or different stimuli (Groves & Thompson, 1970). Philips (1985) did not find evidence of between-session return of fear in anxious subjects until a potent stimulus (i . e . , sound) was introduced. In addition, during-performance SUDS correlated quite strongly with the subjects' rating of their satisfaction with their performances. Similarly, Kendrick et a l . (1982) reported a study by Sanders (1967) in which speech anxiety only reduced as a result of repeated performances before an audience when the speaker performed well. The evaluative standard is an additional dimension to consider in performance tasks in comparison to the passive and non-skilled exposure tasks characteristically employed in other return-of-fear studies. Moreover, the results may be obscured by the failure to.record fear levels at the last few seconds of performance, at which time fear levels may have reduced from during-performance levels. By restricting the analyses to subjects who experienced within-session fear reduction of at least 10 units, based on during-performance SUDS ratings, subjects who could have shown within-session reduction based on an end-of-performance fear rating may have been excluded. Therefore, relatively few subjects were identified as return-of-fear subjects between the second and third sessions (n=14), the third and fourth sessions (n=9), and between the fourth session and post-assessment 185 (n=12). Comments regarding this phenomenon should be considered tentative and in need of further examination. Between-session return of fear was only slightly more apparent in subjects who experienced followup return of fear (50%) than in subjects who did not (40%). This lack of consistency may arise from either false negatives in identifying between-session return of fear due to the above mentioned design faults, or from a difference between the mechanisms involved in short-term and long-term return of fear. The finding that between-session return of fear was more prevalent in high-heart-rate subjects than in low-heart-rate subjects suggested that mechanisms similar to those proposed for long-term return of fear operated for between-session return of fear. However, reverse analyses that compared subjects who displayed between-session return of fear with those who did not failed to demonstrate significant group differences with regard to either efficacy or heart rate. The 26 subjects who displayed at least one between-session return of fear were also equally distributed across the high- and low-tonic-heart-rate groups (13 in each group). However, faulty design may again be the cause of this somewhat surprising finding. It has been shown that high levels of arousal delay within-session habituation (Deitz, 1982; Lader et a l . , 1967; Lader & Wing, 1963; Parkinson & Rachman, 1980). Subjects with high heart rates were even less likely than other subjects to experience within exposure fear reduction in consideration of the relatively short exposure time, and, consequently, may have been inadvertently excluded from the 186 possibility of displaying between-session return of fear. Average heart rates of those displaying between-session return of fear at each interval were lower than corresponding averages for high-heart-rate subjects : 77, 84 and 81 beats per minute at the third and fourth training program sessions and at post-assessment for between-session return-of-fear subjects, compared to the corresponding values of 83, 87 and 89 beats per minute for high-heart-rate subjects. The intriguing finding, however, is that the same process variables suggested by the analysis of long-term return of fear were also suggested by the analysis of between-session return of fear. That i s , subjects who experienced between-session return of fear overestimated their previous level of during-performance fear and reported slightly more anxious thoughts prior to performance than at the completion of their previous performance. Given that high-heart-rate subjects were previously shown to overestimate their fear and to experience thought resensitization, differences between those who displayed between-session return of fear and those who did not on these variables may have been strengthened had the c r i t e r i a for between-session return of fear not inadvertently excluded subjects with high heart rates. The between-session data also suggest that these processes can occur in the absence of high arousal, although the followup return-of-fear data suggest that they are potentiated by the presence of high arousal. It is also possible that the brief exposures used in the current study promoted the overestimation of fear at previous performances and 187 thought resensitization. Again, long exposures may have produced different results. Number of performances (or, relevant practise), which was identified as an important variable in long-term return of fear, was d i f f i c u l t to assess over the week interval between sessions owing to very low frequencies. Systematic differences were not apparent, however, between subjects who demonstrated between-session return of fear and subjects who did not. Relevant practise may be an important variable only over long intervals. An interesting finding to emerge was the significant positive correlation between return of fear across sessions and satisfaction for the previous performance: subjects who experienced between-session return of fear were consistently more satisfied with their previous performance than other subjects. However, when asked to recall their fear levels one week later, they overestimated their level of fear and also thought more fearfully about their current performance. Their responses indicate further the resensitizing nature of the change in cognition over the interval between sessions. They also lend further credence to the hypothesis that the mechanisms underlying return of fear seemed to occur either over the interval or just prior to subsequent performance rather than at the previous performance. On the other hand, high satisfaction ratings could also be an artifact of the selection c r i t e r i a , as only subjects with during-performance SUDS reductions in the previous performance were considered: during-performance SUDS and satisfaction were negatively correlated. 188 The relationship between short-term and long-term return of fear was not adequately tested because of the afore-mentioned design limitations. The study does show that long-term return of fear may be expected when a standard treatment paradigm is provided for a l l subjects and when there is evidence of at least some between-session return of fear (and perhaps more than was evident using the current c r i t e r i a ) . In contrast, Philips (1985) did not find long-term return of fear following individually tailored treatment paradigms in which exposure sessions were continued until between-session return of fear abated. Continued reductions in fear and arousal resulting from repeated behavioural rehearsal may reduce the likelihood of overestimation of previous fear and resensitization of thoughts. Philips' data is also consistent with the data which indicates that dishabituation does not readily occur once habituation reaches an asymptotic level (Groves & Thompson, 1970). DISCORDANCE AND DESYNCHRONY The state of return of fear reflects a state of discordance in which a return in levels of subjective fear is not matched by returns in other response systems. It was observed that changes or improvements in heart rate and s k i l l were largely maintained over the followup period, whereas subjective reports of fear lost their gains in some subjects. However, in view of the fact that the i n i t i a l improvements in s k i l l were relatively minor, the data mainly indicate a discordance between the subjective and physiological systems at followup. At the same time, the interactive nature of these response systems is recognised in 189 the proposition that high levels of physiological arousal may account, to at least some extent, for the elevations in the subjective system known as return of fear. Synchrony was observed over the course of the training program in subjects with high heart rates. The same subjects also displayed the most return of of fear. Desynchrony was observed, to some extent, in subjects with low heart rates. Those subjects displayed the least return of fear. Grey et a l . (1979), Grey et a l . (1981) and Sartory et a l . (1982) recorded return of fear following synchronous reductions in fear. Grayson et a l . (1982), on the other hand, observed return of fear following desynchrony between heart rate and subjective fear during exposure. However, those studies referred to synchrony and desynchrony of within-session fear reductions, whereas evidence from the present study regarding synchrony and desynchrony referred to changes across sessions. Within-session synchrony could not be assessed without continuous measurement of heart rate and subjective fear. Barlow et a l . (1980) examined cross-trial synchrony and desynchrony. They reported "relapse" following a desynchronous pattern of reductions in reported fear and accelerations in heart rate. As mentioned, a similar desynchronous pattern in the present study was not associated with return of fear. The "desynchronous" subjects remained at relatively low levels of arousal despite their pattern of heart rate acceleration. In contrast, Barlow et al.'s "desynchronous" subject attained very high levels of heart rate. Hence, the subsequent relapse they reported may have 190 occurred due to the high level of arousal per se. Grey et a l . (1981) reported that discordance between response systems at i n i t i a l assessment predicts greater return of fear than concordance. It was not possible to assess i n i t i a l discordance in the format specified by Grey et a l . (1981) (i. e . , zero subjective fear and pronounced physiological arousal), as performers were excluded from the current study if their i n i t i a l fear level was not at least at a moderate level. When discordance and concordance were assessed by examination of the congruity between a subject's classification in the upper and lower quartiles of the heart rate and pre-performance SUDS distributions, seven subjects only were found to be discordant, of whom only three displayed followup return of fear. Although that assessment did not represent an adequate test of discordance, i t was consistent with Sartory et a l s . (1982) failure to find any relationship between i n i t i a l discordance and the likelihood of return of fear one week following exposure. Minor i n i t i a l discordance in the present study was not associated with long-term return of fear. Also, the finding that return of fear was more prevalent in high-heart-rate groups following covariation of i n i t i a l fear levels than in low-heart-rate subjects suggested the importance of arousal level per se, as opposed to the importance of the relative status of the physiological and subjective response systems. The high demand level of the exposure in this study may have provoked more discordance than is usually obtained: the musician must be in absolute control of motor coordination involving the finest muscle action, has 191 to trust memory, and at the same time must feel and project the music with authority and conviction ... any concentration lapses or overwhelming feelings of anxiety are likely to lead to immediate and drastic deterioration of performance. (Kendrick et a l . , 1982, p. 354). High demand evoked either by instructions to stay in the . presence of the feared stimulus/situation despite fear, or by highly motivated or "courageous" behaviour (Rachman, 1978) produces discordance: "subjects are able to control a tendency towards flight in spite of autonomic and experiential signs of fear" (Hodgson & Rachman, 1974, p. 321). Musical performance is a situation from which escape is d i f f i c u l t , as evidenced by the fact that on only two occasions did subjects leave the studio before completing their performance. Correlations among the dependent variables were strongest within the subjective response system (pre-performance SUDS, pre-performance thoughts and efficacy). Their strength is consistent, with the report that inter'correlations among measures within a response system are generally stronger than those between systems (Lang, 1971). Cross response-system correlations were strongest in the correlation of heart rate with either pre-performance SUDS or pre-performance thoughts. S k i l l level did not correlate significantly with the subjective report of fear, efficacy or with anticipatory thoughts. It is possible that an additional measure of the behavioural response system, such as the rating of observable signs of anxiety, may have provided additional correlational information, but the r e l i a b i l i t y and validity of such behavioural ratings has been 1 92 questionned. Craske and Craig's (1984) failure to find a strong correlation between performance quality and subjective measures of anxiety was attributed to the uncoupling effect of high demand exposure: behavioural and self report measures typically correlate quite well in nondemanding situations (Rachman, 1978). Demand generated discordance was especially likely in the current study, as subjects were generally quite s k i l l e d , and thus able to perform well in spite of distress. Also, the requirement to play a piece that they had already mastered may have masked some effects of anxiety upon performance quality. The same factors could be used to explain the weak correlation between the behavioural and physiological response systems (Craske & Craig, 1984; Kendrick et a l . , 1982). Others have found that low heart rates accompany superior performance. Cox, Hallam, O'Connor and Rachman (1982), for example, found that bomb disposal operators who had been specially trained and were skilled at handling stressful situations, displayed lower heart rate reactions to a stressful task than control subjects. Also, Fenz (1975) found that parachutists who performed best had significantly lower heart rates at c r i t i c a l points in the jump cycle than those who performed less well. However, those studies were conducted on groups that represented the f u l l range of fear levels in their respective populations. The restriction in the current study to subjects who reported at least moderate fear may have truncated the correlation between heart rate and s k i l l , because heart rate has been shown to relate more closely to subjective fear than to s k i l l (Rachman, 1978). 193 The correlational evidence supports the three-systems model of fear by showing that the response systems are only loosely coupled and can change relatively independently of each other; especially the behavioural from the other response systems. In addition, the presence of slightly stronger correlations in the return-of-fear group than the no-return-of-fear group is consistent with the hypothesis that more concordance than discordance is expected during intense emotional arousal (Hodgson and Rachman, 1974): the return-of-fear group were generally characterised as being highly autonomically aroused. However, the strengths of the correlations did not significantly differ between the return-of-fear groups. Significant differences may have been militated by the low s t a t i s t i c a l power of correlational analyses - based on relatively small sample sizes. SELF-EFFICACY THEORY The data question the value of self-efficacy (or, perceived s k i l l ) as a superordinate variable that mediates changes in a l l response systems. Bandura's self-efficacy model i s , in general, a more interactive model of fear and anxiety than the three-systems model. The latter specifies conditions under which concordance and discordance are most likely to occur. Self-efficacy theory predicts greater concordance amongst the response systems than is characteristic of the return-of-fear phenomenon. Correlations among the major dependent variables in this study indicated that efficacy correlated quite strongly with pre-performance SUDS and pre-performance thoughts when 194 assessed simultaneously, and with number of performances. It did not correlate with either heart rate or s k i l l . The correlational information concurs with evidence presented by Kendrick et a l . (1982) and Craske and Craig (1984). Efficacy and heart rate were related only to the extent that these two variables occurred more often in a concordant than a discordant fashion in the median splits classification: high efficacy was more often found in subjects with low heart rates than in subjects with high heart rates. Hence, groups 1 and 4 were considerably larger than groups 2 and 3. However, e,fficacy did not correlate significantly with heart rate, and the relationship between heart rate and return of fear was not mediated by efficacy. Self-efficacy theory assumes that people are capable of accurately judging their future behaviour. Meyer and Reich (1978), however, found that phobics tended to underestimate their capacity for coping when experiencing threat. Biran and Wilson (1982) reported similar inaccuracies in judgement. The frequent discrepancies found in the current study between confidence for a performance to be conducted in one week and confidence at the time of that performance (seen in the analysis of efficacy differences) illustrated the lack of generalisability of efficacy judgements for a specific performance task over a one-week interval. Confidence estimates may be influenced by current fear levels or by the number of performances over the preceding week. Bandura (1977) briefly described the loss in predictive power of efficacy judgements 195 that may arise from faulty self- knowledge, musjudgement of task requirements, unforeseen situational constraints, disincentives to act upon self efficacy and new experiences that prompt reappraisals of self efficacy. These factors allow for considerable discrepancy between efficacy judgements and future action. They are less likely to be present and influential in the assessment of circumscribed fears in tightly controlled laboratory settings that is characteristic of the majority of self-efficacy research (Lang, 1978) than in "real l i f e " situations. GENERALISABILITY OF THE RESULTS The high demand level of the assessment-performance, which has been shown to enhance discordance and to maintain elevated heart rate, limits the generalisability of these findings to other return-of-fear studies. Borkovec (1974) attributed the failure of anticipatory and reactive physiological arousal to habituate over repeated testing in an analogue social interaction procedure to the high demand level of that setting. Therefore, subjects' heart rates in the current study may not have decreased as much as they would have under low-demand conditions. Hence, return of fear may have been less evident in low-demand conditions than in the present conditions, when the proposed connections between arousal levels and the occurrence of, and processes underlying, return of fear are considered. The type of exposure paradigm also differs from those used in other return-of-fear studies. Previous studies have typically implemented passive exposure, such as observing a 196 snake at varying distances, and have entailed l i t t l e behavioural rehearsal or s k i l l acquisition. Fears associated with the performance of skilled tasks may entail more cognitive components (related to performing) than non-skill fears, especially as research regarding test anxieties has shown that the attentional-cognitive components of treatment are most appropriate for fear reduction (Wine, 1971). Therefore, the data may generalise only to fears that involve s k i l l f u l performance, such as social phobia, test anxiety and agoraphobia. They may have limited generalisability to simple phobias. Moreover, a l l response systems were targetted by the intervention procedure used in the current study (heart rate, cognitions and performance). Other studies that examined return of fear primarily, entailed exposure, and sometimes relaxation. Some authors believe that treatment changes are long lasting only i f a l l response systems are targetted by treatment procedures (Popler, 1977). However, the prevalence of return of fear following the current multi-procedural program suggests that subject-characteristics are an important modifying variable. Researchers are currently addressing the complex interaction between subject response-profile and treatment type (Norton, Dinardo & Barlow, 1983; Ost, Johansson & Jerrelmalm, 1982). The assessment of return of fear at a long-term followup also differs from most previous research. Other studies in this area (excluding Philips, 1985) have involved circumscribed fears 1 97 and short-term treatment outcome. In terms of the generalisability of the findings to fear and anxiety research in general, this study, in most ways, satisfied the conditions outlined by Borkovec and Rachman (1979). The fear-eliciting condition of public performance resembles those that e l i c i t performance anxieties held by socially anxious and test anxious individuals. The intensity of the response, as already discussed, was also comparable to data from other phobias and fears. CONCLUSION In conclusion, the known characteristics of return of fear wil l be reviewed, and areas in need of further investigation will be considered. Return of fear is a partial return in subjective discomfort that formerly underwent decline. It represents the experimental counterpart of the c l i n i c a l phenomenon of relapse. Regardless of the behavioural and physiological components of the fear response, a return in level of subjective discomfort is often sufficient to warrant reinstitution of therapeutic intervention. Return of fear appears to be a robust phenomenon, as indicated by the almost fifty-percent incidence of followup return of fear recorded in this study, and by the consistency with which i t is observed in independent studies. Prevalence in performance anxieties, however, may differ from, and should be compared with, prevalence in simple fears (e.g., small animal fears). The practise of s k i l l s (which was found to be significantly associated with return of fear in performance fears) may be less 198 important in simple fears, and hence, return of fear may be less common than in performance fears. The return-of-fear phenomenon has been observed in simple fears, obsessions, and, as demonstrated herein, performance anxiety. It-has been observed following passive and active exposure, and has been mostly recorded over one-week intervals. This study also recorded return of fear after three months. The relationship between short-term, between-session return of fear and long-term, followup return of fear remains to be c l a r i f i e d . Under individually tailored treatment conditions in which exposure sessions continue until short-term return of fear abates (Philips, 1985), long-term return of fear was not apparent. Under standard treatment paradigms that are typical of research and are sometimes c l i n i c a l l y employed, the presence of between-session return of fear persisting throughout treatment may enhance the likelihood of long-term return of fear. In the latter case, between-session return of fear may serve to identify individuals likely to need "booster" therapy sessions several months following treatment completion. The assessment of the relationship between short-term and long-term return of fear requires a design similar to the one employed herein, but with sufficient exposure duration to allow more extensive within-session habituation. In addition, examination of the ease with which fear is reduced after the appearance of long-term return of fear would provide guidelines for the therapist and client regarding expected number of "booster" sessions. 199 Several characteristics and conditions have been empirically associated with return of fear. Return of fear does not seem to depend upon the speed nor amount of within-session fear reduction (Rachman, 1979). High i n i t i a l fear level (Grayson et a l . , 1982; Philips, 1985; Rachman & Levitt, in press) and high i n i t i a l heart rate (Grey et a l . , 1981) (independent of fear level as shown in the current study) have been recorded from subjects who display return of fear. Distraction during exposure (Grayson et a l . , 1982), and immediately following exposure (Sartory et a l . , 1982), also seems to contribute to return of fear, whereas rehearsal under relaxed conditions (Sartory et a l . , 1982) mitigates return of fear. High demand treatment paradigms were associated with frequent return of fear in Grey et al.'s (1979) study, but were not in Grey et al.'s (1981) study; a discrepancy possibly due to differences in methodology and in need of further investigation. Similarly, the predictive value of i n i t i a l discordance between the subjective and other response systems (Grey et a l . , 1981) has not been substantiated (Sartory et a l . , 1982), and awaits the direct comparison of i n i t i a l l y concordant and discordant groups of subjects. Barlow et a l . (1980) found that desynchronous response-system changes were characteristic of a subject who subsequently experienced followup return of fear, but the role of accelerated arousal level per se was not ruled out. It has also been suggested that overvalued-ideation results in return of fear in obsessional patients (Foa, 1979). The present study lent support to hypotheses proposed by Rachman 200 (1979) by demonstrating that lack of s k i l l , and lack of practise of a skilled task over lengthy intervals, seem to exacerbate return of fear. A reduction in self-efficacy following program completion was also shown to characterise subjects who displayed return of fear, but i n i t i a l self-efficacy was not associated with return of fear. Finally, the theoretically-based expectation that sudden, unpredictable, intense and different stimuli promote return of fear requires empirical investigation. This can be exmained either indirectly, by recording the frequency and type of encounters with fear-related stimuli in individuals who experience return of fear and individuals who do not, or directly, by examining the effect of presenting t r i a l s of sudden, intense or different stimuli between t r i a l s of standard target stimuli. The phenomenon of dishabituation of response strength aptly describes return of fear. Dishabituation is a response tendency that is enhanced by many of the factors predictive of return of fear. The applicability of the habituation model to return of fear can be more fully assessed by examining the frequency of return of fear following manipulations known to produce dishabituation, such as: failure of response decrement within t r i a l s to achieve asymptotic status; presentation of different stimuli between t r i a l s ; and presentation of sudden or intense stimuli. As suggested by the present results, the effects of these manipulations should be potentiated by heightened arousal. In accordance with habituation research, rehabituation or fear reduction should occur more rapidly following dishabituation, or 201 return of fear. Also, the magnitude of dishabituation (as was found by Philips, 1985, in the case of return of fear) should diminish over repeated instances. Several explanatory hypotheses for return of fear have been tested. A reactive inhibition mechanism was suggested on the basis of preliminary evidence regarding demand level of exposure. That hypothesis was not pursued because return of fear was not enhanced by conditions that augmented the build up of inhibition (Grey et a l . , 1981). The significance of i n i t i a l discordance and desynchrony have been speculated, although without theoretical basis, and i t remains to be demonstrated that their effects are not the result of heightened arousal per se. Consolidation models of dishabituation were used to explain the potentiating effects of distraction regarding return of fear and the beneficial effects of cognitive rehearsal. It was suggested that rehearsal strengthens the consolidation of within-session habituaton by lengthening exposure. However, data from the present study question the simplicity of that explanation. The effects of rehearsal seem to interact with level of autonomic arousal: heightened arousal may stimulate the development of fearful cognitive representation, and thereby fa c i l i t a t e return of fear. The power of cognitive representations, which are either consolidated and hence contribute to the maintenance of habituation, or are altered and hence contribute to dishabituation, is strongly suggested by data from this study. Whether fearful cognitive representations influence fear levels directly or, as suggested by Wagner 202 (1974), their subsequent mismatch with actual stimulus presentations results in dishabituation, is not clear. The consistency of the cognitive-representational account of return fear with Bower's (1981) account of mood-dependent cognition is testable: fearful cognitive representations, and hence greater return of fear at re-exposure, should occur in congruent mood states than in incongruent mood states. This was indirectly shown by Sartory et a l . (1982) who found that cognitive rehearsal in relatively relaxed states ( i . e . , incongruent mood) was not associated with substantial return of fear. The study reported herein found evidence for more fearful cognitive representations immediately prior to re-exposure and greater return of fear in highly aroused performers (i.e., congruent mood state) than in less aroused performers. Cognitive representations should be directly compared in relaxed versus agitated states. Moreover, the influence on cognitive representations of different types of encounters with fear related events between treatment t r i a l s , and the expectation that tonic excitability increases the probability that spontaneous cognitive rehearsal of feared stimuli occurs in mood-congruent states, could be assessed. The importance of cognitive representations should be compared in nonskilled-based fears, given that cognition may have greater functional importance in performance anxieties. On a more general level, return of fear can be viewed as a sign of a failure to completely absorb and process emotional reactions, or, incomplete emotional processing (Rachman, 1980). 203 Accordingly, controlled autonomic responses, repeated exposure and predictable stimuli f a c i l i t a t e emotional processing and impede return of fear. Once emotional processing has been more fully operationalised, return of fear may be predicted from the amount of emotional processing that has taken place during treatment. Finally, l i t t l e is known of the treatment of return of fear. It seems that continuation of exposure by increasing the number of exposure t r i a l s (as shown by Philips, 1985), by regular relevant practise between exposure t r i a l s (as suggested by the current study and by other researchers who note the importance of behavioural practise "homework assignments"), and perhaps by lengthening t r i a l duration to ensure that asymptotic levels of fear reduction are achieved, may decrease the likelihood of return of fear. The treatment implications arising more directly from this study, and in need of empirical evaluation, include the beneficial effect of cognitive rehearsal prior to exposure whilst in a relaxed state. Relaxation was brief, self administered, and was induced after the assessment of cognitions and fear in the current study . A complete examination requires the assessment of fear and cognitions before and after intensive relaxation, under conditions that promote cognitive rehearsal compared to conditions that do not, and an assessment of the influence of those conditions on subsequent fear responses. The benefit of memory aids that minimise the distortion of prior experiences should be explored for individuals whose recall and expectation of fear intensify 204 over intervals between exposures. Finally, habituation theory emphasises the importance of stimulus generalisability and suggests that a wide range of stimuli be included in exposure t r i a l s to prepare individuals for encounters with unexpected or different stimuli outside the treatment setting. 205 REFERENCES Agras, W., & Jacob, R. (1981). Phobia: Nature & measurement. In M. Mavissakalin & D. Barlow (Eds.). Phobia:  Psychological & pharmacological treatment. New York: Guilford Press. Andrasik, F., Turner, S., & Ollendick, T. (1980). Self report & physiologic responding during in vivo flooding. Behaviour  Research & Therapy, 18, 593-594. Appel, S. (1976). Modifying solo performance anxiety in adult pianists. Journal of Music Therapy, 13, 2-16. Baddeley, A.D. (1972). Select attention and performance in dangerous environments. British Journal of Psychology, 63 (4), 537-546. Bandura, A. (1977). Self efficacy: Toward a unifying theory of behavioural change. Psychological Review, 84, 191-215. Bandura, A. (1978). Reflections on self efficacy. Advances in  Behaviour Research & Therapy, J_, 237-269. Bandura, A. (1980). Gauging the relationship between self-efficacy judgement & action. Cognitive Therapy & Research, 4(2), 263-268. Bandura, A. (1981). In search of pure unidirectional determinants. Behavioural Therapy, 12, 30-40. Bandura, A. (1982). Self-efficacy mechanism in human agency, American Psychologist, 37, 122-147. Bandura, A. (1982). The assessment and predictive generality of self percepts of efficacy. Journal of Behaviour Therapy  and Experimental Psychiatry, 13(3), 195-199. Bandura, A., & Adams, N. (1977). Analysis of self-efficacy theory of behavioural change. Cognitive Therapy & Research, 1(4), 287-310. 206 Bandura, A., Adams, N., & Beyer, J . (1977). Cognitive processes mediating behavioral change. Journal of  Personality & Social Psychology, 35(3), 125-139. Bandura, A., Adams, N., Hardy, A., & Howells, G. (1980). Tests of the generality of self-efficacy theory. Cognitive Theory  & Research, 4(1), 39-66. Bandura, A., Reese, L., & Adams, N. (1982). Microanalysis of action & fear arousal as a function of different levels of perceived self-efficacy. Journal of Personality & Social  Psychology, 43(1), 5-21. Bandura, A., Taylor, C , Williams, S., Mefford, I., & Barchas, J. , (in press). Catecholamine secretion as a function of perceived coping self-efficacy. Journal of Consulting &  Clinical Psychology. Barlow, D., Mavissakalin, M., & Scholfield, L. (1980). Patterns of desynchrony in agoraphobia: A preliminary report. Behaviour Research & Therapy, _1_8(5), 441-448. Barlow, D., & Mavissakalin, M. (1981). Directions in the assessment and treatment of phobias: In M. Mavissakalin & D. Barlow (Eds.). Phobia: Psychological & pharmacological treatment.. New York: Guilford Press. Bernstein, D., & Paul, G. (1971). Some comments on therapy analogue research with small animal phobias. Journal of  Behaviour Therapy & Experimental Psychiatry, 3_, 225-237. Biran, M., & Wilson, G.T. (1981). Treatment of phobic disorders using coginitive and exposure methods: A self-efficacy analysis. Journal of Consulting and Clinical  Psychology, 49(6), 886-899. Borkovec, T. (1974). The role of expectancy & physiological feedback in fear research: A review with special reference to subject characteristics. Behaviour Therapy, 491-505. Borkovec, T. (1976). Physiological & cognitive processes in the regulation of anxiety. In G. Schwartz & D. Shapiro (Eds.). Consciousness & self regulation: Advances in Research: Vol. 1. New York: Plenum Press. 207 Borkovec, T., & O'Brien, G. (1976). Methodological & target behaviour issues in analogue therapy outcome research. Progress in Behaviour Modification, 3^ 133-172. Borkovec, T. , & Rachman, S. (1979). The u t i l i t y of analogue research. Behaviour Research & Therapy, 17, 253-261. Borkovec, T., & Sides, J . (1979). C r i t i c a l procedural variables related to the physiological effects of progressive relaxation: A review. Behaviour Research &  Therapy, _T7, 119-125. Borkovec, T., Stone, N., O'Brien, G., & Kaloupek, D. (1974). Evaluation of a c l i n i c a l l y relevant target behaviour for analog outcome research. Behaviour Therapy, 5, 503-513. Borkovec, T., Weerts, T., & Bernstein, D. (1977). Assessment of anxiety. In A. Ciminero, K. Calhoun & H. Adams (Eds.), Handbook of Behavioral Assessment. New York: John Wiley & Sons, Inc. Boulougouris, J . , Rabavalis, A., & Stefanis, C. (1977). Psycho-physiological responses in obsessive-compulsive patients. Behaviour Research & Therapy, 15, 221-230. Bower, G. (1981). Mood and Memory. American Psychologist, 36 ( 2 ) , 129-148. Coniotte, M. , & Lichenstein, E. (1981). Self efficacy and relapse in smoking cessation. Journal of Consulting and  C l i n i c a l Psychology, 4_9(5), 648-658. Connolly, J . (1979). Tonic physiological responses to repeated presentations of phobic stimuli. Behaviour Research &  Therapy, J_7, 189-196. Cox, D., Hallam, R., O'Connor, K., & Rachman, S. (1982). An experimental analysis of fearlessness & courage. Unpublished manuscript. Craske, M. , & Craig, K.D. ( 1 984). Musical performance anxiety: The three-systems model and self-efficacy theory. Behaviour  Research & Therapy, 2 2 ( 3 ) , 267-280. 208 Davis, M. (1970). Effects of interstimulus interval length and , variability on startle-response habituation in the rat. Journal of Comparative & Physiological Psychology, 72(2), 177-192. Decker, T., & Russell, R. (1981). Comparison of cue-controlled relaxation and cognitive restructuring versus study s k i l l s counseling in treatment of test-anxious college underachievers. Psychological Reports, 49, 459-469. Deitz, S. (1982). Individual differences in electrodermal response conditioning & self report of discomfort: A phobia analogue. Physiological Psychology, H)(2), 239-245. de Silva, P., & Rachman, S.J. (1981). Is exposure a necessary condition for fear-reduction? Behaviour Research & Therapy, J_9, 227-232. DiClemente, C. (1981). Self-efficacy & smoking cessation maintenance: A preliminary report. Cognitive Therapy &  Research, 5, 175-187. Eich, E., Reeves, J . , Jaeger, B., & Graff-Radford, S. (in press). Memory for Pain: Relation between past & present pain intensity. Nature. Feltz, D. (1982). Pathanalysis of the causal elements in Bandura's theory of self efficacy and an anxiety-based model of avoidance behviour. Journal of Personality and Social  Psychology, 42(4), 764-781. Foa, E. (1979). Failure in treating obsessive-compulsives. Behaviour Research & Therapy, 17, 169-176. Foa, E., & Chambless, D. (1978). Habituation of subjective anxiety during flooding in imagery. Behaviour Research &  Therapy, 16, 391-399. Foa, E., Jameson, J . , Turner, R., & Paynes, C. (1980). Massed versus spaced exposure sessions in the treatment of agoraphobia. Behaviour Research .& Therapy, 18, 333-338. Fremouw, W., & Zitter, R. (1978). A comparison of s k i l l s training and cognitive restructuring-relaxation for the 209 treatment of speech anxiety. Behaviour Therapy, 9, 248-259. Galassi, J . , Frierson, H., & Sharer, R. (1981). Behvior of high, moderate & low test anxious students during an actual test situation. Journal of Consulting & Clin i c a l Psychology , 49(1), 51-62. Gauthier, J . , & Marshall, W. (1977). The determination of optimal exposure to phobic stimuli in flooding therapy. Behaviour Research & Therapy, 15, 403-410. Ginsberg, S. (1982). Effect of attentional & neutral instructions and of forward and backward conditioning analogs on response to change from paired stimulus re p i t i t i o n : more eveidence relevant to the role of orienting reaction recovery in short-interval classical autonomic conditioning. Biological Psychology, 15, 33-48. Gogan, P. (1970). The startle and orienting reactions in man. A study of their characteristics and habituation. Brain  Research, 18, 117-135. Goldfried, M., Linehan, M., & Smith, J . (1978). Reduction of test anxiety through cognitive restructuring. Journal of  Consulting & Clinical Psychology, 46(1), 32-39. Grayson, J . (1982). The eli c i t a t i o n and habituation of orienting and defensive responses to phobic imagery and the incremental stimulus, intensity effect. Psychophysiology, 19 (1) , 104-111. Grayson, J . , Foa, E., & Steketee, G. (1982). Habituation during exposure treatment: Distraction vs Attention-focusing. Behaviour Research & Therapy, 20(4), 323-328. Grey, S., Rachman, S., & Sartory, G. (1981). Return of fear: The role of inhibition. Behaviour Research & Therapy, 19, 135-143. Grey, S., Sartory, G., & Rachman, S. (1979). Synchronous & desynchronous changes during fear reduction. Behaviour  Research & Therapy, 17, 137-147. Gross, R., & Fremouw, W.. ( 1 982) Cognitive restructuring and 210 progressive relaxation for treatment of empirical subtypes of speech anxious subjects. Cognitive Therapy & Research, 6 (4), 429-436. Groves, P., & Lynch, G. (1972). Mechanisms of habituation in the brain stem. Psychological review. , 79(3), 237-244. Groves, P., & Thompson, R. (1970). Habituation: A dual process theory. Psychological Review, 77(5), 419-450. Hallam, R., & Rachman, S. (1980). Courageous acts or courageous actors. Personality & Individual Differences, J_ (14), 341-346. Hassett, J . (1978). A Primer of. Psychophysiology. San Francisco: W.H. Freeman & Co. Hodgson, R., & Rachman, S. (1974). I I . Desynchrony in -measures of fear. Behaviour Research & Therapy, 12, 319-326. Hollandsworth, J . , Glazeski, R., Kirkland, K., Jones, G., & Van Norman, L. (1979). An analysis of the nature & effects of test anxiety: Cognitive, behavioural & physiological components. Cognitive Therapy & Research, 3 ( 2 ) , 165-180. Holroyd, K., Westbrook, T., Wolf, M., & Badhorn, E. (1978). Performance, cognition, & physiological responding in test anxiety. Journal of Abnormal Psychology, 87(4), 442-451 . Hugdahl, K. (1981). The three-systems-model of fear and emotion - A c r i t i c a l examination. Behaviour Research &  Therapy, 19, 75-85. Johansson, J . , & Ost, L. (1982). Perception of autonomic reactions and actual heart rate in phobic patients. Journal  of Behavioral Assessment, 4(2), 133-143. Kazdin, A. (1978). Conceptual & assessment issues raised by self-efficacy theory. Advances in Behaviour Research &  Therapy, 1_, 177-185. Kazdin, A. (1980). Research Design in Cli n i c a l Psychology. 21 1 New York: Harper & Row, Pub. Kendrick, M. (1979). Reduction of musical performance anxiety  by attentional training and behavioural rehearsal: An  exploration of cognitive mediational processes. Doctoral dissertation, University of British Columbia, Vancouver. Kendrick, M., Craig, D., Lawson, D., & Davidson, P. (1982). Cognitive & behavioural therapy for musical performance anxiety. Journal of Consulting & Clinical Psychology, 50 (3), 333-362. Keppel, G. (1973). Design and analysis: A researcher's  handbook. New Jersey: Prentice-Hill Inc. Kirk, R. (1982). Experimental design: Procedures for the  behavioral sciences. California: Brooks/Cole Pub. Co. Kleinknecht, R. (1982). The origins and remission of fear in a group of tarantula enthusiasts. Behaviour Research &  Therapy, 20(5), 437-443. Korchin, S. (1964). Anxiety & cognition. In C. Scheerer (Ed.), Cognition: Theory, research, promise. New Work: Harper & Row. Kovacs, M., Rush, J . , Beck, A., & Hollon, S. (1981). Depressed outpatients treated with cognitive therapy or pharmacotherapy: A one year follow-up. Archives of General  Psychiatry, 38, 33-39. Lader, M., Gelder, M., & Marks, I. (1967). Palmar skin conductance measures as predictors of response to desensitization. Journal of Psychosomatic Research, 11, 283-290. Lader, M., & Mathews, A. ( 1968). A physiological model of phobic anxiety and desensitization. Behaviour Research &  Therapy, 6, 411-421. Lader, M., & Wing, L. (1966). Physiological Measures, Sedative  Drugs, and Morbid Anxiety. London: Oxford University Press. 212 Lande, S.D. (1982). Physiological & subjective measures of anxiety during flooding. Behaviour Research & Therapy, 20 (1), 81-88. Lang, P. (1968). Fear reduction and fear behavior: Problems in treating a construct. In j . Shlien (Ed.). Research in  Psychotherapy Vol. 3. Washington: American Psychological Association. Lang, P. (1971). The application of psychophysiological methods to the study of psychotherapy & behaviour modification. In A. Bergin & S. Garfield (Eds.). Handbook of Psychotherapy & Behaviour Change. New York: John Wiley & Sons, Inc. Lang, P. (1977). Physiological assessment of anxiety & fear. In J . Cone & R. Hawkins (Eds.). Behavioural assessment:  New directions in c l i n i c a l psychology. J . Lone & R. Hawkins (Eds.). New York: Brunner-Mazel Pub. Lang, P. (1978). Self-efficacy theory: Thoughts on cognitions & unification. Advances in Behaviour Research & Therapy, _1_, 187-192. Leitenberg, H., Agras, S., Butz, R., & Wincze, J . (1971). Relationships between heart rate & behavioural change during the treatment of phobias. Journal of Abnormal Psychology, 78(1), 59-68. Leitenberg, H., Agras, S., Edwards, J . , Thomson, L., & Wincze, J . (1970). Practice as a psychotherapeutic variable: An experimental analysis within single cases. Journal of  Psychiatric Research, 1_, 215-225. Lick, J . , & Katkin, E. (1976). Assessment of anxiety & fear. In M. Hersen & A.E. Bellack (Eds.). Behavioral  Assessment: A Practical Handbook. Oxford: Pergamon Press. Likierman, H., & Rachman, S. (1980). Spontaneous decay of compulsive urges: cumulative effects. Behaviour Research &  Therapy, 18, 387-394. Lindsay, P., & Norman, D. (1977). Human Information Processing (2nd ed.). New York: Academic Press. 213 Mandler, G. (1975). Mind and Emotion. New York: John Wiley & Sons Inc. Marks, I., Gray, S., Cohen, D., H i l l , R., Mawson, D., Ramm, E., & Stern R. (1983). Imipramine and brief therapist-aided exposure in agoraphobics having self-exposure homework. Archives of General Psychiatry, 40(2), 153-162. Marks, I., & Huson, J . (1973). Physiological aspects of neutral and phobic imagery: Further observations. British  Journal of Psychiatry, 122, 567-572. Marks, I., Marset, P., Boulougouris, J . , & Huson, J . (1971). Physiological accompaniments of neutral & phobic imagery. Psychological Medicine, J_, 299-307. Marzillier, J . , Carroll, D., & Newland, J . (1979). Self-report & physiological changes accompanying repeated imagining of a phobic scene. Behaviour Research & Therapy, 17, 71-77. Mavissakalin, M. , Michelson, L., Greenwald, D., Kornlith, S., & Greenwald, M. (1983). Cognitive-behavioural treatment of agoraphobia: Paradoxical intention vs. self-statement training. Behaviour Research & Therapy, 2J_( 1 ), 75-86. Meichenbaum, D. (1978). Cognitive Behaviour Modification: An  Integrative Approach. New York: Plexium Press. Meyer, V., & Reich, B. (1978). Anxiety management - the marriage of physiological & cognitive variables. Behaviour  Research & Therapy, 16, 177-182. Morasky, R., Reynolds, C , & Sowell, L. (1983). Generalisation of lowered EMG levels during musical performance following biofeedback training. Biofeedback & Self-Regulation, 8(2), 207-216. Morrison, D. (1976). Multivariate Statistical Methods. New York: McGraw-Hill Book Co. McDonald, R., Sartory, G., Grey, S., Cobb, J., Stern, R., & Marks, I. (1979). The effects of self-exposure instructions on agoraphobic outpatients. Behaviour Research  & Therapy, J_7, 83-85. 214 Neftel, K., Adler, R., Kapelli, L., Rossi, M., Dolder, M., Kaser, H. , Bruggesser, H., & Vorkauf, H. (1982). Stage fright in musicians: A model illustrating the effect of beta blockers. Psychosomatic Mecicine, 44(5), 461-469. Norman, D. (1976). Memory & Attention(2nd ed.). New York: John Wiley & Sons. Norton, R., Dinardo, P., & Barlow, D. (1983). Predicting phobic's response to therapy: a consideration of subjective, physiological and behavioural measures. Canadian Psychology , 24(1 ) , 50-58. Nunes, J . , & Marks, I. (1975) . Feedback of true heart rate during exposure in Vivo. Archives of General Psychiatry, 32 ( 7 ) , 933-936. Ost, L., Jerremalm, A., & Johansson, J . (1981) . Individual response patterns and the effects of different behavioral methods in the treatment of social phobia. Behaviour  Research & Therapy, J_9(1), 1-16. Ost, L., Johansson, J . , & Jerremalm, A. (1982) . Individual reponse patterns and the effects of different behavioural methods in the treatment of claustrophobia. Behaviour  Research & Therapy, 2 £ ( 5 ) , 445-460. Overall, J . , & Spiegel, D. (1969). Concerning least squares analysis of experimental data. Psychological Bulletin, 72 (5), 311-322. Parkinson, L., & Rachman, S. (1980). Are intrusive thoughts subject to habituation? Behaviour Research & Therapy, 18, 409-418. Parkinson, L., & Rachman, S. (1981). Speed of recovery from an uncontrived stress. Advances in Behaviour Research &  Therapy, 3(3), 119-123. Paul, G.L. (1966). Insight vs. desensitization in psychotherapy. California: Stanford University Press. Pearson, E., & Hartley, H. (1972). Biometrika tables for  statisticians. Vol I I . Cambridge: University Press. 215 Philips, H.C. (1985). Return of fear in the treatment of a fear of vomiting. Behaviour Research & Therapy, 2_3(1), 45-52. Popler, K. (1977). Agoraphobia: Indications for the application of the multimodal behavioral conceptulization. The Journal of Nervous & Mental Disease, 164(2), 97-101. Rabavalis, D., Boulougouris, J . , & Stefanis, C. (1976). Treatment of obsessive-compulsive patients. Behaviour  Research & Therapy, .14, 349-355. Rachman, S.J. (1971). Obsessional ruminations. Behaviour  Research & Therapy, 9, 229-235. Rachman, S.J. (1976). The passing of the two-stage theory of fear & avoidance: Fresh p o s s i b i l i t i e s . Behaviour Research &  Therapy, 14, 125-131. Rachman, S.J. (1978). Fear & Courage. San Francisco: W. Freeman & Co. Rachman, S.J. (1979). The return of fear. Behaviour Research  & Therapy, 17, 164-166. Rachman, S.J. (1980). Emotional Processing. Behaviour  Research & Therapy, 18, 51-60. Rachman, S.J. (1981). Part 1. Unwanted intrusive cognitions. Advances in Behaviour Research and Therapy, 2(3), 89-99. Rachman, S.J. (1983). Agoraphobic study: the return of fear: unpublished manuscript. Rachman, S.J., & de Silva, P. (1978). Abnormal and normal obsessions. Behaviour Research & Therapy, 16, 233-248. Rachman, S.J., & Hodgson, R. (1974). Synchrony and desynchrony in fear and avoidance. Behaviour Research & Therapy, 12, 311-318. Rachman, S.J., & Levitt, K. (in press). Panics and their 216 consequences. Rimm, D., Janda, L., Lancaster, D., Nohl, M., and Dittmar, K. (1977). An exploratory investigation of the origin and maintenance of phobias. Behaviour Research & Therapy, 15, 231-238. Robinson, J . (1980). Affect and retrieval of personal memories. Motivation and Emotion, 4(2), 149-174. Sarason, I. (1978). The test anxiety scale: concept & research. In C. Spielberger & I. Sarason (Eds.), Stress &  Anxiety: Vol. 5. Washington: Hemisphere Pub. Corp. Sartory, G., Rachman, S., & Grey, S. (1977). An investigation of the relation between reported fear & heart rate. Behaviour Research & Therapy, 15, 435-438. Sartory, G., Rachman, S., & Grey, S. (1982). Return of fear: the role of rehearsal. Behaviour Research & Therapy, 20, 123-133. Smith, B., Dickel, M., & Deutsch, S. (1978). Overextinction and test stimulus modality determinants of dishabituation. Psychophysiology, j_5(4), 324-329. Spielberger, C , Gorsuch, R., & Lushene, R. (1970). State- Trait Anxiety Inventory Manual. California: Consulting Psycholigists Press Inc. Stern, R., & Marks, I. (1973). Brief and prolonged flooding: a comparison in agoraphobic patients. Archives of General  Psychiatry, 28, 270-276. Sweeney, G., & Horan, J . (1982). Separate and combined effects of cue controlled relaxation and cognitive restructuring in the treatment of musical performance anxiety. Journal of  Counseling Psychology, 2_9(5), 486-497. Thompson, R., & Spencer, W. (1966). Habituation: a model phenomenon for the study of neuronal substrates of behaviour. Psychological Review, 73(1)., 16-43. 217 Thyer, B., Papodorf, J., Himle, D., McCann, B., Caldwell, S., & Wickpert, M. (1981). Invivo distraction-coping in the treatment of test anxiety. Journal of Clinical Psychology, 37(4), 754-764. Tversky, A., & Kahneman, D. (1982). Judgement under uncertainty: heuristics and biases. In D. Kahneman, P. Slovic & A. Tversky (Eds.), Judement Under Uncertainty:  Heuristics and Biases. Cambridge: Cambridge University Press. Tyron, W. (1982). Reinforcement history as possible basis for the relationship between self-percepts of efficacy and responses to treatment. Journal of Behaviour Therapy and  Experimental Psychiatry, 13(3), 201-202. Watson, J., Gaind, R., & Marks, I. (1971). Prolonged exposure: a rapid treatment for phobias. British Medical Journal, _1_, 13-15. Watson, J . , Gaind, R., & Marks, I. (1972). Physiological habituation to continuous phobic stimulation. Behaviour  Research & Therapy, 10, 269-278. Watts, F. (1974). The control of spontaneous recovery of anxiety in imaginal desensitization. Behaviour Research &  Therapy, 12, 57-59. Whitlow, J . (1975). Short term memory in habituation and dishabituation. Journal of Experimental Psychology (Animal  Behavior Processes), 104(3), 189-206. Wine, J . (1971). Test anxiety and direction of attention. Psychological Bulletin, 76(2), 92-103. Winer, B.J. (1971). St a t i s t i c a l Principles in Experimental  Design. Sydney: McGraw-Hill Kogakusha Ltd. Wolpe, J . (1978). Self efficacy theory & psychotherapeutic change: a square peg for a round hole. Advances in  Behaviour Research & Therapy, j_, 231-236. 218 FOOTNOTES 1. Craske, M., Rachman, S.J., & Lopatka, C. (1985). [An examination of the relationship between exposure duration and return of fear in spider phobics]. Unpublished raw data. 2. Rachman, S.J., & Craske, M. (1984). [An assessment of belief strength, habituation, and return of fear in an obsessive-compulsive patient]. Unpublished raw data. 219 Appendix 1 NAME: PHONE NUMBER: Do you perform better in private than when you perform in public? yes no If yes, does this happen even when you are thoroughly prepared to do the best of your ability? yes • no Has this problem affected your career in the f i e l d of music? yes no In what ways: Low grades Poor teacher reviews Courses failed Poor proficiency exams Proficiency exams failed Poor recital reviews Unsuccessful auditions for orchestra seats Unsuccessful auditions for jobs Unsuccessful performances in competitions Is there any difference in your performance when you perform in important versus nonimportant recitals? yes (I perform worse in important recitals) no Have you been told that your a b i l i t y as a performer is greater than your public performances show? yes no Are you interested in acquiring strategies for dealing with performance anxiety? yes no If you are, please be sure to write your name and phone number on the top of this questionnaire. Thankyou 220 Appendix 2 Subject Data Years Years in a Group 1 of Perform- Screening Exper- ance Question-Instrument Age Sex ience Grade Degree na: 1 piano 17 F 8 8 0 9 2 piano 24 F 18 8 0 1 1 3 piano 23 M 17 10 0 9 4 vi o l i n 19 M 14 8 1 1 1 5 vio l i n 26 F 15 7 4 12 6 piano 40 F 8 10 0 9 7 piano 24 F 12 ARCT 0 10 8 piano 37 F 15 ARCT 4 10 9 piano 31 F 10 10 0 1 0 10 piano 20 F 10 10 0 9 1 1 piano 53 F 30 10 1 9 12 piano 30 M 26 10 0 9 1 3 piano 27 F 6 8 0 9 1 4 piano 37 F 5 10 0 9 15 piano 22 M 12 10 2 10 16 piano 1 5 F 9 10 0 9 roup 2 1 vio l i n 22 F 1 1 10 3 1 1 2 piano 20 F 14 ARCT 0 1 1 3 piano 22 F 16 ARCT 3 9 4 piano 15 F 9 10 0 1 1 5 piano 29 M 8 ARCT 1 10 6 vi o l i n 24 F 18 9 4 9 7 piano 20 F 12 ARCT 2 9 8 vio l i n 23 F 6 - 4 10 9 piano 23 F 7 8 3 9 10 piano 18 F 12 ARCT 1 10 1 1 voice 23 F 5 - 3 10 12 v i o l i n 24 M 3 7 2 9 13 piano 19 F 10 9 2 10 1 4 v i o l i n 19 F 4 8 1 9 15 vi o l i n 20 M 8 ARCT 1 9 16 v i o l i n 23 F 6 8 4 10 221 Years Years in a Group 3 , of Perform- Screening Exper- ance Question-Instrument Age Sex ience Grade Degree naire 1 violin 27 M 19 9 6 1 1 2 violin 19 F 1 5 10 2 1 1 3 violin 20 F 1 7 10 3 9 4 piano 18 F 1 4 ARCT 1 9 5 piano 19 F 10 ARCT 2 9 6 piano 23 F 17 7 5 9 7 piano 38 F 13 9 0 9 8 piano 21 F 10 10 0 1 2 9 voice 23 F 5 - 1 9 10 piano 40 F 5 10 2 9 1 1 piano 21 F 9 8 0 9 1 2 piano 21 F 15 10 0 10 1 3 voice 21 F 3 - 3 9 1 4 viol in 47 F 40 ARCT 4 9 1 5 piano 24 F 8 8 1 9 roup 4 1 violin 22 M 12 ARCT 2 10 2 piano 21 F 16 9 3 9 3 violin 20 M 10 9 3 9 4 violin 32 M 10 ARCT 3 9 5 piano 17 F 1 1 10 1 9 6 piano 19 F 12 10 2 9 7 piano 19 F 13 ARCT •1 9 8 piano 20 F 15 1 0 3 9 9 piano 18 F 13 ARCT 1 9 10 voice 23 F 20 - 2 12 1 1 piano 1 5 F 9 ARCT 0 10 1 2 piano 33 F 23 ARCT 4 10 1 3 voice 28 F 7 - 1 10 1 4 piano 15 F 10 ARCT 0 9 1 5 piano 25 M 1 5 10 3 9 16 piano 25 F 1 1 10 2 9 222 Appendix 3 TRAINING PROGRAM MANUAL Session 1 1. attendance. 2. introduction of experimenter. 3. introduction of group members. 4. description of anxiety in performance: discussion of Yerkes-Dodson Law and the e f f e c t s of anxiety (beyond the optimal level) upon a b i l i t y to attend and bodily sensations and their effect on technical and musical a b i l i t y . 5. exploration of history and current status of symptoms (a) how long have you experienced performance anxiety? (b) how much does your anxiety interfere with your performing? (sometimes/every time, able to continue/have to stop) (c) what situations are most anxiety arousing for you? (d) how do you know when you are anxious? If I were l i s t e n i n g to you how would I know i f you were anxious? 6. discuss ways in which performance anxiety can develop Performance anxiety can develop for a number of d i f f e r e n t reasons (a) you may have had one or more bad experiences performing. Perhaps you had a memory lapse in a r e c i t a l and were t e r r i b l y embarrassed, or perhaps you were extremely nervous in an exam and didn't play nearly as well as you could at home. Perhaps you were playing for some friends and f e l t they were bored by the piece you had chosen to play. Negative experiences l i k e these tend to make us less eager to perform and much more nervous and less self-confident when we do perform. These experiences build up negative associations with performing instead of positive ones, so that when you next perform your ;are more l i k e l y to respond negatively and with anxiety-. Have any of you had experiences l i k e these? (b) performance anxiety may also develop after seeing other people perform who are very anxious and make: a lot of mistakes. (c) your anxieties about performing may also be a result of anxieties which you have in other areas of your l i f e . If you have a poor self-image i.e., a low opinion of yourself in general, then you may expect that you won't do well, that y o u ' l l probably have a memory lapse, or that you won't be able to control: your anxiety. (d) do any of you have additional ideas as to how your anxiety about performing may have developed? 223 In any case, i t seems that many performers learn to be anxious while performing as a result of experiences. Therefore, you are able to relearn how not to be anxious. Remember that the goal is not to eliminate feelings of * arousal and alertness a l l together, but to remove the anxiety that interferes with your performance. presentation of treatment rationale In order to "unlearn" the anxiety, over the next four weeks I am going to be teaching you two very effective strategies and help you to practise those strategies while performing in order to reduce your current discomfort. These two strategies are relaxation training, and task focusing training. Each deals with a different aspect of performance anxiety; relaxation training deals with bodily tension that interferes with smooth performance, and task focusing deals with thoughts that distract your attention from the task of performing. Both of these strategies have shown to be very effective in reducing performance anxiety. Today I will be teaching you the method of progressive muscle relaxation- a method of inducing relaxation that can be learned very quickly and that allows you to become more deeply relaxed than ever before. It is effective because the muscles in your body cannot be both tense and relaxed at the same time. So, once you have learned the relaxation technique, i t can be used to counter anxiety, tenseness and other feelings like those you experience when performing. training in progressive muscle relaxation I n i t i a l l y , this technique will take you 15 to 25 minutes, but as you become more sk i l l e d , the time taken to induce relaxation wi l l shorten to a few minutes. Then i t becomes a s k i l l that can be readily applied "on the spot"- just prior to performing. This method involves systematic tensing and relaxing of the different muscle groups in order to heighten your awareness of the feeling associated with each state (of relaxation and of tension). Such heightened awareness then allows you to recognise tension in your body, and to replace i t with relaxation. There are certain preconditions for practising relaxation that you should ensure before you begin. You require either a comfortable chair or bed that supports a l l areas of your body, including your neck, and a quiet room where you know you will not be distracted. You should also loosen any tight clothing such as belts and shoes. Finally, practise at a time that you have set aside particularly for relaxation so that you can devote a l l of your attention to the exercises. Thinking about other things that you should be doing is not conducive to concentrating on feelings of tension and relaxation. It is important to realise that relaxation is a s k i l l and 224 therefore requires practise in order to feel its benefits. Practise of these exercises is essential at least once per day, and twice is preferrable. You should become quite skilled at this technique after one week of practise, so that next session we can begin applying it directly to the performance situation. To begin, I want you to l i e on your back on the floor, and loosen any tight clothing etc. We will work systematically through the major muscle groups by f i r s t tensing the muscles until I say "relax" when you let go of the tension immediately. If the muscles are f i r s t tensed, they will relax more deeply when they are released. Focus your attention on each muscle group as we work through each one. Feel the different sensations. If you have other thoughts coming into, your mind, just let them pass through, and refocus on the sensations. Now, close your eyes and listen to my instructions. Begin by taking three deep breaths - slowly inhale, hold (few seconds) and slowly exhale (repeat 3 times). Make fi s t s with your hands. Make a f i s t and tense the muscles of your hands and forearms; tense until they tremble. Feel the muscles pull across your fingers and lower part of your forearm. (5 to 7 seconds). Now relax, and let your hands open. Pay attention to the muscles of your hands and forearms as they relax. Note how they feel as the relaxation flows through them. (10 to 20 seconds). Again, tense the muscles of your hands and forearms. Pay attention to the muscles involved. (5 to 7 seconds). Now relax and attend only to those muscles, and note how they feel as the relaxation takes place in contrast to the previous feeling of tension. Each time you do this you will'relax even more until your arms and hands are completely relaxed, feeling warm and relaxed and heavy. Proceed to the other gross muscle groups (listed below) in the same manner, in an "hypnotic monotone", repeating the same verbalisations such as "Note how the muscles feel as they relax; feel the relaxation and warmth flow through these muscles; pay attention to these muscles so that later you can relax them again", and emphasising the sensations of warmth, relaxation and heaviness, biceps (imagine l i f t i n g weights) shoulders & neck ( l i f t shoulders towards ears, and then droop) forehead (frown hard or l i f t eyebrows tensing muscles 225 of the forehead and top of head) eyes (squeeze eyes tightly) jaw, mouth and throat (draw back.corners of mouth or clench teeth) chest (take in deep breath & hold for several seconds) back (arch ) abdomen (prepare stomach for a punch) upper legs ( l i f t legs a few inches off floor) calves (push toes away) Now, as I briefly mention each muscle group again, just check that they are s t i l l relaxed - i f they are not relaxed, tense and release those muscles again: hands and forearms, upper arms, shoulders and neck, forehead, eyes, mouth/throat and jaw, chest, back, stomach, upper legs, and calves. Now as you spend a couple of minutes in this relaxed state, I want you to become even more relaxed; let every last bit of tension leave your body.as you become more and more relaxed. Your whole body may feel heavy and warm (continue these suggestions for one minute). Open your eyes and slowly sit up. Never stand up quickly after doing these exercises because you may feel dizzy. 9. home assignment Before next week's session, practise this exercise at least once, and preferrably twice per day so that we can move on to the next stage. The exercises are summarised on this handout (distribute Progressive Muscle Relaxation training sheets). Don't practise for more than 15 minutes at any one time. Each time you practise i t is useful for your own interest to keep a record on these sheets (hand out Relaxation Practise sheets) of the date, the time, and also rate your level of discomfort/tension before you begin and immediately after completing the exercises. Use a scale from 0 to 10 where 0 = complete relaxation and 10 = extreme tension. 10. Remind subjects that they will be performing next session and to bring instruments (violins) i f necessary. 226 Session 2 1. attendance and complete Practise Information questionnaire. 2. brief restatement of rationale. 3. discussion of home assignment (a) how many times were you able to practise? (b) did your ratings of tension change (from before to after, and across days)? 4. progressive muscle relaxtion practise (a) 15 minutes of relaxation instructions. (b) check each group member's level of relaxation by l i f t i n g an arm and observing the way i t drops to the floor. 5. modelling and training in abbreviated relaxation As I mentioned last time, the goal of the tense and release exercises is to enable you to become skilled at inducing relaxation in a very short time, so that your s k i l l can then be applied directly to the performance situation. The "quick" method leaves out the tensing and involves very intense concentration to induce relaxation in the different muscle groups. Therapist demonstrates correct sitting for relaxation (legs uncrossed, arms resting on thighs, and neck resting forward). While you are seated in your chair, I want you to close your eyes and follow my instructions. Begin by taking three deep breaths - showly inhale, hold (few seconds) and slowly exhale (3 times). Now, focus on each muscle group as I say them, and recall the feelings of relaxation that you have learned from the tension-release method. Concentrate f i r s t on your hands and forearms becoming relaxed; feel the muscles relaxing and feel the warmth as relaxation spreads though your hands and forearms. (15-20 seconds). Now move up to your upper arms and concentrate on relaxing those muscles, feel the relaxation and warmth flowing through these muscles, so that now your whole arms are feeling completely relaxed and free from tension. They are just being supported by resting on your legs as there is no muscular tension to support your arms. Proceed to the other gross muscle groups (listed below) in the same manner, in an "hypnotic monotone", repeating the same verbalisations such as "Concentrate on the muscles becoming relaxed, feel the relaxation and warmth spread through these muscles", and emphasising the sensations of warmth, heaviness and relaxation. The whole procedure should take 3 to 4 minutes, shoulders (droop) 2 2 7 forehead (smooth out muscles) eyes mouth, jaw and throat (droop head forward) chest and back (sink into chair to prevent muscular support) abdomen ( l i k e a "big blob of j e l l y " ) legs ("supported only by the chair") calves ("supported only by the floor") If you find i t d i f f i c u l t to relax p a r t i c u l a r muscles, just tense them for a few seconds and then relax. With continued practise at this "quick" form of relaxation, you w i l l be able to relax more and more eas i l y and much more quick l y . 6. imaginal rehearsal This time I want you to imagine yourself going to perform for a group of f r i e n d s . You are s i t t i n g and waiting for a few minutes before performing, so you use that time to relax and to prepare yourself. You begin by taking three deep breaths, and then start focusing on each muscle group and induce r e l a x a t i o n . Concentrate f i r s t on your hands and forearms, saying words to yourself l i k e "relax, warm and heavy", and induce relaxation in those muscles. Then move to your upper arms, following the same procedure, feel i n g the warmth and sensations of relaxation spreading into those muscles. Repeat words to yourself l i k e "relax, warm and heavy", concentrating on relaxing those muscles. Now f e e l the relaxation spreading up as you concentrate on relaxing your shoulders, l e t t i n g them droop as a l l the tension disappears and as the relaxation spreads further and further. Concentrate on your f a c i a l muscles, relaxing f i r s t your forehead, and then your eyes, and then your mouth, jaw and throat - again repeating to yourself "relax, warm and heavy" and r e a l l y concentrate on those muscles becoming relaxed. Let the relaxation spread further and l e t your head droop forward so that i t rests on your chest. Now concentrate on relaxing your chest and back so that you f e e l as i f you sink into the chair as more and more of you becomes relaxed. Now allow your stomach muscles to relax as you concentrate on them and l e t the tension go. Move down to your legs and concentrate on relaxing f i r s t the thigh muscles and then the c a l f muscles, so that your whole legs are completely relaxed and just being supported by the chair and fl o o r rather than by muscular tension. So that now you are fe e l i n g completely relaxed and you are resting in a drooped p o s i t i o n . Now, i t i s your turn to perform so you slowly stand and move towards the front. 228 Remember, this "quick" relaxation method can be carried out even when seated with your instrument in front of you. This w i l l be easy to do as you become more sk i l l e d . By repeating the words "relax, heavy and warm" they will become associated with the state of relaxation and eventually will act as cues that have power to e l i c i t the relaxation. Now, we wi l l practice this s k i l l in actual performance. presentation of treatment rationale We combine the strategies you learn with actual performance for two reasons. 1. So that your are more easily able to apply your strategies in anxious situations, such as real performances - in this way, the strategies become a natural part of your whole approach; they become automatic and like "second nature". 2. rehearsal i t s e l f , as you are probably very well aware, is of great benefit in learning to remove maladaptive or interfering anxiety. A number of well-known teachers and musicians - Arthur Rubinstein for one - have recommended that simply performing in front of others is the key to reducing performance anxiety. They advise you to make yourself do the very thing your fear - play or sing in public as much as possible, even though at f i r s t i t may be d i f f i c u l t because of your anxiety. Each time i t wil l get a l i t t l e easier. Does this make sense to you? During the next sessions, I am going to ask each of you to play or sing a piece of music in front of the rest of us. A l l of you share the same concerns about performing so together we form a sympathetic and understanding audience. Don't worry i f you don't perform your best at f i r s t . By continuing to perform in a comfortable atmosphere, without criticism from those listening to you, you will realize that performing is not such a frightful experience after a l l and your anxiety will lessen, and by implementing the relaxation strategy before playing, you wil l lessen your tension and feel much more relaxed when performing. With repeated practice of the anxiety reducing strategies, you will gradually feel more at ease and wil l concentrate more on the music. Next week we will focus more directly on concentrating on the music. behavioural rehearsal Now I'd like each of you to play or sing for the rest of us. Begin by spending several minutes inducing relaxation while in your chair, and then move up to the front to do your piece. We'll also be obtaining various measures before and after you perform so that over the next few weeks we can observe changes. Everybody can practice relaxing while (RA) helps ( f i r s t group member to perform) with the measures. Each subject performs in the following sequence: 229 (a) heart rate - ask the subject to sit quietly for two minutes, and then record heart rate for two minutes. (b) Perceived S k i l l questionnaire (c) pre-performance Thought questionnaire (d) SUDS The subject is instructed to induce relaxation and then perform. (Experimenter completes answers to memory questionnaire) (e) SUDS (f) post-performance Thought questionnaire Subject is questionned about amount of tension during performing, i f able to relax before performing, etc. 9. home assignment Between now and the next session, I'd like you to practise the "quick" method of relaxation at least twice a day, and before any time you practise your music so that i t becomes something you do automatically. These sheets (hand out Abbreviated Relaxation points) summarise the major aspects of this method of relaxation. In addition, I'd like you to try to perform a piece of music for an audience of at least 2 family members or friends, using the relaxation strategy and recording your levels of tension/discomfort before and after on these sheets (hand out Relaxation Practice sheets). 10. measure tonic heart rate. 11. remind subjects to bring their musical instruments (violin) to the next session. 230 Session 3 1. attendance and complete Memory questionnaire and Practise Information Questionnaire. 2. brief restatement of rationale. 3. discussion of home assignment (a) are you able to relax in a relatively short period of time? (b) when you performed, how did your ratings of discomfort compare to your performance in the session last week? 5. presentation of treatment rationale During this session we are going to begin a new strategy that deals with task irrelevant or task distracting thinking. Often, i t seems that people with performance anxieties frequently have thoughts such as: I wonder what people wil l think of me i f I make mistakes I'm sure I'm playing worse than anyone else I'm a failure Do you have thoughts similar to these before or during playing? Thoughts like these not only distract you from playing but also elevate your anxiety. Relaxation helps to reduce the frequency of such thoughts, but a direct method of counterattack is important to deal with such anxiety-e l i c i t i n g and distracting thoughts when they arise. Your thoughts, or the things you are saying to yourselves in your heads, seem to get in the way of what you are there to do - which is of course.to play the music as beautifully as you can. Is that right? I'm going to c a l l these thoughts or things we say to ourselves in our heads, our self-statements. We know that there is a strong relationship between what people are thinking and how anxious they f e e l . That i s , the anxiety you experience when performing is not only related to physical tension, but also to the kinds of thoughts you have. If you expect to do poorly e.g., "I know I ' l l have a memory lapse" or i f you start to wonder about what people are thinking of you e.g., "They'll think I'm terrible" or even if your mind just wanders, then your attention is taken away from the music and focused onto something else. These self-statements then raise your anxiety and effect the way you play because you are not concentrating as well as you should. First of a l l you should become aware of the anxiety-e l i c i t i n g and distracting thoughts you are saying to yourselves. That awareness wil l then enable you to challenge such thoughts by replacing them with more constructive thoughts that act to reduce anxiety, and by instructing yourself to focus on the task. 231 In short, you will learn how to gain control over your thinking, and not allow i t to take over and lead you to a state of anxiety. Task focusing and anxiety reducing self-statements can be of two major types: (1) they can be comforting and rewarding self-statements. Comforting self-statements are most useful just before performing; e.g., "I'm well prepared for this performance so there's no need for me to be worried."; e.g., "I'm feeling anxious - that is natural. Let me just breath deeply and relax". More examples? Rewarding self-statements are most useful just after performing; e.g., "I feel pleased with the way I played - and I'm glad I was able to control my anxiety." More examples? (2) or, they can be technique oriented and task focused self-statements, which are most useful during performing. e.g., "Now what did my teacher t e l l me to do at this point? Oh yes, I'm supposed to play very softly and gradually increase the tone" or "I'm supposed to play from the shoulders here." When your attention wanders onto things other than the current piece of music, task focused self-statements simply instruct you to stop the irrelevant thoughts and concentrate on performing; e.g., "Dont worry about the next part, concentrate on what I'm doing right now". More examples? Over time you wi l l gradually find that you have more comforting and task oriented self-statements and fewer distracting and anxiety producing ones. However, don't expect to always have only constructive thoughts. Even mildly anxious people don't think only adaptive thoughts. The difference, however, is that they are able to replace the destructive self-statements with more constructive self- statements. 6. audiovisual demonstrations and training in task focusing Present audio visual sequences of musicians stating aloud task irrelevant and anxiety e l i c i t i n g thoughs, replacing them with task focusing and anxiety reducing thoughts before, during and after performing. The f i r s t sequence demonstrates the situation and thoughts associated with performing.for a group of friends, and the second relates to a music festival.. Each is followed by a discussion of the thoughts: Video sequence 1 - performing for friends 232 The performer in the following slides i s . . She/he was together with some friends who asked her/him to play. I'm sure most of you have been in that same situation. has gone back to relive that experience as i t happened and share with us her/his thoughts and feelings. Note the kinds of anxieties which she/he has and the way she/he copes with them. Pay particular attention to the negative things she/he says to herself and the way in which she/he challenges them and is able to replace them with constructive self-statements. A problem that I think many fear when performing is demonstrated here - the memory lapse. What do you do when you have a memory lapse? Are you able to get back on track again, forget about i t and concentrate on doing the best you can for the rest of the piece? Or do you let i t devastate you and make you feel like a failure and make you perform poorly for the rest of the piece? How did cope with it? What constructive things can you say to yourself to help you cope with a memory lapse? Do you remember what she said to herself after she finished performing? Why is that important? What could you say i f you performed well? - or if you didn't perform as well as you would have liked? Video sequence 2 - playing in a music festival The pianist in the following slides is Denella. Denella has gone back to relive the experience of competing for a scholarship in a music f e s t i v a l . I imagine that a lot of you have participated in some of the music festivals throughout the c i t y . Denella wi l l share with us her thoughts and feelings when she performed. She is playing the same peice of music as she did in the festival - two Spanish Dances from a suite by Ginastera. Note the kinds of anxieties which she has and the way in which she copes with them. Again, notice the negative things she says to herself and the way she challenges them, and replaces them with constructive task oriented statements. What kinds of anxieties are special to music festival situations? (competition, audience) What can you say to yourself to cope with that situation? Depending on the group composition, one sequence demonstrated a pianist and the other either a pianist (if a l l group members were pianists), v i o l i n i s t (if some group members were v i o l i n i s t s ) , or vocalist (if some group members were vocalists): no single group contained a l l three types of performers. For this sort of strategy to be most effective, you must choose several constructive thoughts that are specific for 233 you. First I want you each to t e l l me the two to three major anxiety producing or task distracting thoughts that you have. If you find this d i f f i c u l t , recall your most recent performance and the thoughts you had then. Then, I want you to select two to three self- statements to counter those; at least one of these should be task focusing, and one should be comforting or rewarding. E l i c i t the maladaptive and adaptive thoughts from each group member, ensuring that task focusing self-statements are emphasised, while comforting/rewarding self statements are more specific to particular situations. So, whenever you feel yourself becoming anxious about a performance, that is a sign to you to not only relax and reduce your level of physical tension, but also to examine your thoughts, replace specific anxiety e l i c i t i n g thoughts with comforting thoughts, and to instruct yourself to focus on the task i.e., to think about your playing or singing and to ignore a l l of the irrelevant thoughts. behavioural rehearsal Now, I'd like you to combine your strategies in actual performance by playing or singing for the rest of us. Begin by relaxing, and then perform a piece. It's kind of d i f f i c u l t to perform and talk at the same time so save those thoughts and share them with us afterwards. Again, like last week, we wi l l take various measures before and after. Each group member performs in the following sequence: (a) heart rate - ask the subject to s i t quietly for two minutes, and then record heart rate for two minutes. (b) Perceived S k i l l questionnaire (c) pre-performance Thought questionnaire (d) SUDS The subject is instructed to induce relaxation and then to perform. (Experimenter completes Answers to Memory Questionnaire) (e) SUDS (f) post-performance Thought questionnaire The subject is asked to describe their thoughts and ways they did or could have challenged them. home assignment Between now and next time I'd like you to practise combining the relaxation and task focusing self-statements whenever you play. (Attention Focusing training sheet are handed out). 234 Also, perform a piece for family members or friends but this time increase the number of people in your audience from last week., Make i t like a real performance. Again, use the relaxation strategy before you play or sing and rate levels of discomfort on these sheets (Relaxation Practice sheets) before and after. Also, l i s t the major distracting thoughts and major constructive thoughts that you used or could have used. 10. measure tonic heart rate. 11. remind subjects to bring their musical instruments (violin) to the next session. 235 Session 4 1. attendance and complete Memory questionnaire and Practise Information questionnaire. 2. b r i e f restatement of r a t i o n a l e . 3. discussion of home assignment Tai are you able to successfully replace task irrelevant thoughts? (b) when you performed before others, how did your ratings of discomfort compare to your performance the previous week? 5. video modelling Presentation of audiovisual sequences of musicians in exam and r e c i t a l situations followed by discussion of the thoughts demonstrated: Video sequence 3 - performing in an exam The performer in the following s l i d e s i s and she/he is r e l i v i n g the experience of performing in an exam at u n i v e r s i t y . I think most of you have taken Toronto Conservatory or Western Board exams. Note the kinds of anxieties and the way she/he copes with them. Pay pa r t i c u l a r attention to the negative things he/she . says to him/herself and the way in which she/he challenges them and replaces them with more constructive s e l f -statements. Video sequence 4 - playing in a r e c i t a l The pianist in the following s l i d e s i s Vance who i s a teacher and a concert p i a n i s t . He is going to share some of the thoughts and feelings which he has while giving a r e c i t a l because often we think that people who reach that l e v e l never have performance anxiety at a l l . This i s not the case. Notice, however, that Vance's anxieties are not focused on himself and what people w i l l think of him, but whether he w i l l be able to play th music in a way that w i l l communicate something to the audience. As I mentioned e a r l i e r , a certain amount of anxiety i s necessary to give excitement and l i f e to the performance. If there is no anxiety at a l l , i t would be d u l l and l i f e l e s s . I think Vance has just the right amount of performance anxiety. What kinds of anxieties are special to r e c i t a l situations? 6. imaginal rehearsal I'm going to describe a scene to you. Each time I come to a destructive self-statement I w i l l pause and I'd l i k e you to challenge i t and replace i t with a coping s e l f -statement. Now close your eyes and relax. Imagine that you are p a r t i c i p a t i n g in a music f e s t i v a l . You a r r i v e at the h a l l where you're going to play or sing and while standing at the door you can hear someone 236 performing inside. You show your competitor's card to the woman at the desk and s l i p inside as the audience is applauding the performer who is just leaving the stage. You walk slowly down to the front and take a seat right at the a i s l e . You notice from your program that there is one more person to perform before i t ' s your turn. The next performer is walking up to the stage. She prepares herself and waits for a signal from the adjudicator to begin. She starts. It's a piece which is more d i f f i c u l t that the one you have chose to do. And she does i t beautifully - her technique is flawless and she performs musically too. You are wondering how you will measure up to her - you are thinking you couldn't possibly beat her but then you say to yourself . She is finished and returns to the audience. You get up and ascend the five stairs to the stage. The light is so bright and everything feels strange. You look out towards the audience. It's f a i r l y dark, you can't see much, but towards the back of the hall you see the adjudicator sitting at a long table. He seems to be reading something - then he looks up. You hear the bell and know you must begin. You wonder if you can remember the f i r s t note but you say to yourself . You begin. You're feeling kind of shakey and are just hoping you'll get to the end without messing i t up but then you say to yourself . You're remembering the louds and softs, the parts to play staccato parts and the smooth parts, and you're concentrating on the beautiful phrasing. You're doing so well you wonder if you can keep it up but you say to yourself . You find you are actually enjoying your performance. You come to the closing notes. You f i n i s h . The audience is applauding. You feel good inside and say to yourself . 7. behavioural rehearsal Today I'd like each of you to play or sing a piece of music only this time imagine yourself in an exam or festival or r e c i t a l . Again try to focus on task oriented thoughts. If you have distracting thoughts challenge and replace them with task focused ones. After you fi n i s h , share with us how successful you've been at controlling your thoughts. Each subject performs once in the following sequence: (a) heart rate - the subject is instructed to sit quietly for two minutes, and then heart rate is recorded for two minutes. (b) Perceived S k i l l questionnaire (c) pre-performance Thought questionnaire (d) SUDS The subject is instructed to induce relaxation and to perform (Experimenter completes answers to Memory questionnaire). 237 (e) SUDS (f) post-performance Thought questionnaire 8. home assignment Before now and the post assessment, I'd like you to again perform a piece before an audience, but this time increase the number of people. Use the relaxation before hand and silent control of your thoughts. 9. measure tonic heart rate. 10. arrange post assessment. Appendix 4 238 TRAINING PROGRAM OUTLINE FOR RESEARCH ASSISTANTS  Session 2 1. check attendance of group members on attendance sheet 2. handout Practise Information questionnaires to be completed 3. relaxation exercises (a) ensure a l l members lying in correct position (legs uncrossed, arms by sides) (b) check relaxation levels by gently l i f t i n g each person's arm by the wrist about six inches of the floor and letting i t drop. 4. behavioural rehearsal (a) measure pulse rate 1. turn on pulse metre 2. attach sensor to earlobe - light at the back 3. ask subject to sit quietly for a few minutes 4. ensure "E" signal is off and heart signal is on 5. wait for 2 minutes (use timer on d i g i t a l display - top left) 6. record the number every 10 seconds for two minutes (i. e . 12 readings) and write heart rate on the recording sheet under the performance section. 7. remove sensor and turn pulse metre off (b) ask subject to complete questionnaires in following order: Perceived S k i l l , Pre-performance Thought Questionnaire, and SUDS. (c) instruct subject to relax for a few minutes and then perform (d) ask subject to complete questionnaires in the follow-ing order: SUDS and Post-performance Thought Quest-ionnaire. 5. hand out Quick Relaxation Outline and practise sheets 6. measure tonic pulse rate (same as 4. (a)) 1. turn on pulse metre 2. attach sensor to earlobe - light at the back 3. ask subject to s i t quietly for a few minutes 4. ensure "E" signal is off and heart signal is on 5. wait for 2 minutes (use timer on d i g i t a l display - top left) 6. record the number every 10 seconds for two minutes (i. e . 12 readings) and write heart rate on the recording sheet under the performance section. remove sensors and turn pulse metre off 240 Session 3 1. check attendance of group members on attendance sheet 2. hand out Practise Information and Memory Questionnaires to be completed 3. control the caramate projector - audio visual sequence (a) plug in (b) insert audiotape and ensure rewound to start (c) ensure slides begin with number 1 (d) turn power switch on (e) turn cassette tape to play (f) rewind tape at end of f i r s t sequence & turn power off (g) change cassette tape & ensure rewound to start (h) turn power on for next sequence following discussion (i) rewind tape and turn power off 4. behavioural rehearsal (a) measure pulse rate 1. turn on pulse metre 2. attach sensor to earlobe - light at the back 3. ask subject to sit quietly for a few minutes 4. ensure "E" signal is off and heart signal is on 5. wait for 2 minutes (use timer on digital display - top left) 6. record the number every 10 seconds for two minutes ( i . e . 12 readings) and write heart rate on the recording sheet under the performance section. 7. remove sensors and turn pulse metre off (b) ask subject to complete questionnaires in following order: Perceived S k i l l , Pre-performance Thought Questionnaire, and SUDS. (c) instruct subject to relax for a few minutes and then perform (d) ask subject to complete questionnaires in the follow-ing order: SUDS and Post-performance Thought Quest-ionnaire. 5. hand out Attention-Focusing Outline and Practise sheets 6. measure tonic pulse rate (same as 4. (a)) 1. turn on pulse metre 2. attach sensor to earlobe - light at the back 3. ask subject to sit quietly for a few minutes 4. ensure "E" signal is off and heart signal is on 5. wait for 2 minutes (use timer on digital display - top' left) 6. record the number every 10 seconds for two minutes ( i . e . 12 readings) and write heart rate on the recording sheet under the performance section. r e m o v e s e n s o r s a n d t u r n p u l s e m e t r e o f f 242 Session 4 1. check attendance of group members on attendance sheet 2. hand'out Practise Information and Memory Questionnaires to be completed 3. control the caramate projector - audio visual sequence (a (b (c (d (e (f (g (h (i plug in insert audiotape and ensure rewound to start ensure slides begin with number 1 turn power switch on turn cassette tape to play rewind tape at end of f i r s t sequence & turn power off change cassette tape & ensure rewound to start turn power on for next sequence following discussion rewind tape and turn power off 4. behavioural rehearsal (a) measure pulse rate 1. turn on pulse metre 2. attach sensor to earlobe - light at the back 3. ask subject to s i t quietly for a few minutes 4. ensure "E" signal is off and heart signal is on 5. wait for; 2 minutes (use timer on dig i t a l display - top left) 6. record the number every 10 seconds for two minutes ( i . e . 12 readings) and write heart rate on the recording sheet under the performance section. 7. remove sensors and turn pulse metre off (b) ask subject to complete questionnaires in following order: Perceived S k i l l , Pre-performance Thought Questionnaire, and SUDS. (c) instruct subject to relax for a few minutes and then perform (d) ask subject to complete questionnaires in the follow-ing order: SUDS and Post-performance Thought Question-naire. 5. measure tonic pulse rate (same as 4. (a)) 1. turn on pulse metre 2. attach sensor to earlobe - light at the back 3. ask subject to s i t quietly for a few minutes 4. ensure "EV signal is off and heart signal is on 5. wait for 2 minutes (use timer on digital display - top left) 6. record the number every 10 seconds for two minutes ( i . e . 12 readings) and write heart rate on the recording sheet under the performance section. I 7. remove sensors and turn pulse metre off Appendix 5(a) 243 PROGRESSIVE MUSCLE RELAXATION 1. either s i t in a comfortable chair or l i e on a bed (but do not f a l l to sleep during the exercises) 2. loosen tight clothing (belts, shoes etc) and eliminate distractors (noise, l i g h t , etc.) 3. take three slow and deep breaths: inhale for 5 sees, hold for 5 sees, and exhale for 5 sees 4. as you work systematically thoughout your body tensing and relaxing the different muscle groups, remember to - tense for approx. 10 sees - relax for approx. 20 sees - try to tense only one muscle group at a time - concentrate on the feelings of tension and relax-ation in the particular muscle group. If other thoughts come into your mind, don't focus on them and don't struggle against them, but just let them pass through 5. tense and then relax the following muscle groups. While relaxing ensure that your breathing is slow and regular. Tense and relax each muscle group twice. hands and forearms (make a f i s t , clench your hands) biceps (tense upper arm) shoulders and neck ( l i f t shoulders towards ears) forehead (frown or l i f t eyebrows) eyes (squeeze eyes - tense for only approx. 5 sees) mouth, jaw and throat (clench teeth or draw corners of mouth back) At this stage, quickly check each of the muscle groups that have been tensed and relaxed to ensure that the tension has not returned: i f you are aware of tension, you may quickly tense and relax that particular muscle group again. chest (take in a big, deep breath and hold for 5 sees, and slowly exhale) back (pull shoulder blades back and arch) abdomen (tighten stomach muscles -prepare for a punch) upper legs ( l i f t legs a few inches off the floor) calf muscle (push toes away from you) shin (pull toes towards you) Again, quickly check each of the muscle groups to ensure that they are s t i l l relaxed. Then spend a couple of minutes in this relaxed state — you may use relaxing imagery (eg. floating on a cloud). Let yourself become more and more relaxed. 6. Slowly become alert again. practise the tensing and relaxing more often with those areas of your body where you experience most tension. practise these exercises for approximately 15 minutes at least once and preferably twice per day. Appendix 5(b) NAME: 245 S e s s i o n : 1 R e l a x a t i o n P r a c t i s e Sheet 0 = complete r e l a x a t i o n 10 = extreme t e n s i o n Date Time SUDS before SUDS a f t e r • 246 Appendix 6(a) QUICK RELAXATION METHOD 1. Sit in a comfortable position, drooping your head forward but keeping your back straight. Rest your hands on the arms of the chair or your thighs. 2. Take three slow and deep breaths. 3. Systematically work through each muscle group, and as you concentrate on each group recall the feelings of relaxation that you experienced in the tense-relax method, (see previous handout). 4. As you proceed through the muscle groups: (a) repeat words to yourself like "relax, heavy and warm". Through repetition, those words will become associ-ated with the state of relaxation and will in them-selves induce this relaxation. (b) use relaxing imagery to either "visualise" the relax-ation spreading throughout your body or to aid the induction of relaxation. 5. Spend approx. 10 to 15 sees on the following muscle groups: hands and forearms upper arms (feel the heaviness and warmth in your whole arm) shoulders and neck (drooping) forehead (smooth) eyes mouth, jaw, throat (feel the relaxation across your whole face) chest (breathing is smooth and easy) abdomen (feel the warmth) back (sink into the chair) upper legs lower legs (feel the heaviness and warmth in your legs) 6. Now you should be feeling relaxed and resting in a drooped position. This set of exercises can be carried out in a few minutes and is very useful just prior to a performance. These exercises can be carried out sitting or standing. Appendix 6(b) NAME: Session: 2 Relaxation Practise Sheet Number of people in audience Did you play from memory? YES NO Level of discomfort: 0 = no discomfort or anxiety 50 = moderate discomfort or anxiety 100 = extreme discomfort or anxiety BEFORE DURING Appendix 7(a) ATTENTION FOCUSING TRAINING HANDOUT Explore what you are saying to yourself. You will probably find i t is something distracting and destructive. Note: your anxiety is a cue to start applying the s k i l l s you have learned. Challenge those destructive thoughts. Replace the destructive thoughts with constructive ones: (a) task focused - focus your attention on the music and not on the audience or a d i f f i c u l t spot two pages ahead, etc. technique oriented - concentrate on the touch, expression, and a l l the other things you have been practising. (b) comforting - reassure yourself before you perform and think about the music rather than what could go wrong or how anxious you feel rewarding - say something good to yourself after you have performed (if you did well and even i f you didn't do as well as you would have liked). 249 Appendix 7(b) NAME: Session: 3 Relaxation and Attention Focusing Practise Number of people in audience Did you play from memory? YES NO Level of discomfort: 0 = no discomfort or anxiety 50 = moderate discomfort or anxiety 100 = extreme discomfort or anxiety BEFORE DURING List your major distracting anxiety e l i c i t i n g thoughts whilst performing List your major task focusing & anxiety reducing thoughts whilst performing 250 Appendix 8 NAME: Assesst: pre post followup Self Evaluation Questionnaire A number of statements which people have used to describe themselves are given below. Read each statement and then blacken in the appropriate circ l e to the right of the statement to indicate how you feel right now; that i s , at this moment. There are no right or wrong answers. Do not spend too much time on any one statement but give the answer which seems to describe your present feelings best.. Moder- Very Not Some- ately Much At a l l what So So 1. I feel calm. 1 2 3 4 2. I feel secure. 1 2 3 4 3. I am tense. 1 2 3 4 4. I am regretful. 1 2 3 4 5. I feel at ease. 1 2 3 4 6. I feel upset. 1 2 3 4 7. I am presently worrying 1 over possible misfortunes. 2 3 4 8. I feel rested. 1 2 3 4 9. I feel anxious. " 1 2 3 4 10. I feel comfortable. 1 2 3 4 1 1 . I feel self confident. 1 2 3 4 12. I feel nervous. 1 2 3 4 13. I am j i t t e r y . 1 2 3 4 14. I feel "high strung". 1 2 3 4 15. I am relaxed. 1 2 3 4 16. I feel content. 1 2 3 4 17. I am worried. 1 2 3 4 18. I feel over-excited & rattled 1 2 3 4 19. I feel pleasant. 1 2 3 4 20. I feel calm. 1 2 3 4 Appendix 9 NAME: 251 Assesst: pre post followup Session 2 3 4 Perceived S k i l l How confident are you that you will perform your piece without experiencing discomfort or anxiety to a degree that will-interfere with your performance? For example, 50 (indicated on the line below) would represent a moderate degree of confidence in one's ability to perform without experiencing discomfort or anxiety to an extent that would interfere with one's performance. Write a number on the line below to indicate your level of conf idence. 0 50 100 completely unconf ident completely conf ident 0 100 completely unconf ident completely conf ident 252 Appendix 10(a) NAME: Assesst: pre post followup Session 2 3 4 Pre-Performance SUDS Please rate your level of discomfort or anxiety on a scale from 0 to 100, where 0 = no discomfort or anxiety 50 = moderate discomfort or anxiety 100 = maximum discomfort or anxiety Write a number on the line below to indicate how much anxiety or discomfort you are experiencing at this moment. 0 100 completely unconf ident completely conf ident 253 Appendix 10(b) NAME: Assesst: pre post followup Session 2 3 4 During-Performance SUDS Please rate your level of discomfort or anxiety you experienced while playing on a scale from 0 to 100, where 0 = no discomfort or anxiety 50 = moderate discomfort or anxiety 100 = maximum discomfort or anxiety Write a number on the line below to indicate how much anxiety or discomfort you experienced while performing. 0 100 completely unconf ident completely conf ident Appendix 11(a) NAME: 254 Assesst: pre post followup Session: 2 3 4 Pre-Performance Thought Questionnaire The items below refer to your thoughts before you begin your performance. A. Tick one option in each item i.e. tick either (a), (b), (c) or (d) 1. Are you thinking mostly about your (a) performance (b) discomfort (c) enjoyment (d) other 2. Are you thinking mostly about (a) previous performances (b) the current performance (c) future performances (d) other 3. If you are thinking about the current performance, which moments are you mostly thinking about (a) comfortable moments (b) mildly uncomfortable moments (c) very uncomfortable moments 4. If you are thinking about the current performance, are you mostly thinking about (a) the parts where you will perform most s k i l l f u l l y (b) the parts where you will perform least s k i l l f u l l y 5. Are you mainly aware of feelings of (a) physical relaxation (b) physical tension 6. If you are thinking about the current performance, are you thinking mostly about (a) your technique (b) the impression you will make on the listeners (c) what musical experts wil l say and think about your audiotape and the mark they 255 will give your performance ((c) was deleted from session-performance questionnaires) 7. Are you thinking mostly about (a) your own performance (b) your performance in comparison to others' performances 256 Appendix 11(b) NAME: Assesst: pre post followup Session: 2 3 4 Post-Performance Thought Questionnaire The items below refer to your thoughts now that you have completed your performance. A. Tick one option in each item i.e. tick either (a), (b), (c) or (d) 1. Are you thinking mostly about your (a) performance (b) discomfort (c) enjoyment (d) other 2. Are you thinking mostly about (a) previous performances • (b) the current performance (c) future performances (d) other 3. If you are thinking about the current performance, which moments are you mostly thinking about (a) comfortable moments (b) mildly uncomfortable moments (c) very uncomfortable moments 4. If you are thinking about the current performance, are you mostly thinking about (a) the parts where you performed most s k i l l f u l l y (b) the parts where you performed least s k i l l f u l l y 5. Are you mainly aware of feelings of (a) physical relaxation (b) physical tension 6. If you are thinking about the current performance, are you thinking mostly about (a) your technique (b) the impression you made on the listeners (c) what the musical experts w i l l say and think about your audiotape and the mark they will give your performance 257 ((c) was deleted from session-performance questionnaires) Are you thinking mostly about How satisfied were you with your performance? Indicate on a scale from 0 to 1 0 0 ( 0 = no at a l l satisfied and 1 0 0 = completely satisfied) by writing a number on the line below: How confident are you that when you perform next you will perform without experiencing discomfort or anxiety to a degree that w i l l interfere with your performance? Write a number on the line below to indicate your level of conf idence: (a) your own performance (b) your performance in comparison to others' performances 0 1 0 0 not at a l l satisfied completely satisfied 0 1 0 0 not at a l l conf ident completely conf ident Having finished performing, are you s t i l l thinking about your performance or about other things unrelated to the performance ? 258 Appendix 12 General Information Name Age Address Postal Code: Phone Number Instrument Years of Experience Toronto or Western Board Grade completed and/or number of years in college or univerity music program grade: years: Have you played any other instruments? YES: NO: If yes, what were they and when did you play them (years)? from to from to from to Has there been any length of time when you have been unable to perform or practise? YES: NO: If yes, when from to why (e.g. i l l n e s s , vacation etc.) Also, what was i t like when you began practising again in private SAME: DIFFERENT: If different, was i t easier or more d i f f i c u l t were you less anxious or more anxious other Also, what was i t like when you began performing before others again? SAME: DIFFERENT;: 259 If different, was i t easier or more d i f f i c u l t were you less anxious or more anxious other How many times have you performed before audiences? Do you believe that you perform better when: (tick one option in each line) A. relaxed a l i t t l e anxious very anxious B. practise l i t t l e moderate practise overpractise C. small audience (5-10) large audience (50-100) D. performance is not evaluated performance is evaluated E. audience is familiar (friends etc.) audience of strangers F. following a break from practise immediately following practise Appendix 13(a) NAME: Assesst: , pre Practise Information Have you practised as much as you normally would over the past 4 weeks? yes no If yes, how often did you and do you normally practise? number of days per week length of practise per day Have you been unable to practise over this time period? yes no If yes, how many days were you unable to practise . Also, briefly state the reason: Have you performed before an audience over this time period? yes • no If yes, on how many occasions _. 261 Appendix 13(b) NAME: Assesst: post Practise Information Have you practised as much as you normally would over the past week? yes no If no, have you practised more or less ? Please specify: number of days per week length of practise per day Have you been unable to practise over this time period? yes .. no . If yes, how many days were you unable to practise . A l s o , briefly state the reason: ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ * Have you performed before an audience over this time period? yes no If yes, on how many occasions . Please estimate the average amount of time spent practising the piece you played today since the previous assessment. number of days per week number of minutes per day 262 Appendix 13(c) NAME: Assesst: followup Practise Information Have you practised as much as you normally would over the past 4 weeks? yes no If yes, how often did you and do you normally practise? number of days per week length of practise per day If no, have you practised more or less ? Please specify: number of days per week length of practise per day Have you been unable to practise over this time period? yes no If yes, how many days were you unable to practise . Also, briefly state the reason: • Have you performed before an audience over this time period? yes no If yes, on how many occasions . Also, briefly describe the types of performances: Were you able to apply the s k i l l s you acquired from the training program in those performances? yes no Do you think acquiring these s k i l l helped you in those performances? yes no Please estimate the average amount of time spent practising the piece you played today since the previous assessment. number of days per week number of minutes per day 264 Appendix 14 NAME: Session: 1 2 3 4 Session Practise Information Have you practised as much as you normally would over the past 4 weeks? yes no If yes, how often did you and do you normally practise? number of days per week length of practise per day If no, have you practised more or less ? Please specify: number of days per week length of practise per day Have you been unable to practise over this time period? yes no If yes, how many days were you unable to practise ' Also, briefly state the reason: Have you performed before an audience over this time period? yes no If yes, on how many occasions? . 265 Appendix 15 NAME: Session: 3 4 post Memory Questionnaire Please answer the following in reference to your performance in the previous session: 1. what were your levels of reported discomfort, where 0 = no discomfort, 50 = moderate discomfort and 100 = extreme discomfort: before playing after playing how distressed were you (indicate by writing a number on the line) 100 not at a l l extremely distressed distressed 3. how distressed do you expect to be this session 0 100 not at a l l extremely distressed distressed 4. how strong were your physical reactions (indicate by writing a number on the line) 100 not at a l l extremely strong strong 5. how many people listened to you perform? 6. how many minutes did you perform? 7. who performed before you? 8. who performed after you? 9. who sat closest to you while you performed? 10. who sat closest to the door while you performed? 266 Appendix 16 NAME: Month: Followup Practise Information Have you practised as much as you normally would over the past 4 weeks? yes no If yes, how often did you and do you normally practise? number of days per week _ length of practise per day If no, have you practised more or less? more less How often do you normally practise? number of days per week length of practise per day How often have you been practising? number of days per week length of practise per day Have you been unable to practise over this time period? yes no If yes, how many days were you unable to practise . Also, briefly state the reason: • Have you performed before an audience over this time period? yes no If yes, on how many occasions . Also, briefly describe the types of performances: Were you able to apply the s k i l l s you acquired from the training program in those performances? yes no 267 Do you think acquiring these s k i l l helped you in those performances? yes no How confident are you that when you next perform, you wi l l do so without experiencing anxiety or discomfort to a degree that will interfere with your performance? Indicate on a scale from 0 to 100, where 0 = completely unconfident 100 = completely confident 268 Appendix 17 Definitions of Performance Quality Dimensions (1) Touch technical efficiency of the performer to achieve clarity and sound quality (slurs, staccatos, legato lines and accents, and pedalling). (2) Phrasing - c l a r i t y , subtlety, control, length, originality and variety of phrasing; degree to which the performer is aware of and shapes phrases -expressiveness. (3) Pitch and Omission - incorrect, omitted or blurred notes or chords (taking into account the relative technical d i f f i c u l t y of the piece and the importance of the incorrect note to the presentation). (4) Rhythm - time value of note or rest; clarity with which the indicated metrical rhythm is rendered (misplaced barlines or stresses); clarity of the rhythmic harmonic progression; rubato; overall rhythmic timing of sections (placements and arrivals of phrases); coordination of passages. (5) Tempo - steadiness of tempo; s t y l i s t i c appropriateness of tempos; consistency and appropriateness of speeds of passages. (6) Dynamics - relative effectiveness of dynamics chosen (timing, variety, s t y l i s t i c appropriateness and dynamic range); balancing of chords and lines; interpretation of indicated (in score) dynamics or accents. (7) Memory - abili t y to remember the piece, and to recover from a memory s l i p (taking into account the importance of the s l i p to the passage). (8) Overall Quality - the extent to which the performance is convincing in its technical, s t y l i s t i c and musical qualities; the emotional conviction of the performer. Voice Ratings (1) Vocal Production technical efficiency of the performer to achieve clarity of sound quality (slurs, staccatos, legato lines and accents). 269 Appendix 18 Performance Quality Rating Rater: -Subject # : Performance # : Rate each dimension on the following scale: 1 2 3 4 5 6 7 8 9 10 Totally Average Excellent Unacceptable 1 . Touch/vocal production 2. Phrasing 3. Pitch and omission 4. Rhythm 5. Dynamics 6. Phrasing 7. Memory 8. Overall Quality 270 Appendix 19 I n i t i a l Contact My name is Michelle Craske and I am a PhD student in c l i n i c a l psychology. I am interested in musical performance anxiety and ways to reduce i t . As an i n i t i a l survey of the sort of experiences you have when performing, i t would be of great help to me i f you could complete this questionnaire. For those of you who are interested in quickly learning effective strategies for reducing performance anxiety, I'm going to be running a training program from to , involving 4 weekly sessions in a group format of 4 to 8 people with similar concerns. During these sessions, I will be teaching a method of relaxation, and ways to focus your attention on the task of performing. These techniques have been implemented before with good success. To assess the effects of these s k i l l s and to provide me with some information, three assessments will also be conducted; immediately before the program, immediately after and then three months after, so those who wish to participate would have to be available for a l l of those assessments. If you are interested, be sure to place your name and phone number on the questionnaire and then I will contact you soon. Thankyou. Appendix 20(a) 271 Dear piano, v i o l i n , or voice teacher, I am a PhD student in c l i n i c a l psychology at the University of British Columbia and I will be offering a free program for pianists, v i o l i n i s t s and vocalists suffering from performance anxiety which prevents them from demonstrating their s k i l l s and musicianship to their best a b i l i t y , and makes the performing experience unpleasant for them. As music techer, I'm sure you are aware that this is a common d i f f i c u l t y among many music students. This program is directed toward students who do well at home or for their teachers, but in performance situations tend to perform at a level below their own standard. I am not attempting to reduce anxiety completely, as that would result in a dull l i f e l e s s performance. I am concerned with reducing i t to the point where i t is under control and enhances rather than interferes with the performance. The behavioural science have successfully demonstrated ways of reducing anxiety in this area, as shown by the work of Margaret Kendrick in 1979 who focused on the attentional component of performing. I expect that including a muscle relaxation component into the program used by M. Kendrick will maker her program even more effective. This program consists of four 1 and 1/2 hour evening sessions over the course of four weeks, to be held in . Participants may choose the night they would like to attend. In order to assess the benefits of the program, participants wi l l be asked to do a pre-program and post-program performance and a second post-program performance three months Appendix 21 274 OUTLINE OF ASSESSMENT PROCEDURES FOR SUBJECTS I'm calling about the training program for musical performance anxiety. It entails 4 weekly sessions of 1 1/2 to 2 hours length each with a group of other musicians (piano, v i o l i n & voice) where I will teach you a method of muscle relaxation, and attention focusing and then let you practise these s k i l l s in performances. There are also assessments before and after the program that will allow us to determine i t s effectiveness and individual differences in response to the program. These assessments will require you to perform a piece that you feel comfortable playing/singing from memory and which is at your grade level of performance, before an audience of 5 people and which will also be audiotaped so i t can be rated by musicians. The assessment will be repeated three months following the program to examine long term effects. Are you interested in participating? Please choose a piece from 3 to 5 minutes in length, no longer -it must be something you are able to perform from memory and which is from your grade level of performance. It should not be a piece that you are going to perform in the near future (ie. within the next 4 months). Also, please bring a copy of the music for us to keep and remember to put your name and the t i t l e on the top and if you plan anything different from the way i t is written (eg. tempo, dynamics) be sure to make the changes on the sheet. 275 The f i r s t assessment is being held and the training sessions will begin a week later. What time can you come? After that we'll arrange for the group meetings. So, I ' l l see you at on with a copy of your music. Appendix 22 276 PROCEDURAL OUTLINE FOR ASSESSMENTS RA 1 1. collect music copy - ensure that name and t i t l e are written - if subject failed to bring a copy of their music, write name on the sheet provided, (deleted at post-assessment and followup assessment) RA 1 2. seat subject RA 1 3. Read the following instructions: Before you perform we will be taking a measurement of your heart rate by attaching these three elec-trodes to your arms and leg. There is no danger nor discomfort involved in this procedure. You are required to sit quietly for a few minutes. Then, while we remove the electrodes, you can begin f i l l i n g in some questionaires. Then we will take you into the next room and you can perform your piece. Remember that your performance is being audiotaped so that i t can later be rate by musical experts. Are there any questions? RA 1 4. Give subject the consent form to sign. (deleted at Postassessment and Followup Assessment) RA 1 5. Attach electrodes: (a) right inside forearm (b) lef t inside forearm (c) l e f t leg, just below the knee (if subject is' wearing stock-ings, ask them to remove them in the washroom). I. abrade skin with redux paste t i l l redness appears (40 sees) II. wipe off excess paste with tissue III. peel collar off reel IV. place collar evenly over electrode V. f i l l electrode with electrode paste (no air bubbles) VI. attach electrode firmly to the skin VII. ask subject to hold electrode leads and direct to the polygraph. Experi-6.. Heart rate recording - 5 minutes menter 277 RA 2 7. Give subject questionnaires in the following order': Self Evaluation Questionnaire, Perceived S k i l l , Pre-Performance Thought Questionnaire, SUDS. (Memory Questionnaire is included at Postassessment). and say "First impressions are the most appropriate responses to give to these questionnaires." RA 2 8. While subject completes the questionnaires remove the electrodes: I. pull off the collar and electrode together (with one finger on the back of the electrode) - do not pull off by the lead. II. wipe skin clean with tissue III. pull collar off the electrode (again with finger on the back of the electrode) and throw collar away. IV. remove paste from the electrode with a tissue (do not scratch the silver surface). RA 2 9. Inform subject: Now we'll take you in to the audience, and someone in there wi l l turn on the audiotape when you are ready. After the tape has begun, announce your name and the t i t l e of your piece quite clearly so it can be heard on the tape. When you have f i n -ished, the person who turned on the tape will walk with you back to this room. Have you any questions? RA 1 10. seat the subject RA 1 11. Give subject questionnaires in the following order: SUDS, Post-Performance Thought Questionnaire, General Information Questionnaire and Practice Information Questionnaire. (General Information Questionnaire is deleted at Postassessment and Followup Assessment) Experi-12. Ask subject to indicate choices for days and times menter for group meetings on the sheets provided, (deleted at Postassessment and Folloup assessment. Arrange-ments for Followup Practise-Information phonecalls included at Postassessment). Experi-13. Inform the subject that he/she will be phone con-menter tacted to finalise these arrangements. Experi-1 4... menter Give subject the practise instructions. 278 Assessment: pre post followup ASSESSMENT SCHEDULE TIME NAME TIME NAME 9 : 0 0 - 9 : 3 5 1:00 - 1 :35 9 : 1 0 - 9 : 4 5 1 : 1 0 - 1 :45 9 : 2 0 - 9 : 5 5 1 : 2 0 - 1 :55 9:30 - 1 0 : 0 5 1:30 - 2 : 0 5 9 :40 - 1 0 : 1 5 1:40 - 2 : 1 5 9 :50 - 1 0 : 2 5 1 :50 - 2 : 2 5 10:00 - 1 0 : 3 5 2 : 0 0 - 2 : 3 5 10:10 - 1 0 : 4 5 2 : 1 0 - 2 : 4 5 10:20 - 1 0 : 5 5 2 : 2 0 - 2 : 5 5 10:30 - 1 1 : 0 5 2 : 3 0 - 3 : 0 5 TIME NAME TIME NAME 9:00 - 9 : 3 5 1 :00 - 1 :35 9 : 1 0 - 9 : 4 5 1 : 1 0 - 1 : 45 9 : 2 0 - 9 : 5 5 1 :20 - 1 : 55 9:30 - 1 0 : 0 5 1:30 - 2 : 0 5 9 : 4 0 - 1 0 : 1 5 1:40 - 2 : 1 5 9 : 5 0 - 1 0 : 2 5 1 :50 - 2 : 2 5 10:00 - 1 0 : 3 5 2 : 0 0 - 2 : 3 5 10:10 - 1 0 : 4 5 2 : 1 0 - 2 : 4 5 10:20 - 1 0 : 5 5 2 : 2 0 - 2 : 5 5 10:30 - 1 1 : 0 5 2 : 3 0 - 3 : 0 5 Appendix 23 CONSENT FORM I have agreed to participate in this program t i t l e d Training Sk i l l s for Performance Anxiety being conducted by Professor S.J. Rachman and Michelle Craske of the Psychology Department, U.B.C. I have been informed about the requirements of participating in this program for reducing my musical performance anxiety. I have made a commitment to attend three performances and four training sessions, a l l of which wil l entail approximately eight hours of time. I am aware that my three performances wil l be audiotaped and that a measure of my heart rate will be taken, and that I will complete several questionnaires before and after performing. I also know that similar heart rate and questionnaire measures will be taken during the training session, where I will learn relaxation and attention focusing techniques to use in performance situations. I know that this and a l l other information gathered from me in the course of the program will be kept confidential by means of numerical coding. It will be used to assess the benefits of the program. I know that I am able to make inquiries concerning the procedures to ensure that they are fully understood. If for some reason I wish to terminate my participation in the program, I know that I am free to do so at any time without jeopardising my class standing. I have read this consent form and agree to participate in this project. Signature 280 Appendix 24 PRACTISE INSTRUCTIONS In order to better determine the effects of your participation in this program, please try to restrict practising of the piece you performed today to no more than 15 minutes per day between now and next assessment. Also, if possible, do not perform this piece again until then. Appendix 25 NAME: 281 Session: 2 3 4 Answers to Memory Questionnaire 1. Levels of discomfort before during 2. Number of people in audience 3. Number of minutes played 4. Previous player 5. Subsequent player 6. Closest member of audience 7. Audience member closest to door 282 Appendix 26 Debriefing Letter Dear • , It has now been approximately two months since I completed my last "performance anxiety" program. I have analysed most of my data and would like to explain to you more fully the nature of the research study in which you participated, and describe some of the findings I have obtained. Essentially, I was interested in (a) the extent to which improvements were felt in level of comfort while performing as a result of the training program, and (b) the extent to which those improvements were maintained, and, more specifically, in whom such improvements were maintained more fully than others. I was attempting to determine i f heart rate and self efficacy (or, confidence) related to maintenance of improvement. The results seem to show that most people who participated in the program reported less anxiety and greater s k i l l when performing in front of the audience at the post-program assessment. In addition, many maintained their improvements at the three month followup assessment. In agreement with one of my hypotheses, improvements felt by those people who had higher heart rates tended to be less durable than those experienced by musicians with lower heart rates. That i s , arousal levels seemed to influence the permanence of the program effects. In contrast to my second hypothesis, however, self efficacy did not seem to have a significant effect on maintenance of improvements. The number of performances in the interval between the post and followup assessments, however, was also important - those who performed more often generally experienced Appendix 27 284 Instrumental Group Differences Variable Pre-performance SUDS During-perf ormance SUDS Heartrate State Anxiety S k i l l Screening Group Singers Pianists/ Violinists Singers Pianists/ Violinists Singers Pianists/ Violinists Singers Pianists/ Violinists Singers Pianists/ Violinists Singers Questionnaire Pianists/ Violinists Years of experience Variable Singers Pianists/ Violinists Mean 54.00 60.78 60.00 65.35 82.88 87.74 49.20 48.74 5.57 6.32 1 0.00 9.57 7.20 12.55 Standard Deviation 5.48 1 2.42 22.36 18.56 1 1 .09 12.32 6.76 12.43 0.54 1 .42 1 .23 0.84 7.50 6.33 Homogeneity of Variance Pre-SUDS During-SUDS Heart Rate State Anxiety S k i l l Screening questionnaire Years of experience t df p -1.20 61 >0.1.0 -0.61 61 >0.10 -0.85 61 >0.10 -0.08 61 >0.10 1.17 61 >0.10 1.06 61 >0.10 -1.79 61 >0.05 5.1 4 1 .45 1 .23 3.38 6.95 2.13 1 .40 >0. 1 0 >0. 10 >0. 10 >0. 1 0 >0. 10 >0. 10 >0.. 1 0 Appendix 28 285 Classification Group Differences on Subject Variables Standard Variable Group Mean Deviation F df p_ Screening 1 9.69 0.95 0.41 3,59 >0.10 questionnaire 2 9.13 0.71 3 9.53 0.99 4 9.44 0.81 Years of 1 12, .86 6. ,56 1.18 3,59 >0 experience 2 9. .30 4. .43 3 13. 30 9. .82 4 12, .91 4. .41 Age 1 2 3 4 25.36 22.81 25.52 21 .40 8.39 4.62 9.61 4.39 2.20 3,59 >0.10 Years in 1 0. 75 1 .39 3.02 3,59 Performance 2 2. 13 1 .36 Program 3 2. 00 1 .89 4 1 . 94 1 .18 Variable X X df E Sex 6 .13 3 >0.05 Instrument 5 .65 6 >0.05 Group 1 = high heart rate, low. efficacy 2 = high heart rate, high efficacy 3 = low heart rate, low efficacy 4 = low heart rate, high efficacy Appendix 29 286 I n i t i a l Classification Group Differences on Dependent Variables Standard Variable Group Mean Deviation F df E pre- 1 72.19 1 0.95 1 1 .39 3,59 <0 .01 perf ormance 2 59.06 10.36 SUDS 3 56.33 1 1 .72 4 53. 13 5.44 pre- 1 9.63 1 .20 4.62 3,59 <0 .01 perf ormance 2 9.13 1 .02 Thoughts 3 7.93 3.35 4 7.06 2.29 S k i l l 1 6.45 1 .38 3.66 3,59 <0 .05 2 5.57 1 .53 3 5.99 1.01 4 7.02 1 .94 Groups 1 = high heart rate, low efficacy 2 = high heart rate, high efficacy 3 = low heart rate, low efficacy 4 = low heart rate, high efficacy Appendix 30 287 Classification Group Means on Heart Rate and Efficacy Time - Mean (standard deviation) Variable Group A B C Heartrate 1 95. 34 (9.07) 84. 30 (9.81) 80.91 (8.42) 2 98. 90 (6.14) 87. 34 (9.82) 85.73 (12.33) 3 76. 31 (6.07) 73. 99 (8.71) 72.13 (9.15) 4 78. 18 (6.39) 73. 67 (8.73) 73.24 (5.36) D E F 1 87. 11 (8.01) 85. 86 (9.80) 90.48 (11.45) 2 88. 46 (9.47) 92. 46 (8.96) 90.29 (10.07) 3 75. 75 (8.55) 75. 33 (6.71) 78.35 (8.62) 4 75. 28 (7.37) 76. 43 (8.. 11) 80.86 (6.36) A B C Eff icacy 1 23. 13 (9.64) 43. 56 (18.45) 54.19 (16.10) 2 60. 00 (13.29) 63. 13 (10.78) 59.25 (17.37) 3 33. 00 (18.01) 45. 47 (15.06) 54.00 (20.63) 4 70. 31 (11.03) 69. 06 (9.70) 73.06 (11.41) D E F 1 51 . 25 (13.84) 61 . 56 (14.57) 50.31 (17.75) 2 50. 56 (19.62) 68. 13 (15.04) 64.69 (19.10) 3 54. 33 (22.90) 64. 67 (16.74) 60.00 (19.09) 4 62. 38 (15.80) 75. 67 (11.32) 73.00 (18.38) Appendix 31 288 Summary of HRxEFxTime Univariate Analysis of pre-performance SUDS Effect SS F df E HR 6336.61 7.28 1,51 <0.025 EF 4146.18 4.76 1 ,51 <0.05 HRxEF 1685.57 1 .94 1,51 >0. 1 0 HRxEFxTime 746.46 0.93 5,287 >0. 1 0 A-BCDEF 10.16 0.07 1 ,57 >0.1 0 BCD-EF 114.00 0.52 1 ,57 >0.10 E-F 0.67 0.01 1 ,57 >0.10 EFxTime 1636.90 2.04 5,287 >0.05 A-BCDEF 49.73 0.34 1 ,57 >0.1 0 BCD-EF 635.93 2.89 1 ,57 >0.05 E-F 109.18 0.81 1 ,57 >0. 1 0 HRxTime 2690.04 3.36 5,287 <0.025 A-BCDEF 143.72 0.98 1 ,57 >0. 1 0 BCD-EF 770.39 3.50 1 ,57 >0.05 E-F 1439.79 1 0.70 1 ,57 <0.01 Time 14370.62 17.94 5,237 <0.001 A-BCDEF. 11465.09 77.79 1 ,57 <0.001 BCD-EF 572.28 2.60 1 ,57 >0. 1 0 E-F 536.51 3.99 1 ,57 <0.05 Residual planned orthogonal contrasts not listed Error df corrected for 8 missing cases Appendix 32 289 Summary of HRxEFxTime Univariate Analysis of Covariance of pre performance SUDS Effect SS F df P. Covariate 8318.63 10.49 1 ,50 <0.01 HR 184.41 1 .69 1 ,50 >0.1 0 EF 233.33 0.23 1 ,50 >0.1 0 HRxEF 148.15 0.29 1 ,50 >0.1 0 HRxEFxTime 736.30 1.13 4,228 >0.1 0 E-F 0.67 0.01 1 ,57 >0.1 0 EFxTime 1587.17 2.43 4,228 <0.05 E-F 109.18 0.81 1 ,57 >0.1 0 HRxTime 2546.32 3.90 4,228 <0.01 E-F 1439.79 10.70 1 ,57 <0.01 Time 2905.52 4.45 4,228 <0.01 E-F 536.51 3.99 1 ,57 <0.05 Planned orthogonal contrast of interest only is listed Error df corrected for 8 missing cases 290 Appendix 33 Simple Effects Analyses of HRxTime and EFxTime of pre-performance SUDS Effect SS F df E HRxTime Time A 1887.15 18.68 1 ,59 <0.001 Time B 707.39 2.68 1 ,58 >0. 1 0 Time C 822.14 2.56 1 ,56 >0. 1 0 Time D 27.59 0.10 1 ,55 >0.1 0 Time E 348.97 1 .06 1 ,59 >0.1 0 Time F 5233.41 1 3.85 1 ,59 <0.001 EFxTime Time A 116.978 11.41 1 ,59 <0.001 Time B 2069.44 8.17 1 ,58 <0.01 Time C 8.36 0.05 1 ,56 >0.10 Time D 0.23 0.00 1 ,55 >0. 1 0 Time E 1006.22 3.16 1 ,59 >0.05 Time F 2243.11 6.37 1 ,59 <0.05 Error df adjusted for missing cases Appendix 34 291 Summary of HRxEFxTime Univariate Analysis of Variance of Pre-Performance Thoughts Effect SS F df E HR 1 35. 08 20, .06 1 ,51 <0. ,001 EF 75. 08 1 1 . .15 1 ,51 <0. ,01 HRxEF 2. 25 0. .33 1 ,51 >0. ,10 HRxEFxTime 9. 87 0, .75 5,287 >0. ,10 A-BCDEF 0. 02 0, .01 1 ,57 >0. ,10 BCD-EF 2. 57 0. 82 1 ,57 >0. ,10 E-F 0. 72 0. 51 1 ,57 >0. ,10 EFxTime 23. 1 1 1 , .76 5,287 >0. ,10 A-BCDEF 0. 83 0, .22 1 ,57 >0. ,10 BCD-EF 0. 16 0, .05 1 ,57 >0. ,10 E-F 5. 00 3. .57 1 ,57 >0. ,05 HRxTime 19. 55 1 . .49 5,287 >0. ,10 A-BCDEF 9. 1 4 2, .39 . 1 ,57 >0. ,10 BCD-EF 7. 45 2. 38 1 ,57 >0. ,10 E-F 1 . 28 0. 91 1 ,57 >0. ,10 Time 71 . 89 5. 48 5,287 <0. ,001 A-BCDEF 44. 50 1 1 , .65 1 ,57 <0. ,001 BCD-EF 5. 83 1 . .86 1 ,57 >0. , 1 0 E-F 2. 87 2, ,05 1 ,57 >0. ,10 Residual planned orthogonal contrasts not listed Error df corrected for 8 missing cases Appendix 35 292 Summary of HRxEFxTime Univariate Analysis of Covariance on Pre Performance Thoughts Effect SS F df 2 Covariate 24.26 4.33 1 ,50 <0.05 HR 42.28 7.55 1 ,50 <0.01 EF 54.41 9.71 1 ,50 <0.01 HRxEF 1 .20 0.21 1 ,50 >0.1 0 HRxEFxTime 9.86 1 .06 4,228 >0. 10 E-F 0.72 0.51 1 ,57 >0. 1 0 EFxTime 22.28 2.30 4,228 >0.05 E-F 5.00 3.57 1 ,57 >0.05 HRxTime 10.41 1.12 5,228 >0. 1 0 E-F 1 .28 0.91 1 ,57 >0. 10 Time 27.40 2.94 4,228 <0.05 E-F 2.87 2.05 1 ,57 >0. 1 0 Planned orthogonal contrast of interest only listed Error df corrected for 8 missing cases 293 Appendix 36 Means & Summary of HRxEFxTime Univariate Analysis of Practise Effect SS F df p HR 627.63 0.16 1,51 >0.10 EF 8042.48 2.00 1,51 >0.10 HRxEF 790.14 0.20 1,51 >0.10 HRxEFxTime 1242.21 0.17 5,287 >0.10 EFxTime 6212.14 0.85 5,287 >0.10 HRxTime 7480.94 1.02 5,287 >0.10 Time 15145.63 2.07 5,287 <0.10 Time - Mean (standard deviation) Group A B C 1 77.06 (62.53) 60.25 (39.53) 60.50 (38.97) 2 83.44 (32.14) 79.25 (40.99) 82.25 (30.91) 3 91.33 (51.38) 62.07 (32.03) 57.27 (39.75) 4 92.38 (33.34) 74.44 (29.75) 84.44 (37.29) D E F 1 64.19 (40.10) 74.56 (34.32) 81.63 (73.70) 2 66.31 (43.64) 87.50 (40.84) 91.88 (55.25) 3 78.47 (34.68) 85.87 (42.72) 75.40 (41.11) 4 81.00 (56.68) 82.75 (34.39) 73.13 (45.17) Error df corrected for 8 missing cases Appendix 37 294 Summary of HRxEFxTime Univariate & Covariate Analysis of S k i l l Effect SS F df 2 HR EF HRxEF 1 .06 8.11 10.04 0.30 2.29 2.83 1 ,59 1 ,59 1 ,59 >0.1 0 >0.1 0 >0.05 HRxEFxTime 6.23 5.21 2,118 <0.05 EFxTime 3.09 2.58 2,118 >0.05 HRxTime 3.81 3.19 2,118 <0.05 Time 8.72 7.29 2,118 <0.00 Summary of HRxEFxTime Covariance Analysis of S k i l l Effect SS F df 2 Covariate HR EF HRxEF 54.68 2.78 10.06 2.45 31 .77 1 .62 5.85 1 .43 1 ,50 1 ,50 1 ,50 1 ,50 <0.001 >0.1 0 <0.05 >0.10; HRxEFxTime 0.46 1.17 1 ,51 >0. 1 0: EFxTime 0.03 0.07 1 ,51 >0.1 0: HRxTime 0.76 1 .93 1 ,51 >0.10\ Time 0.47 1 .20 1 ,51 >0.10 Appendix 38 295 Pre-, Post- and Followup-Assessment of Pre-performance SUDS and Heart Rate Variable Effect SS F df p_ Pre-performance SUDS HR 6026.88 13.30 1,59 <0.005 EF 4058.98 8.96 1,59 <0.01 HRxEF 542.22 1.20 1,59 >0.10 HRxEFxTime 58.82 0.19 2,118 >0.10 EFxTime 133.89 0.43 2,118 >0.10 HRxTime 1442.65 4.57 2,118 <0.05 Time 11872.22 37.64 2,118 <0.001 Heart rate Time 1004.09 11.34 2,124 <0.001 Appendix 39 296 Summary of Univariate Analyses of Heart Rate by Memory for Time B Effect Variable SS- F df p HR Pre-performance 1346.68 4.53 1,58 <0.05 SUDS accuracy During-peformance 178.82 0.48 1,58 >0.10 SUDS accuracy Arousal recall 4051.33 8.59 1,58 <0.01 Stimuli accuracy 0.08 0.04 1,58 >0.10 Error df corrected for one missing case Appendix 40 297 Summary of Univariate Analyses of Tonic Heart Rate Variable Effect SS F df Pre-performance SUDS Tonic 236. 07 0. 23 1 ,53 >0. 10 TonicxTime 769. 47 0. 91 4.32,255. 67 >0. 1 0 A-BCDEF 136. 49 0. 95 1,51 >0. 10 BCD-EF 1 . 76 0. 01 1,51 >0. 10 E-F 487. 14 3. 31 1 ,51 <0. 10 Time 14127. 80 16. 72 4.32,255. 67 <0. 001 A-BCDEF 11320. 40 78. 80 1 ,51 <0. 001 BCD-EF 574. 00 2. 42 1 ,51 >0. 10 E-F 487. 74 3. 31 1 ,51 <0. 10 Pre-performance Thoughts Tonic 39. .34 4, .21 1 ,53 <0. 05 TonicxTime 10. .73 0, .80 4.32,257. 71 >0. 10 A-BCDEF 6. .44 1 , .72 1 ,51 >0. 10 BCD-EF 0. .26 0, .08 1 ,51 >0. 1 0 E-F 1 . . 1 3 0, .78 1,51 >0. 10 Time 70. .64 5, .28 4.32,257. 71 <0. 001 A-BCDEF 42. .80 1 1 , .41 1,51 <0. 005 BCD-EF 5, .70 1 , .79 1,51 >0. 10 E-F 3. .03 2, .09 1 ,51 >0. 10 Residual planned orthogonal contrasts not listed df adjusted using Greenhouse-Geisser and for 8 missing cases Appendix 41 298 Summary of Univariate analysis of Upper and Lower Extreme Tonic Heart Rate Groups on pre-performance SUDS Effect SS F df E Tonic 295. 1 5 0. .27 1 ,20 >0. 10 TonicxTime 1152. 42 1 . .08 5,128 >0. 10 A-BCDEF 138. 1 5 0. .96 1 ,25 >0. 10 BCD-EF 3. 10 0. 01 1 ,25 >0. 10 E-F 809. 49 5, .84 1 ,25 <0. 05 Time 8803. 36 8. .25 5, 128 <0. 01 A-BCDEF 7568. 23 52. .77 1 ,25 <0. 01 BCD-EF 308. 41 0, .98 1 ,25 >0. 10 E-F 622. 41 4. .49 1 ,25 <0. 05 Residual planned orthogonal contrasts no listed Error df adjusted for 7 missing cases Appendix 42 299 Summary of Return of FearxTime Univariate Analyses Var iable Effect SS F df Heart Rate Group 6223.26 15.19 1 ,42 <0.001 GroupxTime 182.24 0.79 4. 4,208. 2 >0. 10 A-BCDEF 18.41 0.33 1,41 >0. 10 BCD-EF 2.37 0.04 1,41 >0. 10 E-F 116.97 3. 18 1,41 <0. 1 0 Time 2482.20 10.80 4. 4,208. 2 <0.001 A-BCDEF 1350.27 24. 12 1,41 <0.001 BCD-EF 713.44 10.59 1 ,41 <0.005 E-F 96.93 2.64 1 ,41 >0.1 0 Efficacy Group 8825.26 8.91 1 ,42 <0.005 GroupxTime 300.37 2.96 4. 7,212. 8 <0.025 A-BCDEF 343.03 1 .26 1 ,42 >0. 1 0 BCD-EF 135.89 0.61 1 ,42 >0.10 E-F 2384.91 22.51 1 ,42 <0.001 Time 16491 . 14 16.26 4. 7,212. 8 <0.001 A-BCDEF 10515.70 38.46 1 ,42 <0.001 BCD-EF 4578.63 20.58 1 ,42 <0.001 E-F 753.89 7.11 1 ,42 <0.05 Pre-performance Thoughts Group 58.56 6.00 1 ,42 <0.025 GroupxTime 23.96 2.05 4. 3,205. 9 <0. 10 A-BCDEF 3.67 1 .05 1 ,41 >0. 10 BCD-EF 1 1 .53 5.59 1 ,41 <0.05 E-F 7.08 5.25 1 ,41 <0.05 Time 73. 10 6.26 4. 3,205. 9 <0.001 A-BCDEF 50.86 1 4.57 1 ,41 <0.001 BCD-EF 5.19 2.52 1 ,41 >0.10 E-F 2.79 2.07 1,41 >0. 10 Efficacy Group 378.29 0.90 1 ,42 >0. 10 Differ- GroupxTime 2524.67 1 .87 4, 183 >0. 10 ences dif4-dif5 2360.01 17.52 1 ,45 <0.001 Time 13578.17 10.05 4, 183 <0.001 dif4-dif5 1991.03 14.78 1 ,45 <0.001 Thought Group 1 9.48 2.51 1 ,42 >0.10 Differ- GroupxTime 21 .55 1.15 4, 183 >0.10 ences dif4-dif5 14.02 3.18 1 ,45 <0. 10 Time 61 .60 3.29 4, 183 <0.025 dif4-dif5 1.80 0.47 1 ,45 >0.10 300 S k i l l Group 14.66 GroupxTime 3.62 A-E 2.28 E-F 1.35 Time 7.54 A-E 6.96 E-F 0.58 4.95 1,42 <0.05 3.08 2,89 <0.10 3.00 1,44 <0.10 3.23 1,44 <0.10 6.40 2,89 <0.005 9.16 1,44 <0.005 1.38 1,44 >0.10 Major planned orthogonal contrasts only listed df adjusted using Greenhouse-Geisser for Heartrate, Efficacy and Prethought variables, and error df corrected for 5 missing cases Appendix 43 301 Multiple Regression of Return of Fear Variables Beta T 2 Heart Rate A -0. 30 -2.54 0 .01 Efficacy A 0. 1 3 1.11 0 .27 Pre-performance Thoughts E-F 0. 23 1 .96 0 .05 0.19 4.61 3,59 <0.025* No. of performances Amount of practise Years of experience Grade 0.31 2.36 0.02 0.13 0.99 0.33 0.06 0.45 0.65 0.06 0.4-7 0.64 0.13 2.03 4,53 >0. 10 Appendix 44 302 Summary of Between-Assessments Return of Fear Univariate Analyses of Variance Return of  of Fear Interval riable SS F df p B-C Thought differences 57.40 15.67 1,57 <0.001 Pre-performance SUDS accuracy 34.33 0.11 1,57 >0.10 During-performance SUDS accuracy 2805.36 8.77 1,57 <0.005 Stimuli accuracy 3.81 1.76 1,57 >0.10 D-E Thought differences 23.84 4.01 1,57 <0.10 Pre-performance SUDS accuracy 24.17 0.10 1,57 >0.10 During-performance SUDS accuracy 1347.05 6.13 1,57 <0.025 Stimuli accuracy 2.51 1.00 1,57 >0.10 Error df corrected for 4 missing cases Appendix 45 303 Correlations Among Major Dependent Variables at each Assessment Assessment Occasion I A II I.II I B II III I C II III Pre-SUDS Heart Rate Efficacy Prethoughts S k i l l .32 .47 .39 .03 .30 -.23 .62 -.19 .15 -.51 .13 .21 .27 -.51 .50 .29 -.65 .53 .02 -.27 „ .45 .04 -.19 .27 .37 -.36 .30 -.20 -.14 .35 Heart Rate Efficacy Prethoughts S k i l l .09 .35 .13 .24 .50 .13 -.09 .13 -.31 -.18 .26 -.14 .34 .01 -.03 -.19 -.01 -.29 .32 . 1 2 -.35 Efficacy Prethoughts-S k i l l .17 . 1 1 . 1 1 .15 -.40 -.02 -.42 -.39 -.41 -.47 -.33 -.62 Prethoughts S k i l l .10 -.01 -.08 I D II III I E II III I F II III Pre-SUDS Heart Rate Efficacy Prethoughts S k i l l .28 . 1 3 .55 .39 -.42 .72 .12 -.08 .49 -.03 -.49 .42 -.12 . 1 1 -.74 .66 .05 . 1 1 -.53 .51 -.07 .41 -.57 .63 -.28 .34 -.59 .73 -.19 .27 -.64 .44 -.09 Heart Rate Efficacy Prethoughts S k i l l .17 .34 -.23 .44 -.26 .03 .02 .19 .14 .18 .28 .09 -.11 . 1 6 .46 -.28 .31 -.27 -.21 .38 -.36 -.04 -.11 .09 Efficacy Prethoughts S k i l l .01 -.30 -.12 -.44 -.10 -.53 -.10 -.66 -.09 -.39 .38 -.42 .32 -.46 . 1 3 Prethoughts S k i l l -.08 -.08 -.08 -.13 -.10 -.19 I = Total Sample II = Return of Fear III = No Return of Fear 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0096539/manifest

Comment

Related Items