Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Social anxiety and memory deficit for information about others Biggs, Edward Eugene 1985

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

Item Metadata


831-UBC_1985_A2 B53.pdf [ 8.19MB ]
JSON: 831-1.0054303.json
JSON-LD: 831-1.0054303-ld.json
RDF/XML (Pretty): 831-1.0054303-rdf.xml
RDF/JSON: 831-1.0054303-rdf.json
Turtle: 831-1.0054303-turtle.txt
N-Triples: 831-1.0054303-rdf-ntriples.txt
Original Record: 831-1.0054303-source.json
Full Text

Full Text

SOCIAL ANXIETY AND MEMORY DEFICIT FOR INFORMATION ABOUT OTHERS by EDWARD EUGENE BIGGS M.A., San Franscisco State College, 1967 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTORATE OF EDUCATION in THE FACULTY OF GRADUATE STUDIES (Department of Counselling Psychology) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA October 1985 © EDWARD EUGENE BIGGS, 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 it 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 is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. The University of British Columbia 1956 Main Mall Vancouver, Canada Department of V6T 1Y3 DE-6(3/81) ABSTRACT Cognitive factors have been identified as c r i t i ca l variables in the origin and maintenance of interpersonal dysfunction associated with high social anxiety. Although evidence of a memory deficit accompanying general anxiety states i s abundant, studies of memory accompanying social anxiety have failed to demonstrate a def ic i t . Previous studies of memory deficit in social anxiety have measured only retention of evaluative feedback, the present study investigated memory more typical of interpersonal encounters, the recall of information about others. Forty-eight high socially anxious males and forty-eight low anxious males were asked to l isten to a tape recording of self-disclosures either during an interaction with the self-discloser or in private. Following an interim task, each subject was then asked to recall the information from the tape either in the presence of the female sel f -discloser or in private. This design allowed for social anxiety provoking manipulation at encoding to be completely crossed with social anxiety manipulation at retrieval. Multiple measures of memory were taken and analyzed with a multivariate procedure. It was hypothesized that a situational deficit would occur for the high socially anxious subjects when they were encoding the other-referent information in a social context. Additionally, i t was hypothesized that high socially anxious subjects would recall more affective as opposed to neutral information, and more negative items than positive or neutral. The results confirmed that memory is disturbed for high socially anxious subjects when in a social context, and specifically the disturbance occurs at the encoding phase. Results regarding the recall of affective material were contrary to prediction and suggest that high socially anxious subjects selectively process less affective material than do low socially anxious subjects. The results are supportive of a cognitive perspective arguing that dysfunctional interpersonal experiences may stem from impoverished, incomplete, and barren schema that guide the social behavior. The presence of a recall deficit along with intact recognition memory suggests that information about others is attended to but not processed 'deeply' or elaborately enough to be available on a free recall basis. The identification of memory deficit as a component of social anxiety provides a variety of new intervention possibilities including social memory enhancement programs, interventions aimed at unearthing poorly encoded memories, and strategies focussed on attention to affective messages. i i i TABLE OF CONTENTS Page Abstract i i List of Tables v i i List of Figures ix List of Appendices x Acknowledgements x i CHAPTER 1: SHYNESS AND SOCIAL ANXIETY 1 Social Anxiety As A Cl inical Problem 1 Defining Shyness and Social Anxiety 3 Theoretical Frameworks for Social Anxiety 8 Discriininating Measures in High and Low Social 10 Anxiety The Promise of Cognitive Approaches to Social Anxiety 15 CHAPTER 2: MEMORY AND SOCIAL ANXIETY 18 The Social Impact of Memory Deficits 18 Anxiety and Memory 20 Theoretical Formulations Predicting Memory Deficits With Heightened States of Anxiety 22 The Need to Identify Where in the Information Processing Sequence the Memory Deficit Occurs 24 Encoding Processes in Anxiety Related Memory Deficits 27 Attention and Encoding in Social Anxiety 29 Evidence of Memory Deficit in Social Anxiety 30 The Role of Affect in Memory Performance 36 iv Experimental Questions Emerging from the Review Regarding the Impact of Social Anxiety on Memory 38 The Scope of the Present Study 39 CHAPTER 3: METHODOLOGY 40 Design 40 Subjects 45 Confederate 47 Procedure 49 Instrumentation 54 CHAPTER 4: PRELIMINARY RESULTS 62 Development of the Social Recall Stimulus 62 Description of the Subject Population 75 Determination of Subject Equivalence with Regard to SAD Scores, Age, High School GPA, and Confederate Effect 78 Pre and Post Measures of Self-Report Anxiety 86 Coding and Judging of Recall Responses 92 Analysis of Item Serial Order Effect 94 CHAPTER 5: EXPERIMENTAL RESULTS 96 MANOVA I: Analysis of Recall and Recognition Measures 98 MANOVA II: Analysis of Affective Content Recall 106 Analysis of Total Affective Recall Percentage 108 Summary of Results 112 CHAPTER 6: DISCUSSION OF RESULTS 114 Memory Deficit 114 Encoding or Retrieval 115 v Attention vs. Depth of Processing 118 Recall vs. Recognition 119 The Impact of Social Context 124 Affective Messages 125 Memory Deficit , Performance Deficit or Cognitive Distortion 130 Cl inical Implications 134 The Role of Memory in the Treatment of Social Anxiety 136 Memory Enhancement Intervention: Possible New Directions for the Treatment of Social Anxiety 139 Directions for Further Research 144 Conclusion 145 Footnotes 147 References 148 Appendices 159 v i LIST OF TABLES Page 1. Means, Standard Deviations and Rater 71 Consensus Percentages for Items Selected as Negative, Positive, and Neutral for the Social Recall Stimulus Array 2. Means, Standard Deviations, and Comparison 73 Tests for the Imagery Ratings of the Positive, Negative, and Neutral Items Used in the Social Recall Stimulus Array 3. Means, Standard Deviations, and Quartile 76 Ranges of the Social Avoidance and Distress Scale Scores for the Volunteer Subject Population, Reported for the Fal l 1983 Intake, Spring 1984 Intake, and the Intakes Combined 4. Means and Standard Deviations of Social 79 Avoidance and Distress Scale for Both High and Low Anxiety Groups in A l l of the Encoding x Retrieval Treatment Combinations 5. ANOVA Results for the Social Avoidance and 80 Distress Scale Scores of High Anxious Subjects (Encoding x Retrieval) 6. ANOVA Results of Social Avoidance and 80 Distress Scale Scores of Low Anxious Subjects (Encoding by Retrieval) 7. Means and Standard Deviations of Subject Age 82 for Anxiety Group by Encoding by Retrieval 8. Means and Standard Deviations for Subject's 83 High School Grade Point Average for Anxiety Group by Encoding by Retrieval 9. Means and Standard Deviations of Recall 85 Totals for Anxiety Group by Encoding Condition by Retrieval Condition as a Function of the Confederate Factor 10. Pre and Post Anxiety Ratings of High and Low 88 Socially Anxious Subjects for Each Encoding by Retrieval Treatment Combination (Means and Standard Deviations) v i i 11. Evaluation of Judges' Acceptability Rating 93 Reported by Anxiety Group for Each Treatment Combination 12. The Effects of Items Order on the Recall 95 Performance of High and Low Socially Anxious Subjects by Treatment Condition 13. MANOVA I: Analysis of Anxiety Group by 98 Encoding Condition by Retrieval Condition Effects on Three Dependent Measures, Total Recall, Total Recognition and Total False Positives 14. Standardized Discriminant Coefficients 100 for the Statist ical ly Significanat Main and Interaction Effects Reported in MANOVA I 15. Computed Mean Standardized Discriminant 102 Scores for the Statist ical ly Significant Interactions Reported in MANOVA I Using the Discriminant Coefficients from Table 15 16. Cell Contrasts and Significance Estimates 104 for the Encoding x Anxiety Group Interaction and the Encoding x Retrieval Interaction Using the Computed Standard Discriminant Means 17. MANOVA II: Analysis of Anxiety Group x 107 Encoding Condition x Retrieval Condition Effects on the Three Affective Dependent Measures (Percentage of Positive, Negative, and Neutral Items Recalled) 18. 3-Way ANOVA for the Percentage of Affective 108 Items (Positive plus Negative) Recalled by High and Low Socially Anxious Subjects as a Function of Encoding and Retrieval Conditions 19. Means for Percentage of Affective Recall 110 Reported for the Significant Encoding by Retrieval Interaction v i i i LIST OF FIGURES Page 1. Schematic of the 4-way Factorial Design 41 2. Plotting of Discriminant Means for the 103 Encoding by Anxiety Group Interaction and the Encoding by Retrieval Interaction 3. Graphing of the Computed Estimates for 111 the Encoding x Retrieval Interaction of the 3-way ANOVA for Percentage of Affective Recall ix LIST OF APPENDICES Page A. Social Recall Stimulus List 159 B. Recognition Form 160 C. Recruitment Leaflet 162 D. Guidelines for the Confederate Interaction 163 E. Background Information Sheet 164 F. Consent Form 165 G. Instructions for Private Encoding 166 H. Instructions for Private Retrieval 167 I. Social Behavior Scale (the Social 168 Avoidance and Distress Scale) J . STAI Self-Report Questionnaire 169 (pre and post forms) K. Bett's QMI Visual Imagery Scale 171 L. Recall Form 173 M. Means and Standard Deviations for 176 the Three Dependent Measures, Total Recall, Total Recognition and Total False Positives for Anxiety Group by Encoding Condition by Retrieval Condition N. Means and Standard Deviations 177 for the Mean Percentage Positive, Negative, and Neutral Recall for Anxiety Group by Encoding Condition by Retrieval Condition x ACKNOWLELXJEMENTS I wish to express my affection and gratitude to those individuals whose effort helped produce this document. Special thanks goes to my research supervisor, Dr. Leslie Greenberg, a gifted cl inician and an inspiring theoretician. Thanks also to the members of my committee, Dr. Sharon Kahn, Dr. Rich Young, and Dr. Walter Boldt, for their continuing support and careful consideration of my writing. I wish also to identify the individuals who acted as 'confederates' in my study, Kim Redding, Dale Hunter, Ann Parfitt-Lewis, and Nina Vivieros, for very special thanks. Finally, I am profoundly grateful to my loving family, Deborah, Tollef, and Zannie, for their extended patience, nurturance and support. x i CHAPTER I. SHYNESS AND SOCIAL ANXIETY Social Anxiety As A Cl inical Problem Distress in social interactions has been documented widely as a significant c l in ica l problem (Argyle & Kendon, 1967; Bandura,1969; Borkovec, Stone, O'Brien, & Kaloupek 1974; Lawlis,1971; Martinson & Zerface, 1970; Pilkonis, 1977 Zimbardo, 1977). Arkowitz, Hinton, Perl and Himadi (1978) found that one third of their college population reported being 'somewhat' or 'very' anxious about dating. Martinson and Zerface (1970) report that a significant number of adolescents suffer discomfort or failure in dating interactions. Glass, Gottman and Schmurak (1976) report that dating constituted over half of the situations tagged by undergraduate men as a source of d i f f icul ty . Social deficits have long been linked to pervasive psychological d i f f icu l t ies . Argyle, Trower and Bryant (1974) see personal isolation and inadequate social s k i l l s as both causes and consequences of emotional problems. Lawlis (1971) documents the prevalence of shyness in patient populations. Kraft (1971) indicates that one out of every four psychiatric patients are judged to be socially inadequate. Shyness in the extreme can be a pervasive, complex and debilitating anxiety reaction. Borkovec, Stone, O'Brien and Kaloupek (1974) 1 indicate that interpersonal performance anxiety is frequently accompanied by strong physiological arousal which i s slow in returning to baseline levels. Disturbances in sexual activity (Bandura, 1969) distortions in memory (O'Banion & Arkowitz, 1977) and a higher incidence of negative self-statements (Cacioppo, Glass, & Merluzzi,1979) also have been indicated in populations exhibiting high social anxiety. Zimbardo (1977) in his extensive survey (N = 5000) of shyness at Stanford University found that 40% of the university population described themselves as shy. Pilkonis (1977) validated this finding in his study, and also ascertained that 24% of the self-reported shy sample view their shyness as intense enough to seek help in overcoming this problem. A consistent response pattern of avoidance and experience of anxiety in the face of social encounters can lead to a serious disruption of daily functioning. In extensive interviews with shy people, Zimbardo has documented the pain, inhibition, and self-defeating spiral which shyness can engender. Lowered social and vocational achievement, decreased likelihood of interpersonally satisfying relationships, limited availabil i ty to positive feedback from others, lowered cognitive functioning, depression, isolation, and chronic dissatisfaction are some of the debilitating consequences of continued social awkwardness, distress, and withdrawal. In an attempt to better understand this widely shared social disabil i ty , the following sections w i l l attempt to define shyness and 2 social anxiety more clearly and w i l l present the prevailing theoretical frameworks which have been used to conceptualize the phenomenon. Finally, new directions in the area of cognitive functioning which appear to be promising for the understanding and treatment of social anxiety w i l l be identified. Defining Shyness and Social Anxiety The construct of social anxiety can be traced to research in psychopathology and to the search for individual differences in personality research. Rodnick and Garmezy (1957) demonstrated the negative effects of social censure on the perceptual-motor performance of psychotic patients. Zigler and Phi l l ips (1962) demonstrated social s k i l l s deficits as a common deficiency in hospitalized psychiatric patients which they termed "social incompetence". The indication that individuals who are not patients show differing degrees of anxiety in social situations has also been demonstrated ( Byrne, McDonald, & Mikawa, 1963; Diggory, 1966; Sears, 1967). The development of scales to measure social-evaluative anxiety (Social Avoidance and Distress Scale and the Fear of Negative Evaluation Scale) by Watson and Friend (1969) established social anxiety as a well-validated, highly homogeneous construct. Watson and Friend define social-evaluative anxiety "as the experience of distress, discomfort, fear, anxiety, etc. , in social situations; as the deliberate avoidance of social situations; and f inal ly as a fear of receiving negative evaluations from others.". This scale has 3 received repeated validation as an accurate discriminator of high and low socially anxious populations (see instrument validation section below). In an attempt to refine the nature of the social anxiety experience, Richardson and Tasto (1976) factor analyzed 166 items drawn from Behavior Therapy treatment hierarchies used with socially anxious clients. Their analysis revealed seven factors; (1) fear of diasapproval or crit icism from others, (2) fear of social assertiveness and social v i s ib i l i t y , (3) fear of confrontation and anger expression, (4) fear of heterosexual contact, (5) fear of intimacy and interpersonal warmth, (6) fear of conflict and rejection by parents, and (7) fear of interpersonal loss (experienced as jealousy or rejection by mate). This identification of factors and resulting inventory, the Social Reaction Inventory (SRI) goes beyond the general t rait social anxiety scales developed by Watson and Friend, and identifies more specific social stimuli that trigger the experience of anxiety. This approach i s consistent with Mischel (1968) and others (Leary,1983; Pilkonis,1977) who contend that for many people the experience of social anxiety i s situation specific. It i s of interest to note, however, that measures of state and t ra i t social anxiety correlate highly (Leary,1983). Factor one of the SRI is similar to the Fear of Negative Evaluation scale of Watson and Friend. Factors two through seven can be seen as specific amplifications of Watson and Friend's more general Social Avoidance and Distress scale. It i s interesting to note that in addition to identifying the anticipation of specific events as triggers for 4 anxiety (demands for assertiveness, confrontation, v i s ib i l i t y ) this scale also identifies fantasy expectations (fear of intimacy, rejection by loved one) which are more notably internal events, as catalysts for anxiety and subsequent rationales for avoidance behavior. A third attempt to delineate the factors underlying intense social distress comes from the work of Pilkonis (1977) and Zimbardo (1977) at the Stanford Shyness Cl in ic . They define shyness as "a tendency to avoid social interactions and to f a i l to participate appropriately in social situations". Although this definition requires stipulation of what constitutes "appropriate social participation", the emphasis is clearly a behavioral one stressing an individual's response tendencies in reaction to fe l t social awkwardness or anxiety. Pilkonis (1977) asked shy subjects to rank order several different aspects of social anxiety which had previously been incorporated in social anxiety inventories. The aspects are: (1) internal discomfort experienced in social situations, (2) fear of negative evaluation, (3) avoidance of social situations, (4) failure to respond appropriately in social situations (relative to talk, avoidance of eye contact), (5) awkward behaviors arising from attempts to respond (e.g. inabil i ty to articulate or be fluent, physical clumsiness). These categories were to be ranked according to how much of a problem this aspect of their shyness posed. The results showed that shy people were most bothered by behavioral deficits ( items 4 and 5) and ranked the avoidance of social situations as the least severe consequence of their shyness. This study i s valuable for several reasons. For one, the results underscore the notion that the 5 embarrassment and humiliation arising from public failure i s the most punishing aspect of social anxiety. Secondly, i t indicates that avoidance of social contact i s a relatively minor inconvenience as seen by the shy person. Pilkonis points out that this i s quite paradoxical, as the withdrawal and avoidance of social contact may well be the most sinister aspect of the shyness syndrome. In the extreme, the agoraphobic may not be able to tolerate any social contact without experiencing marked anxiety which, in turn, leads to escalating isolation. Gradual social withdrawal i s also diagnostic of pre-psychotic conditions. The absent, quiet, non-visible aspects of shyness have contributed to i t s status as a non-problem c l in ical ly in our culture. It i s only recently that the potentially paralyzing effects of intense social anxiety have been articulated. To further complicate the definition of this phenomenon, Zimbardo (1977) also identifies several aspects of shyness which both the shy individual and those interacting with him or her perceive as desirable or positive traits (eg. modest, unassuming, discreet). It i s evident that the constructs of shyness and social anxiety remain conceptually blurred. This i s a condition which has continued to be problematic even for those investigators who operationally define the domain as 'social s k i l l s deficits ' (Hersen & Bellack, 1977). Mark Leary (1983) has reviewed the literature attempting to delineate social anxiety and shyness, and concludes that the " . . . l i s t of social anxiety items is nearly identical to the l i s t of shyness-producing situations obtained in Zimbardo's (1977) extensive survey research". Leary asserts that the quality of these experiences which 6 set them apart from other anxiety-provoking circumstances is the dimension of 'interpersonal evaluation'. Leary (1983,p.23) has recently abandoned his view that shyness and social anxiety are interchangeable. He prefers to view shyness as a response syndrome containing a variety of experiential and behavioral characteristics one of which is social anxiety. He believes that this shyness syndrome has failed to emerge as a unitary construct as i t has been used by various investigators. Leary refers to social anxiety as the relatively consistently described subjective experiences of apprehension and nervousness, and shyness refers to a variety of responses to that apprehension. Descriptions of dating anxiety, public speaking anxiety, and heterosocial anxiety are best seen as situationally specific sub-sets of general social evaluative anxiety; accordingly, specific s k i l l s deficits refer to the absence of particular behaviors (assertiveness, dating) which are sub-sets of social s k i l l s . Schlenker and Leary (1982) offer an impression management model of social anxiety which identifies two c r i t i ca l components of the experience: (1) an individual's motivation to make a particular impression on others, and (2) the degree to which that individual doubts that they w i l l be successful in making the desired impression. This conceptualization of social anxiety i s largely a cognitive formulation. Distress arises from an individual's expectations about their own performance and from the assessment of the likelihood their expectations w i l l be f u l f i l l e d . For the purposes of this study social anxiety has been adopted as the focus of investigation and the term shyness w i l l be reserved as a more impressionistic descriptor 7 which w i l l refer to the problematic or undesirable dimensions of shyness only. Theoretical Framework For Social Anxiety Treatment programs logically flow from the conceptualization of the factors responsible for the origin and maintenance of the condition. One position, historically, suggests that shyness or "threctia" i s a stable dispositional t rait (Cattell , 1973), and this carries with i t the possibil ity of genetic determination. Although several twins studies (Freedman, 1965; Gottesman, 1962; Shields, 1962) have demonstrated greater similarity in sociability between identical twins as compared to fraternal twins, considerable doubt exists with regard to this disposition accounting for extremes in adult social anxiety. The on-going social learning history of any chi ld, coupled with the evidence for effective remediation of social anxiety, argue for the acquired and therefore reversible nature of social inhibition. Curran (1977) describes the three major approaches which attempt to account for the development of social anxiety: the conditioned anxiety model, the social s k i l l s deficit model, and the cognitive-evaluative model. In Curran's review of the s k i l l s deficit approach to one dimension of social anxiety, heterosexual social anxiety, he acknowledges that each of these three models has demonstrated effectiveness in treating aspects of social anxiety: 8 There appears to be evidence supporting a l l three concepts regarding the origin and maintenance of heterosexual social anxiety. This is not unusual because most disordered behavior tends to be multidimensional upon inspection and multiply determined in origin. (Curran, 1977, p. 143.) With regard to the sk i l l s deficit hypothesis in particular, Curran finds reason for optimism after reviewing the available research, although his optimism is tempered by the prevalent methodological problems and inconsistent results. A study by Christensen, Arkowitz and Anderson (1975) exemplifies Curran's position. In this study 30 male and female volunteers participated weekly in a practice dating program lasting six weeks. Three groups were established: practice dating with feedback, practice dating without feedback, and a wait-l i s t control group. Subjects were matched by scores on social anxiety inventories (SAD, SRAI). Results indicated that both of the treatment groups differed significantly at the end of the program from the wait-list controls on the self-report measures of anxiety, on the frequency of dating, on the number of conversational silences, and on pulse rate. There was no difference in social anxiety or social s k i l l behaviors as perceived by partners or observers. Curran is encouraged by the results here, but guarded concerning their validity due to inadequacies in methodology. As Curran points out, the study f a i l s to adequately screen subjects to insure symptom severity; f a i l s to assess a l l groups at the pre-test phase; and f a i l s to employ adequate follow-up procedures. 9 The Christensen, Arkowitz and Anderson study is characteristic of several of the studies dealing with the training of social s k i l l s (Glasgow & Arkowitz, 1975; Rehm & Marston, 1968) in finding significant changes in self-report measures of social anxiety, but finding no differences in observable behaviors. Finding clear differences between shy and not-shy populations would be a valuable step forward in conceptualizing the dysfunctional processes which underlie this condition. The following section reviews the experimental pursuit of these elusive differences. Discriminating Measures in High and Low Social Anxiety Several authors have found that self-report measures of social anxiety differentiate low frequency daters from high frequency daters (Arkowitz, Lichtenstein, McGovern, & Hines, 1975; Glasgow & Arkowitz, 1975); high anxious from non-anxious (Borkovec, Stone, O'Brien, & Kaloupek,1974; Caldwell, 1978); and shy from not-shy (Pilkonis, 1977; Twentyman & McFall, 1975). In addition, observer judges and partner ratings of subjects social s k i l l s and level of social anxiety have proved to be predictive of social anxiety level (Curran, 1975; Curran & Gilbert, 1975; Twentyman & McFall, 1975) but not consistently (Christensen, Arkowitz & Anderson, 1975). Isolating consistent behavioral measures which differentiate high and low social anxiety, however, continues to be elusive. Borkovec, Stone, O'Brien, and Kaloupek (1974) found overt signs of anxiety as rated on a timed behavioral checklist to be non-discriminating, but pre-test heart rate measures did distinguish socially anxious from 10 non-anxious subjects. Twentyman and McFall (1975) also found that pulse rate during analogue heterosexual interactions was significantly lower for confident subjects. In addition, shy subjects following treatment were much more like confident subjects on scores of self-reported anxiety and their interaction performance. Glasgow and Arkowitz (1975), attempting to delineate micro behaviors which correlate with social anxiety, orchestrated simulated interactions between high and low frequency daters and opposite sex partners. They collected self and partner ratings of the subjects' social s k i l l s and anxiety level as well as observer ratings of gazing time, eye contact, total talk time, number of silences, and initiations of interactions. The self and partner ratings were discriminating for the two groups, but the behavioral measures failed to demonstrate predictive power. A study by Arkowitz, Lichtenstein, McGovern and Hines (1975), similarly attempting to define behaviors associated with social competence, measured a variety of interactional dimensions (smiles, nods, talk time, etc.) as well as coding and rating conversational content. Of the eight dimensions tapped, only the number of conversational silences differentiated low and high dating populations. Content rated quality of self-disclosure was not significant. Craighead, Kimball, and Rehak (1979) using female subjects who scored high or low on the Social Approval scale of the Irrational Beliefs 11 Test (a test devised to tap the i r r a t i o n a l bel ief systems described by Albert E l l i s ) compared scores of self-reported distress , physiological arousal, and frequency of negative self-statements. Only the frequency of negative self-statements reached significance i n d i f f e r e n t i a t i n g the two groups. This f inding i s contrary to a previous study by Goldfried and Sobocinski (1979) who found s ignif icant changes i n arousal and self-reported anxiety as well as the self-statement difference under s imilar s i tuat ional demands. Both experiments measured responses to imagined scenes of soc ia l reject ion, and while there i s v a r i a b i l i t y i n arousal measurement, both studies ident i fy the frequency of negative self-statements as a s ignif icant dimension for groups who d i f f e r i n their constant need for soc ia l approval. Trower (1980) rated groups of high anxious and low anxious non-psychotic hospital patients on behavioral dimensions s imilar to those used by Arkowitz, Lichtenstein, McGovern, and Hines (1975). Trower found that s o c i a l l y s k i l l e d patients spoke, looked, smiled, gestured, and moved their body posture more than unski l led patients. Trower's success i n obtaining significance on these measures where other investigators have not i s important, although the comparability of the population he used i s i n question. As hospitalized patients, the potent ial ly confounding effects of medication, length of stay i n hospi ta l , and diagnostic c l a s s i f i c a t i o n of the subjects are c r i t i c a l issues and are not addressed i n the study. I t may be that the extremes i n social s k i l l s found i n populations requiring hospi ta l izat ion are much broader than the extremes of subjects drawn from functioning sett ings. Trower also indicates that speech cues 12 are the most prominent d e f i c i t s i n es t a b l i s h i n g impressions of s o c i a l incompetence. This f i t s with a p o s i t i o n which characterizes the problem of shyness as an i n a b i l i t y to i n i t i a t e speech; however, as language i s the most v i s i b l e manifestation of thought, t h i s d e f i c i t may be meaningfully seen from a cognitive perspective as a crude i n d i c a t i o n of cognitive disorganization. P i l k o n i s (1977) asked self-described shy and not-shy persons to inter a c t i n a staged opposite-sex i n t e r a c t i o n which included a p u b l i c speaking exercise. Each subject was rated by the confederate, an observer, and the experimenter on a v a r i e t y of verbal, non-verbal and a f f e c t i v e dimensions. A discriminant analysis produced a weighted formula f o r p r e d i c t i n g shyness with a c o r r e l a t i o n of .84. The formula included the following variables i n order of importance: judged speech anxiety, percentage of time t a l k i n g , number of smiles, number of questions asked the experimenter, a f f e c t rated at time one, breaking of s i l e n c e s , judged speech q u a l i t y , self-manipulation, head nods, and physical attractiveness. Curiously, self-reported anxiety was not a predictor, nor was the percentage of time looking, a f f e c t rated at time two, the number of gestures made, or the distance the subject was from the experimenter. In the research attempting to i d e n t i f y observable behavior associated with s o c i a l anxiety a s t r i k i n g lack of consistency emerges. As noted elsewhere (Kupke, Hobbs, & Cheney, 1979; Twentyman & McFall, 1975) there are s u r p r i s i n g l y few overt behaviors which seem to d i s t i n g u i s h t h i s group. With regard to s o c i a l anxiety, s e l f - r e p o r t measures are 13 consistently the best predictors; occasionally arousal measures have been discriminative; and infrequently are observed behaviors predictive. V/hen such observable differences are available, as with the frequency of conversational silences, they contribute l i t t l e to our understanding of the complex multidimensional experience of social anxiety which would be of value in developing treatment interventions. It i s important to note as well that the studies which have reported significant changes in behavioral measures as well as self-report measures (Curran, 1975; Curran & Gilbert, 1975; Curran, Gilbert, & L i t t l e , 1976; Twentymen & McFall, 1975) tend to involve much more intensive training programs using a wide variety of overlapping experiences. Curran's (1975) program included: didactic information, modelling, behavioral rehearsal, coaching, video and group feedback, and ' in vivo' assignments. It could be argued that a massive program such as this offers opportunity for change at many levels and conceivable for many reasons, and therefore does not offer evidence exclusively for a s k i l l s deficit model. This kind of experience may also provide a major assault on belief systems, offer alternative ways of viewing heterosexual contact, involve esteem building ego support, confront c l ient 's irrational ideas, immerse the subject implosively in sensitized surroundings, or encourage approach to previously avoided situations. A study by Christensen and Arkowitz (1974) demonstrated that anxiety could be reduced in inhibited subjects by the mere exposure to dating situations. That i t i s the intensive multifaceted programs which produce behavioral change suggests that perceptual and attitudinal change may be the c r i t i ca l components in successful interventions. This issue as to 14 whether or not Behavior Therapy programs are actually cognitive in nature has spawned l ively debate in recent years (Ledwidge, 1979; Locke, 1979; Meichenbaum, 1979). The Promise of Cognitive Approaches to Social Anxiety Cognitive processes have been both directly and indirectly implicated as c r i t i ca l variables in the mediation of social anxiety; the results from cognitive treatment programs as well as non-treatment sources are promising in this regard. Successful treatment of conditions similar in many ways to social anxiety have been reported using cognitive interventions; with speech anxious subjects (Meichenbaum, Gilmore, & Fedoravicious, 1971; Trexler & Karst, 1972), with test anxiety (Goldfried, Linehan, & Smith, 1978; Meichenbaum, 1972), with unassertive behavior (Thorpe, 1975; Wolfe & Fodor, 1978), and with clients experiencing heterosexual anxiety (Glass, Gottman, 8c Schmurak, 1976). The findings of Goldfried, Linehan, and Smith (1978) that reductions in test anxiety generalized to social-evaluative situations, suggests an underlying conmonality between these conditions and social anxiety, and that these sets of experiences are amenable to cognitive strategies. The implication of cognitive factors in social anxiety i s indirectly suggested from a number of sources. Phi l l ips (1980) using a s k i l l training approach to address what he termed a 'reticence syndrome' found that one third of the shy subjects increased their anxiety 15 subsequent to the s k i l l s acquisition. This suggests that the behavioral deficit may not have been the source of anxiety and the increased expectancy to perform may have been the c r i t i ca l variable aggravating the level of distress. Goldfried and Sobocinski (1975) found a significant correlation between self-report measures of social anxiety and the presence of irrational beliefs as measured by the Irrational Beliefs Test. O'Banion and Arkowitz (1977) demonstrated that high socially anxious subjects recall more negative information regarding themselves than do low socially anxious subjects. Smith and Sarason (1975) found that high socially anxious subjects interpret feedback regarding themselves less favorably than do low socially anxious subjects, and they have a greater expectancy that others w i l l evaluate them negatively. This i s consistent with Clark and Arkowitz (1975) and Glasgow and Arkowitz (1975) who found that socially anxious male subjects as compared to non-anxious males underestimate their own performance and expect greater negative evaluation from others even though their s k i l l s were not seen to be inferior by independent judges. These results have led some investigators to conclude that social anxiety is not a deficiency in the social s k i l l s repertoire, but rather results from mediational processes: selective attention, biased encoding, selective information retrieval, distortions in reasoning, and negative self-evaluation. As mentioned earlier, Schlenker and Leary's impression management model identifies 16 expectation of failure as central to the creation of social apprehension. Additionally, Galassi and Galassi (1978) conclude, in their review of social s k i l l s modification programs, that "there i s more evidence that heterosocial problems are characterized (not necessarily caused) by deficits in cognitive rather than behavior s k i l l s . " The cautionary note here i s well advised, for as promising as the results of cognitive treatment programs are, and, as suggestive as the parallel literature i s , as Cacioppo (1979) points out, they only provide "indirect support for the role of cognition in the development and maintenance, and el ic i tat ion of heterosocial anxiety." Although research indicates that cognitive differences are more discriminating than behavioral ones, the role of cognition in shyness requires a more detailed, systematic, and thorough exploration. With the goal of delineating the crucial cognitive variables underlying social anxiety more closely, the following chapter w i l l examine the role of memory in social interaction. Memory w i l l be identified as a fundamental cognitive process implicated in the dysfunctions associated with social anxiety, and a cognitive process which has received relatively l i t t l e attention in relation to social anxiety. 17 CHAPTER II. MEMORY AND SOCIAL ANXIETY The Social Impact of Memory Deficits Memory functions are central to the array of information processing events implicated in the experience of interpersonal distress. Mischel (1973) presents a taxonomy of cognitive social learning variables which he sees as essential for interpersonal competency. His system identifies individual differences in attentional capacity, cognitive style, encoding processes, and retrieval strategies as central processes which are, in turn, influenced by each individual's unique values, interests, needs, personal constructs, expectations and goals. A similar set of cognitive events involved in social encounters has been described by Markus (1980). At the core of these systems i s the need for attentional s k i l l s , s k i l l s in sampling and symbolizing one's experience, and s k i l l s in accessing previous personal and social experiences. Social competence i s inextricably bound with social memory. The significance of memory in social interactions becomes apparent when the components of satisfying interactions are examined. Communication s k i l l s are essential in achieving and maintaining interpersonal contact, and communication s k i l l s are dependent on social memory. 18 The a b i l i t y to spontaneously s e l f - d i s c l o s e has been i d e n t i f i e d as a process which deepens intimacy. S e l f - d i s c l o s u r e i s contingent on an i n d i v i d u a l ' s a b i l i t y to access autobiographical memory quickly, and i n a v a r i e t y of circumstances. In addition, an i n d i v i d u a l ' s a b i l i t y to demonstrate att e n t i v e l i s t e n i n g i s dependent on h i s or her a b i l i t y to o f f e r accurate and relevant interpersonal feedback. Providing feedback i s a function which draws d i r e c t l y on e f f i c i e n t memory s k i l l s . The i n a b i l i t y to o f f e r t h i s feedback may be perceived as lack of i n t e r e s t or lack of empathy by the other person, and i s l i k e l y to hinder the development of intimacy. On a more concrete l e v e l , chronic i n a b i l i t y to remember faces or names contributes to f e e l i n g s of awkwardness and embarrassment f o r everyone involved. S a t i s f y i n g s o c i a l discourse requires continual cross-checking of personal and s o c i a l information without t h i s , there i s confusion and a pervasive lack of s o c i a l coherence. People who are remembered by others f e e l f l a t t e r e d and reinforced; i n d i v i d u a l s with poor s o c i a l memory lack t h i s powerful r e i n f o r c i n g t o o l . Individuals who^ are capable of making coherent connections between the present and previous information are l i k e l y t o be seen as more engaging conversationalists than those f o r whom few connections are a v a i l a b l e . Clear memories of one's own past are at the core of personal i d e n t i t y . Individuals with constricted or incomplete contact with t h e i r previous thoughts and experiences may s u f f e r from greater 19 identity diffusion, a state which Erikson (1968) associates with confusion, conformity and emotional vulnerability. Carl Rogers (1951) puts this notion of a well-articulated self-system at the core of his concept of the healthy personality. And f inal ly , consider the social handicap of individuals with chronic memory deficits: mild cases in which individuals who tend to unknowingly repeat themselves are often experienced as tedious; more severe amnesic conditions found in dissociative disorders and organic brain syndromes are tragic and incapacitating. Memory would appear to be an important factor in social interacting, but, i s there a memory deficit in individuals who experience high social anxiety? There i s very l i t t l e research which addresses this question specifically, yet the expectation of a deficit i s clearly available from related sources. The following section reviews some of the evidence for memory dysfunctions occurring with heightened levels of non-specific anxiety. This research frequently involves manipulation of the anxiety level of normal subjects in laboratory settings. This is a situation which i s clearly different from naturally occurring, situation-specific anxiety, but the results are instructive and suggestive in the absence of more specifically relevant research. Anxiety and Memory Reviews of the impact which anxiety exerts on memory and learning (Dutta & Kanungo,1975; Eysenck,1977; Levitt,1980) document the 20 reality of this relationship. As Mueller (1976) indicates the negative effects of anxiety on memory are so commonly assumed that memory deficits seen in intelligence testing have become diagnostic of anxiety. Memory deficits resulting from anxiety have been investigated using a variety of dependent and independent measures including eyelid conditioning under conditions of applied electric shock (Spence, 1964), paired-associate learning using electric shock (Spence, Farber, & McFann, 1956), recall dysfunctions with high test anxious subjects (Sarason, 1972), reconstruction of eye witness events under stressful conditions (Siegel & Loftus, 1978) and free recall of autobiographical childhood memories with emotionally arousing content (Ansbacher, 1947). In addition, improvement in problem-solving abi l i t ies in high state anxious subjects has' been demonstrated by providing memory aids to these subjects (Gross & Mastenbrook, 1980). Additional support for a memory decrement associated with heightened states of stress comes from the considerable literature regarding the impact of heightened arousal on learning, although difference of opinion exists with regard to the validity of interpreting these studies as analogous to anxiety studies (Eysenck,1977). The memory/anxiety relationship i s widely documented, but theoretical formulations of the origin and the location of the predicted deficit (encoding,retrieval) differ greatly depending on the the theoretical framework endorsed. 21 Theoretical Formulations Predicting Memory Deficits With Heightened States of Anxiety Historically, the 'repression hypothesis', central to Psychoanalytic theory, was the earliest formulation of retrieval difficulties which stemmed from the anxiety-provoking content of these memories. Kihlstrom (1980) has reviewed the research which addresses the repression hypothesis and finds the impact to be ambiguous and inconclusive. Kihlstrom concludes that there is evidence which indicates selective memory for successful and completed experiences (Zeigarnik Effect) which is relevant to the repression model; however, Kihlstrom indicates that a more productive approach is to recast repression in information-processing terms. To this end, and consistent with other authors (Erdelyi & Goldberg, 1979; Mandler, 1975) Kihlstrom states that: ... a plausible redefinition of repression liberates i t from Freudian psychosexual theory and holds that individuals are capable of defensively biasing information processing functions so that threatening material available in both the perceptual field and memory are not represented in phenomenal awareness. (Kihlstrom, 1981, p. 130) Within a learning theory framework, Spence (1958) hypothesized that heightened drive states would increase the probability of the strongest response tendencies occurring; this would improve 22 performance when no competing response i s available and impair performance when competing responses are present. This theory has generally been supported using paired-associate learning tasks (Eysenck, 1977). Some investigators (Saltz, 1970; Weiner, 1972) find the Spence model oversimplified and have invoked more cognitive explanations involving the subject's anticipation of success or failure. Goulet and Massei (1969) identify a further confounding variable which they term 'confidence threshold'. Using a paired-associate task these investigators found that highly anxious subjects did not respond as early as did low anxious subjects, indicating that they may require a much higher level of internal certainty before coranitting themselves to a response. With regard to free recal l , Drive theory would argue that high anxiety raises more responses above threshold therefore i t would predict high anxious subjects to produce better recal l . This hypothesis has not been confirmed. Mueller (1976) and Zubrycki and Borowski (1973) both found low anxious subjects to be superior in free recal l . Andrew Mayes (1983) concludes his review of the impact of anxiety on memory with this summation: In sum, there i s no clear theoretical understanding of the effects of anxiety on learning and memory. It i s probable that anxiety i s more l ikely to reduce the efficiency of internal processing operations than i t i s to impair performance Surprisingly, what limited evidence there i s (for a memory deficit accompanying 23 high anxiety) indicates that anxiety affects primarily the i n i t i a l encoding of information; i t has usually been found that retrieval per se is unaffected by anxiety... (p.289) Mayes identifies the constriction of attention and impairment of encoding operations (most l ikely from the intrusion of irrelevant cognitions) as the primary effects which anxiety exerts on learning and memory. The Need to Identify Where In The Inforimtion-Processing Sequence the Memory Deficit Occurs The recognition or recall of previous information occurs relatively late in the information-processing sequence. Hastie and Carlston (1977) present a model for person memory which identifies the operations involved in the memory process. This model i s useful for i l lustrating the possible sites involved in memory dysfunction. Hastie and Carlston (1977) have identified six component substages which they see as the minimum elements necessary for an adequate model of person memory: 1. The parsing and encoding of information from the on-going behavioral stream. 2. The encoding transformations that convert raw perceptual information into mentally significant symbol structures. 3. A description of the symbols and relations that constitute the "mental language"in which social information i s represented and retained. 4. The events occurring during retention which render some information accessible and other 24 information inaccessible. 5. The decoding transformations that probe, search or integrate the memory structure. 6. Specification of the nature of decisions that are made when a response i s generated and performed. These cr i ter ia incorporate the traditional acquisition, retention, and retrieval processes, articulate specific cognitive domains within each, and include the behavioral implications which retrieval or failure to retrieve might engender. If the point in this memory sequence at which the deficit or distortion occurs can be identified, the c l in ica l usefulness of identifying memory di f f icul t ies would be significantly enhanced. Retrieval i s necessarily dependent on prior encoding efficiency, but retrieval d i f f icul t ies could also result from ful ly encoded memories which cannot be accessed on demand. Recently cognitive schema theorists (Hastie, 1980; Neisser, 1976) have re-emphasized the constructive and reconstructive quality of memory, and view errors of omission or comission in memory as flowing from incompatibility or compatibility of the information with existing guiding themes. In this view, r igid or poorly-formed schemata could constrict the way new information i s encoded or retrieved. The suggestion of possible schema deficits occurring in individuals low in social competency i s offered by Ostrom (1980): It may be that people with a chronic inabi l i ty to relate to others as individuals suffer from a deficit in their capacity to organize social information by person. If so, i t would be worthwhile to develop training programs to faci l i tate person organization, (p.33) 25 Other investigators interested in social anxiety have posited a preponderant negative self-schema as the biasing structure distorting information sampling, encoding, and retrieval. (Breck & Smith, 1983; Clark & Arkowitz, 1975; Smith & Sarason, 1975). When asking whether a memory deficit exists in socially anxious populations, i t becomes increasingly important c l in ical ly to ask, as well, whether the focus of that deficit i s occurring early in the sequence or later at the retrieval phase. Eysenck, in his review of "Anxiety, Neuroticism, and Memory", concludes that very l i t t l e effort has been made to address the point at which anxiety affects memory: An obvious approach would be to separate presentation and test t r ia ls in time, and to present stressful stimulation (e.g. electric shock, ego-threatening instructions) either at input or at output. According to the models of the Spence-Spence type, the effects of stress at output should be greater than at input. (Eysenck, 1977,p.217) Eysenck's suggestion here i s relevant and important i f this deficit is to be addressed therapeutically. Determining this aspect of the memory deficit would clarify the theoretical model most useful to the conceptualization of cognitive dysfunction in social anxiety, as well as faci l i tate the development of relevant treatment interventions. 26 Encoding Processes in Anxiety Related Memory Deficits The sequence of processes identified by Hastie and Carlston in their model for person memory is compatible with the emerging trend to identify the control processes as c r i t i ca l to efficient cognitive functioning. The conclusions of Mayes above as well as the refomulation of repression theory suggested by Kihlstrom lend support to this emerging perspective. Recently Greenberg and Safran (1980) have articulated an information processing model specifically designed for c l in ica l application. They have effectively adapted Neisser's conceptualization of a "perceptual cycle" which links perception (sampling and symbolizing processes) to cognition (transformational and relational processes). Their model identifies attending and encoding processes as central functions in adaptive and maladaptive disorders. The implication of encoding processes in the dysfunctional impact which anxiety exerts on memory also has been formulated by Mueller (1976). Mueller emphasizes the encoding process as the point in the information-processing sequence which i s vulnerable to state specific anxiety. He suggests that information sampling i s constricted in i t s accessing of available information (a position similar to the response availabil ity aspect of Drive theory) and i s , in turn, encoded in less meaningful ways. The f i r s t aspect of this model incorporates the cue-utilization hypothesis (Easterbrook, 1959) which posits a reduced sampling capacity; the second aspect of Mueller's 27 model ut i l izes the 'levels of processing' notion of Craik and Lockhart (1972) suggesting that high anxious subjects are less l ikely to encode material semantically (according to meaning) as opposed to structurally (according to physical attributes) and engage in less elaborative rehearsal. Mueller's own research involving the free recall of word triads which can be processed either in terms of rhyme or meaning lend tentative support to his hypothesis. Mueller's procedure involves presenting to anxious subjects a series of three words (e.g. TABIJE-CHAIR-SHARE) which can be encoded using a shallow feature (rhyme) or a deep feature (associative). Using this procedure the amounts of deep and shallow clustering i s interpreted as a measure of the depth of cue ut i l izat ion. Mueller (1976) has demonstrated much less clustering of both types with high anxious subjects. This result i s supportive of his prediction of deep processing deficiencies with high anxious states, but the deficiency in shallow processing as well , i s d i f f icu l t to interpret. It may indicate that high anxious subjects demonstrate diminished cue uti l ization as well as more shallow cue ut i l izat ion. Alternatively, i t may be that the process of rhyme encoding actually involves more elaborate encoding- than was i n i t i a l l y believed. This latter concern has led Nelson (1977) to argue that rhyme i s an inappropriate task to use as an alternate to 'depth' encoding. Despite the presence of some unexplained results here, Mueller's model does provide a coherent alternate perspective which combines Drive theory analysis with encoding depth, breadth, and specificity notions involved in both the acquisition and retrieval of memories. 28 Mueller's perspective i s congruent with the recent identification of encoding processes as particularly vulnerable sites for information-processing distortions to occur. Greenberg and Safran (1980) argue that the encoding phase has been largely ignored in cognitive therapies, that this phase i s amenable to perceptual retraining, and that this encoding phase may well be a more appropriate intervention focus for selected dysfunctions. Attention and Encoding in Social Anxiety L i t t le investigation of information sampling and encoding operations has occurred in the area of social anxiety, yet the relevance of this perspective i s evident. In Zimbardo's important exploration of shyness he suggests that shy people miss much more of what i s going on because they are so se l f -preoccupied. The compelling reality of this distraction i s i l lustrated by the description of a shy male who responded to Zimbardo's Stanford Shyness Survey: "...I often find myself unable to follow what's going on in a conversation because I've been so nervously conscious of myself." (Zimbardo,1977, p.72) This example vividly demonstrates that shy people do in fact exhibit memory defic i ts , and that the deficits are due primarily to the inattention to available information. Glass and Merluzzi (1980) suggest that the anxiety/attention model put forward by Wine (1976) to describe behavioral deficits in test 29 anxious subjects, might be f ru i t fu l ly extrapolated to problems with social anxiety. This theory, based on Easterbrook1s cue uti l ization hypothesis, holds that heightened anxiety increases self-monitoring and triggers competing cognitions which impede the reception of information. Wine's hypothesis i s similar to the formulation of encoding dysfunction by Mueller, but places greater emphasis on the distracting self-rumination which test anxious subjects are believed to engage in . Support for the 'direction of attention' hypothesis comes from studies showing increased self-focussed attention in test anxious subjects; studies showing task irrelevant cognitions during test taking in test anxious subjects; and studies which demonstrate a reduction in test anxiety with attentional training interventions. (Wine, 1976) Applying Wine's attention theory to social anxiety, one would predict that the anxiety produced by social situations would narrow the range of information sampled and impoverish one's access to important social cues. The paralysis or inappropriate responding seen in socially awkward individuals would be due to a reduced access to information and the faulty assessments based on that incomplete information. Sampling would be restricted by the particularly well organized negative self schema which directs attention to schema confirming information and excludes important, but non-self referent cues. Evidence of Memory Deficit in Social Anxiety Limited and selective access to information either newly sampled or 30 from previous memory could account for many of the demonstrated cognitive distortions believed to be at work in the highly socially anxious individual. Evidence and theory from general anxiety frameworks, as discussed earl ier, as well as from situationally specific anxiety forms, suggest similar processes working within the shy individual. A search of the literature, however, reveals only a few studies which address this topic directly. O'Banion and Arkowitz (1977) had low and high socially anxious female subjects interact briefly with a male confederate. Following the interaction they gave and received feedback concerning the interaction. The feedback they received was supposedly their partner's evaluation of their performance. The results showed no evidence of a memory deficit at a l l , but did find a selective enhancement of memory for negative self-referent material in the group of high socially anxious females. These results do not support any of the theories which would predict a memory deficit occurring in shy individuals, but they are consistent with the body of research which shows socially anxious subjects to be particularly sensitive to negative self-evaluation. It i s possible to see the design of the O'Banion and Arkowitz study as masking any possible memory def ic i t . The experimenters arranged opposite-sex dyadic interactions followed by affectively-toned feedback regarding performance. This highly self-relevant information which could reinforce self-ruminations of inadequacy can be viewed as feeding the dominant self-monitoring pre-disposition of 31 socially anxious individuals. Receiving affect-laden sel f -evaluation is not typical of most social interaction, particularly f i rs t encounters. The question remains as to whether increased se l f -monitoring interferes with receiving neutral or other-relevant information. In addition, receiving affect-ladened self-evaluations from an opposite-sex partner is anxiety provoking for anyone regardless of their general comfortableness or discomfort in social settings. By heightening the anxiety level of the control group (Low Social Anxiety) O'Banion and Arkowitz may have inadvertently masked any differences occurring between the groups with regard to social anxiety alone. The average recognition for both high and low groups was 23 items out of 28 (82%) showing a high attentiveness and a restricted range for demonstrating differences. By definition, socially anxious subjects should experience more anxiety in social settings than socially non-anxious individuals. The O'Banion and Arkowitz study confounds evaluation anxiety with the social variable. It may be more meaningful to conclude that both high and low socially anxious subjects are vigilant to self-referent material, but high socially anxious individuals appear to be hypervigilant to negative self-referent feedback. Additionally, restricting memory measures to recognition and excluding recal l , tests only memory for information for which strong stimulus cues are available. We do not know whether the low and high anxiety groups in the O'Banion and Arkowitz study were equally able to generate stimulus items on a free recall basis. 32 In order to answer the question regarding informational access deficits in shy people i t i s necessary to measure recall abi l i ty to a variety of informational materials, and in a condition where social anxiety is operationally uncoupled from a strong evaluative component. In a more recent study Breck and Smith (1983) have attempted to separate the evaluative component of social distress from another important feature, the presence of a negative self-schema. Breck and Smith measured recall of self-descriptive traits in groups of socially anxious and non-anxious females. Some of the subjects were led to expect later interactions and/or evaluations involving a male stranger. Their results indicated that the high socially anxious females acknowledged as many positive traits as self-descriptive as did the low anxious subjects, but remembered significantly more of the negative traits than did the low anxious women. Negative t ra i t recall was found to be enhanced for the high anxious subjects when they were anticipating a future interaction with a male, but this was not influenced by the prospect of being evaluated in that situation. Breck and Smith interpret these results as supportive of the presence of a negative self-schema which i s situationally activated in high anxious subjects by the prospect of interactions and not expl icit ly from fear of negative evaluation. 0 The Breck and Smith (1983) study offers further indication that HSA subjects describe themselves consistently more negatively and, when 33 anticipating social contact, become more vigilant of information confirming their negative self-schema. This study used t rai t adjectives which were negative and positive in tone. A question relevant to the attentional strategies of anxious populations i s whether the inclusion of neutral items would affect recall in any way. Would HSA subjects highlight more or less affective material i f competing neutral material was available? The Breck and Smith study has separated evaluation features from the schema component and has demonstrated vigilance to negative se l f -descriptors. With regard to social memory deficits , however, the question of memory function regarding information about others remains unanswered. In another recent study which contributes some refinement to the original O'Banion and Arkowitz study Smith, Ingram, and Brehm (1983) attempted to assess the presence of cognitive excesses or deficits in the self-relevant information processing of socially anxious subjects. They hypothesized that a deficit would occur most l ikely on material which was encoded shallowly (structurally) under stressful conditions as opposed to material encoded with se l f -reference. Additionally, i t was expected that cognitive excesses (hypervigilance to self-referent material) would occur in the socially stressful conditions. Manipulation of the focus of the self-referent information would indicate whether hypervigilance was for (a) information about the self as perceived by others or (b) information about the self as viewed by oneself. Smith, Ingram, and Brehm adopted a depth of processing procedure in which taped 34 adjectives were to be processed according to various instructional sets which guided processing to surface features or self-comparison. The finding of significantly greater recall of public self-referent material supports the notion that HSA subjects are particularly sensitized to social-evaluative material rather than str ict ly se l f -preoccupation. In contrast to the Breck and Smith (1983) study the evaluative aspect of the experience in this study surfaced as c r i t i c a l . Consistent with the Breck and Smith study, however, no indications of memory deficits were apparent for the high socially anxious group. Two features of the Smith,Ingram, and Brehm research present some concern with regard to the presence of a memory def ic i t . The processing task occurred after the socially stressful event, not during the event. This methodology could allow for fluctuating attention between the task situation and self-ruminating internal dialogue without distracting from the task to a marked degree. Actual social interaction requires attention to social task demands with less opportunity for self-focussing episodes. This design may not be sensitive to memory interference which occurs in the ongoing processing of social interaction. In addition, the range of recall performance may be cause for concern. The total number of items possible for recall was 12 per condition. The range of mean recall was .42 to .64 out of 12 for the structural processing task, and 1.50 to 2.48 out of 12 for the public self-referent conditions. These scores indicate minimal retention 35 for the structural encoding (Was the word read by a male or female?), and quite limited retention for the most potent encoding sets (Does the word describe you?). Identifying memory differences between low and high socially anxious groups for messages low in personal relevance may require tasks which are designed to reflect actual social discourse more closely. The studies presented above which have investigated memory function in social anxiety have a l l used memory for self-referent information, and have consistently shown a heightened sensitivity to negative self-referent material which i s processed in a socially stressful context. No evidence is available which addresses the question of memory deficit for material about other people. The present study w i l l investigate the memory function of socially anxious individuals with regard to this unexplored area of other-referent social memory. The Role of Affect in Memory Performance Socially anxious individuals have been shown to recall more negative self-referenced material (O'Banion 8s Arkowitz, 1977). This finding supports the notion that shy individuals have relatively stable negative self-schema, but i t te l l s very l i t t l e about how information about other people is processed. Most social interactions contain very l i t t l e negative feedback and contain relatively more neutral statements or affectively toned information about others. An important question here i s whether socially anxious individuals process information about others through their own negative sel f -system by selectively attending to information that would have 36 negative implications for themselves. The attributional style of men who are low in social self-esteem (Girodo,Dotzenroth & Stein,1981) lends support to this notion of heightened sensitivity to negative information which is schema confirming. Another possibility is that socially anxious individuals are sensitized to negative information be i t relevant to themselves or to others. The affective tone of the message is likely to be an important determinant of recall as i t contributes to the meaningfulness of the message. Higgins (1980) concludes his review of the field of 'Social Cognition' with the suggestion that future research efforts incorporate notions of 'affect' and 'personal relevance' as integral parts of the evolving models of social cognition. Memory research has frequently used nonsense syllables or single adjective traits as stimuli; but as Hastie and Carlston (1980) point out, encoding procedures occurring with complex information are likely to be quite different. Procedures involving clustering, schema matching, egocentric review, and elaborate rehearsal are more likely to occur with association rich stimuli. The research question regarding affective content is included in the present study to tap potential information qualities which might affect memorability. The complexity of information available in real l i f e social situations cannot be underestimated. Manipulation of the message tone in the present study is expected to maximize the extent of cognitive dysfunction occurring with individuals for whom social involvements are troublesome. Socially relevant information i s also more characteristic of meaningful social exchanges and inclusion in this stimulus array should increase the ecological validity of the 37 study. Experimental Questions Emerging from the Review Regarding the Impact of Social Anxiety on Memory Several questions regarding the socially anxious person's access to both new and previously learned material have emerged from this review of memory function: a. Do socially anxious people in fact demonstrate a reduced access to previously experienced information? b. If a recall deficit exists, is i t a stable dispositional characteristic or is i t evoked only in specific anxiety arousing conditions? c If such a deficit exists, is i t a function of disruption at the i n i t i a l attentional/encoding phase or at the later retrieval phase? d. If a recall deficit exists, i s i t content specific? Is there a deficit for neutral information in favor of negatively-toned affective messages? The attentional and encoding formulations outlined earlier coupled with available data from memory research in social anxiety present a coherent framework for understanding possible memory difficulties occurring with socially anxious individuals. 38 Scope of the Present Study The present study w i l l address the four questions presented above regarding the reduced access to information experienced by s o c i a l l y anxious i n d i v i d u a l s . The design of the present study i s an attempt to balance concerns f o r empirical control with those of eco l o g i c a l v a l i d i t y , and i n t h i s process obtain r e s u l t s which are both t h e o r e t i c a l l y and c l i n i c a l l y meaningful. The present study w i l l address the following hypotheses: 1. I t i s hypothesized that memory i s disturbed i n s o c i a l l y anxious people only when they are i n sit u a t i o n s which evoke t h e i r anxiety. 2 . I t i s hypothesized that the impact of the anxiety-arousing circumstances w i l l i n t e r f e r e with information retention at the i n i t i a l a t t e n t i o n a l and encoding phases of t h e i r experience and not as a r e s u l t of anxiety aroused at r e t r i e v a l . 3. I t i s hypothesized that s o c i a l l y anxious subjects w i l l demonstrate a s e l e c t i v e attention to a f f e c t i v e material over neutral information, and p a r t i c u l a r l y messages construed as negative i n tone. Answers to these questions w i l l contribute s i g n i f i c a n t material t o the emerging conceptualization of dysfunctional cognitive a c t i v i t i e s which accompany intense s o c i a l anxiety. 39 CHAPTER III. METHODOLOGY I.DESIGN To address the three experimental hypotheses proposed in Chapter Two this study employed a design suggested by Eysenck (1977) and used by Nasby and Yando (1982) in which anxiety arousing conditions at stimulus exposure were completely crossed with anxiety arousing conditions at the stimulus retrieval phase. This design was able to isolate the effects of anxiety arousing manipulations at different points in the memory process (encoding, retrieval) on memory recall abi l i ty . Specifically, high and low socially anxious subjects were exposed to information which varied in content, while they were either in a social context or in a private context. The social context (receiving personal information about another person while interacting with that person) was assumed to raise the anxiety level of the high socially anxious subjects, but not of the low socially anxious subjects. Following a brief interim task which allowed material to enter long term memory while discouraging rehearsal, subjects were asked to recall the information they were exposed to. This retrieval once again occurred in either a social or private context. 40 The manipulation of the context in which information i s received and retrieved constituted the experimental treatments in this study. The other major independent variable in this design is the subject factor, high and low socially anxious males, as determined by a sel f -report social anxiety measure. Additionally, a stimulus item order factor was included as a control for possible differences in memorability due to stimulus item sequence. Significant primacy and recency effects are not anticipated in this study based on previous findings using recall of meaningful material in a stressful context (Brower & Mueller, 1978; Nasby & Yando, 1982). This i s consistent with the assertion by Hastie (1980) that schema based recall i s less affected by order in general. Figure 1 presents a schematic of this factorial design with four completely crossed factors. INSERT FIGURE 1 ABOUT HERE Dependent Measures In the retrieval phase of the experimental sequence, subjects were asked to complete two memory forms. Scores derived from these instruments constituted the dependent variables; six scores were computed from the subjects responses. Descriptions of these six measures follows: 41 F i g u r e ( 1 Schematic o f the 4-way F a c t o r i a l Design SOCIAL ANXIETY GROUP (A) ENCODING CONDITION (B) RETRIEVAL CONDITION (C) ITEM ORDER ( Q ) D l D2 A 1 (n=48) High S o c i a l A n x i e t y S o c i a l Encoding C l S o c i a l R e t r i e v a l n=6 n=6 C 2 P r i v a t e R e t r i e v a l n=6 n=6 B 2 P r i v a t e 'Encoding n=6 n=6 C 2 n=6 n=6 A 2 (n=48) Low S o c i a l A n x i e t y B l C l n=6 n=6 C 2 n=6 n=6 B 2 C l n=6 r.=6 C 2 n=6 n = 6 42 1. Total Recall The primary measure of memory was the number of correct stimulus items (see Social Recall Task Development in the Preliminary Results chapter) generated in the retrieval period during a free recall test. Each response generated was evaluated by judges to determine whether the production was an acceptable replication of the original item (see Appendix A). 2. Total Recognition A second measure of memory, less demanding in i t s retrieval requirements, required a subject to correctly identify the task items when they were presented in a checklist form in which an equal number of bogus items were embedded, (see Appendix B). 3. Number of False Positives A score derived from the recognition task indicating the number of items incorrectly identified as correct items was computed as a check on the accuracy of the total recognition scores. This measure offers some indication of signal 43 detection precision when memory is measured using a recognition format. 4. Number of Neutral, Positive, and Negative Items Recalled Subscores of the Total Recall score were computed as measures of the pattern of item retention with regard to affective-toned content. This set of scores was computed to address the hypotheses dealing with content specific recal l . 5. Percentage Affect Percentage Affect i s a derived score indicating what percentage of the total number of items correctly recalled i s represented by items judged to be affective (positive, negative) in tone. 6. Certainty A subject's estimate of his confidence that an item which he generated during the free recall task was, in fact, a member of the original stimulus l i s t . This score i s a control measure indicating performance differences 44 possibly related to willingness or reluctance to offer low probability responses. II. SUBJECTS Groups of high and low socially anxious males were recruited as subjects from the population of male students at Douglas College. Ninety-six (96) male students between the ages of seventeen and thirty were chosen to participate in the study from the 219 males who volunteered. Inclusion in the sample was determined by scores on the Social Avoidance and Distress Scale (Watson & Friend, 1969). Following the procedure by Clark and Arkowitz (1975) and O'Banion and Arkowitz (1977) subjects scoring in the 75th percentile and above constituted the High Social Anxiety Group and those in the 25th percentile or below were included in the Low Social Anxiety Group. This yielded six subjects per ce l l when Anxiety Group x Encoding Condition x Retrieval Condition x Serial Order were completely crossed (see Figure 1 for a schematic design). The decision to use male subjects was made solely on pragmatic grounds. Using both male and female subjects would add a gender variable which was beyond the scope of this study. Some findings suggest that a greater proportion of males describe themselves as socially anxious and report greater extremes in their experience of 45 social anxiety (Pilkonis,1978). The choice of males therefore was made to maximize the chances of recruiting the necessary socially anxious group within the constraints of the accessible male population. A. Recruitment: Volunteers were recruited using leaflets distributed during college registration and in Introductory Psychology classes during the f i rs t and second weeks of class. The leaflet (Appendix C) described the study and invited volunteers to complete the attached "Social Behaviour Questionnaire" (actually the Social Avoidance and Distress Scale) and return i t to the Impression Study office i f they were wil l ing to participate. Recruitment was planned for Douglas College registration periods starting with Fal l 1983 and continuing forward unt i l the cumulative number of volunteers allowed for the formation of the two anxiety groups. The representativeness of the obtained samples w i l l be addressed in Chapter Four. B. Subject Allocation Subjects determined to be within the criterion were coded and placed in a pool of either high or low socially anxious subjects. Phone appointments for the experimental sessions were made by a secretary who randomly chose subjects from each group, with the proviso that an equal number of subjects from each group be scheduled for each testing block. As this proviso was not always 46 possible, and as some cancellations and no shows occurred, this procedure needed to be modified somewhat near the end of the experimental phase to insure that appropriate numbers from each group appeared in each of the experimental conditions. Additionally, an informal attempt to assign older subjects somewhat evenly across a l l treatment conditions was made to avoid potential sub-sample age disparity. This procedure resulted in a modified random sampling in which approximately 90% of the subjects were assigned on a totally random basis. In a l l cases experimenter and confederates remained unaware of the anxiety group classification of the subjects. III. CONFEDERATES USED IN THE SOCIAL MANIPULATIONS The social mode of the Encoding Condition and the Retrieval Condition required an interaction sequence in which a female assistant was introduced to the male subject as a fellow student. An opposite-sex confederate was used in this social manipulation to enhance the social anxiety provoking potential of these interactions; this i s in keeping with the findings that heterosocial anxiety i s a substantial component of general social anxiety (Leary, 1983; Martinson & Zerface,1970; Richardson & Tasto, 1976). For use in this study four women between the ages of 21 and 26 were selected from nine women who volunteered to assist in this social interaction sequence. A l l of the assistants were second year psychology majors and judged to be relaxed, personable, and spontaneous during a selection interview. Personalities are impossible to regulate in experimental situations so explicit 47 instructions and training for the interaction sequences were conducted to insure some degree of uniformity of the subjects experiences in the social mode. A l i s t of the guidelines used in confederate training i s presented in Appendix D. These guidelines were discussed thoroughly and each confederate practiced her approach in role-play sessions with the experimenter. Prior to contact with the confederate's f i rs t experimental subject, each confederate had a session with a pre-experimental subject drawn from the unused volunteer group who scored in the mid-quartiles on the subject selection instrument. The practice session was thoroughly debriefed regarding the confederate's impressions and concerns. During the experimental t r i a l s , informal aural monitoring of the interactions from an adjacent room revealed no differences in the length of conversations (5 minutes plus or minus 2) for the sampled interactions, but did reflect wide variation in the direction of conversation and the level of observer perceived comfortableness or uncomfortableness of the participants. These latter substantive concerns are d i f f icu l t to orchestrate, and at the same time, maintain a natural and genuine flow to the interaction. It i s assumed that random assignment of confederates to conditions and subject group would control for any systematic bias operating. More elaborate monitoring of this aspect of the study was considered and rejected for two reasons: (a) the cumbersomeness and obtrusiveness of videotaping interactions, and (b) the belief that uniformity of interaction was not essential or desirable. Even i f marked differences between the interaction patterns of the confederate with high and low anxious subjects occurred, this difference could not 48 necessarily be attributed to personal differences or bias of the confederate. By definition, socially anxious subjects experience and respond differently in social situations than do low anxious individuals. The moment by moment communication of warmth, openness, tension, aloofness, faci l i tat ing reactions, or alienating reactions cannot and should not be standardized. The social and private variations in this study were used to establish different contexts within which high and low anxious subjects would process information. If the impact on information retrieval i s a result of a subject's unique response to this context variation, i t i s seen as a reflection of real l i f e reactions and supportive of the experimental hypotheses. It i s important, however, to insure that a particular confederate did not exert a systematic bias within a particular treatment mode or with a particular anxiety classif ication. The "Confederate" section of the Preliminary Results chapter addresses this concern stat ist ical ly . IV. PROCEDURE The experimental design involved four treatment combinations: A. Social Encoding/Social Retrieval B. Social Encoding/Private Retrieval C. Private Encoding/Social Retrieval D. Private Encoding/Private Retrieval 49 Subjects within both treatment groups (high and low social anxiety) were assigned randomly to one of the four conditions. Subject assignment occurred as described in the section above describing subject allocation. Subjects arrived at the Psychology Lab and were given a brief introduction which was identical for a l l conditions. The rationale for the study emphasized the importance of impression formation in the processes of social interaction and stated specifically that this study was designed "to explore how men form impressions of the women they meet". Hamilton, Katz and Leirer (1980) found that instructional sets involving impression formation instructions as compared to recall instructions actually produced better retention of personal information, and without the potentially interfering tendency to rehearse the newly experienced material. Hamilton et. a l . suggest that this impression formation response set may evoke 'deeper' levels of i n i t i a l processing which result in more completely formed person schema. If this i s in fact the case, the impression formation mode should e l i c i t a more genuine and ecologically valid experience than a recall instructional set. Following the introduction subjects were asked to complete some forms: a background information form (see Appendix E), a consent form (see Appendix F), and a modified version of the State Trait Anxiety Inventory (see Instrument section below). The primary purpose of the background information form was to introduce subjects gently to the experimental process by having them answer questions which were 50 non-threatening and of a broad enough range so as not to pre-dispose the subject to specific notions about the nature of the study. Embedded in the background questionnaire were questions regarding the subject's age and high school academic performance for later use in determining subject group equivalence. The State-Trait Anxiety Inventory is described in the Instrumentation section. The experimental manipulations involved placing subjects in either a social or private context at two different times: while they were exposed to the task stimulus material and while they retrieved that material. For the Social Encoding condition, subjects were told that in a short time they would be introduced to another student and that the two of them were to chat for a few minutes "as i f they were getting to know each other for the f i rs t time", which for a l l subjects was, in fact, the case. Subjects were further told that after chatting a few minutes the other student would play a tape on which she had read a l i s t of things about herself which she was wil l ing to share. The subject was told to l isten carefully to the tape. When the tape had finished, the subject was instructed to open the door as a signal to the experimenter. At that time the experimenter returned, asked the confederate to leave the room, and presented the subject with the interim task. In the Private Encoding condition, the subject was shown directly to 51 the experimental room and given a sheet of instructions (Appendix G) to read which explained that they were to listen to the tape in the tape recorder in front of them and to turn off the machine and signal the experimenter when the tape had ended. When the subject signalled the experimenter returned and presented the interim task. For the Social Retrieval condition, the confederate entered the room with the experimenter and is reintroduced to the subject (Social/Social condition) or is introduced to the subject for the firs t time (Private/Social condition). In the Social Retrieval condition in which the previous exposure had been in a private mode, the experimenter explained that the student being introduced is the one who had recorded the tape listened to before. The experimenter then explains that he would like the subject to try and recall as many of the items from the tape as he can, and that the student (confederate) would record these on the forms provided. A four minute time period was provided for this recall phase and following this the confederate presented the item rating and recognition components. When the forms were collected they were quickly scanned for possible omissions by the experimenter who then explained that the experiment was over. After the confederate had left, each subject was asked to indicate again on the anxiety scale how they felt in the previous situation. The subject was then debriefed with regard to the experiment. For the Private Retrieval condition, subjects were given an instruction sheet (Appendix H) following the interim task which instructed them to write down on the form provided as many of the 52 descriptions from the tape as they could remember. Four minutes were allowed for this recall at which time the experimenter returned and presented the recognition form. Following the recognition form, the subject completed the anxiety rating and was told that the experiment had ended. This was followed by debriefing as in the other condition. Debriefing After the memory forms and anxiety rating form had been collected, subjects were asked for any comments or impressions they might have had about their experience. Subjects were asked what they believed the study was trying to determine and whether they had any prior knowledge of the purpose of the experiment. In the conditions involving confederates, subjects were introduced to the female student and i t was explained that she was an assistant in the experiment and that the items on the tape were not really about her. Subjects were asked whether they believed that the items on the tape were actually about the student or whether they were suspicious about her or her recorded disclosures. The actual purpose and design of the experiment was explained and subjects were encouraged to ask about any part of the process. Subjects were also told that when the results were available they were welcome to return and discuss the study again. Every effort was made to dispell any anxiety or misgivings which subjects might have had with regard to the experience. By far the dominant response from subjects in both anxiety classifications and in every treatment condition was one of interest and enjoyment. While a few subjects indicated that some of the self-disclosures seemed to be unusual things for a person to 53 snare, none of them indicated that this detracted from their involvement in the study or their abi l i ty to recall the items in any significant way. Finally, subjects were asked not to discuss their experiences with other students in order to maintain confidentiality regarding the procedures. V. INSTRUMENTATION A. The Social Avoidance and Distress Scale The Social Avoidance and Distress Scale (SAD) scale i s a wel l -documented and widely used instrument for differentiating quickly high and low socially anxious populations. The SAD (see Appendix I) was designed to measure one's endorsement of relatively specific behavioral and experiential components of social anxiety. In considering previous measures of anxiety four dimensions of social anxiety were noted by the authors of this scale, Watson and Friend (1969): subjective distress, execution of avoidance responses, impaired performance, and physiological signs. The authors intentionally excluded items from the last two categories so that the scale would accurately reflect dimensions of class "anxious" membership and not the resultant behaviors which correlate with class membership. The resultant scale i s a 28-item, true-false questionnaire which reflects specific t rait anxiety. The scale was i n i t i a l l y evaluated using 205 students from the University of Toronto. The mean biserial correlation of selected SAD items i s .77. Another measure of internal consistency, the KR-20, yielded a co-54 efficient of .94 . It appears that the test i s highly homogeneous. Correlation with the Crowne-Marlowe Scale of Social Desirability was - .25 (N=205, p.01) indicating a minimal relationship with social desirability which, according to Watson and Friend (1969) would have been even less had the Crowne-Marlowe scale i tse l f been more homogeneous. Supportive convergent and discriminant relationships have been reported by Watson and Friend using correlations with Manifest Anxiety (.54), the social evaluation parts of the SRIA (.45), and Paivio's Audience Sensitivity Inventory (.76). Other validational support has come from research where the SAD has discriminated socially anxious from non-anxious populations (Arkowitz, Lichtenstein, McGovern & Hines, 1975; Caldwell, Calhoun, Humphreys & Cheney, 1978; Clark & Arkowitz, 1975) and with research demonstrating changes in anxiety following treatment (Goldfried, Linehan, & Smith, 1978; Kanter & Goldfried, 1979). The SAD appears to be a valid and reliable instrument for establishing groups which clearly differ on the dimension of social anxiety. B. Situational Anxiety Rating As a way of monitoring subjects' anxiety reactions to the experimental conditions, a quick self-rating anxiety scale was administered prior to the experiment and at the end of the experimental sequences. The instrument used was a modified version of the State-Trait Anxiety Inventory: Form Y (Spielberger, Lushene, Vagg & Jacobs, 1983). 55 The State-Trait Anxiety Inventory (STAI) i s widely viewed as the most carefully constructed of the general anxiety measures (Levitt,1980; Wine, 1973). In addition Buros (1978) indicates that this instrument was the most widely used scale for research with anxiety over the preceding decade. Spielberger (1983) has published a bibliography of research use with the STAI cit ing over 2000 references. The popularity of the STAI in psychological research stems from the scale's demonstrated psychometric properties. Test-retest re l iabi l i ty coefficients are high for the Trait Anxiety component of the scale but not for the State Anxiety component. A better indication of re l iab i l i ty for situational instuments i s internal consistency. Computed alpha coefficients using the STAI-S scale with high-school and college students range from .86 to .94. Spielberger (1983) indicates that alpha re l iab i l i t y i s consistently higher when administered under conditions of psychological stress. Construct validation for the STAI-S scale has been examined by exposing subjects to high and low stress conditions between administrations of the scale. Significant mean differences and correlations between .60 and .73 were obtained between anxiety scores and experimental conditions (Lazarus & Opton, 1966; Spielberger, 1970). This instrument emerges as a highly reliable and valid measure. In addition, two additional characteristics of the scale made i t a desirable choice for this study. F i rst , the state anxiety component 56 of the scale can be modified to evaluate state anxiety in the specific situation of interest to the experimenter. In the present study, instructions were worded to direct the subject to indicate his feelings (a) at the present time, and (b) how he fe l t in the previous task situation. Secondly, Spielberger (1983) suggests that when time is a crucial factor and i t i s useful to obtain repeated measures of state anxiety, a ten-item short version of the STAI-S scale w i l l present a reasonably valid estimate of anxiety level and one which interferes less with other experimental act iv i t ies. The modified version of the STAI-S scale (see Appendix J) adopted for this study consisted of ten items reflecting feeling states to be endorsed by the subject on a four point continuum ranging from 'very much so' to 'not at a l l ' . Half of the items were anxious states and half were statements reflecting pleasant or comfortable feelings. Scores on the pleasant items were reversed and totalled with the scores on the anxious items to obtain a total score. C. Interim Task Ronald Johnson (1980) in his review of the concept of rehearsal in memory research indicates that the repetition of stimulus information or the formation of new encoding strategies occurring covertly following experimental stimulus presentation are important factors influencing the extent of stimulus recal l . Several strategies for preventing rehearsal have been used with the goal of limiting any potentially faci l i tat ing covert processes, and at the same time not introducing memory interference. 57 In the present study two procedures were used to mnimize the possibil ity of i l l i c i t rehearsal. The task instructions directed the subject to monitor the sentences and 'form an impression' of the stimulus person, and secondly, the interval between exposure and recall involved a distracting task. The absence of the expectation of having to recall the stimulus acts as a control for rehearsal. This procedure lessens the tendency to rehearse (Marshall & Werder, 1972) and appears to have no detrimental effects on memory. In fact, Hamilton, Katz and Leirer (1980) found a greater degree of sentence recall with 'impression formation' instructions than with memory instructions using an interim arithmetic problem-solving task. As rehearsal was limited by the interim task, this difference i s attributed to the more meaningful encoding strategies employed with impression formation instructions and not to rehearsal opportunity. The choice of interim task was made with regard to the following requirements: 1) The task be sufficiently long to allow the experimental stimulus material to enter long term memory. 2) The task be sufficiently engaging to occupy the subject's interest and attention, thereby minimizing rehearsal of previous material. 3) The task should appear to be somewhow related to the processes and rationale of 58 the study i tse l f so as not to appear as an irrelevant excercise. 4) The task not be too d i f f icu l t or taxing. 5) The task does not require activit ies markedly similar to the experimental task. The task chosen for use as the interim exercise which seemed to meet a l l of these cr i ter ia to a reasonable degree was an image formation task; specifically, the Visual Imaging section of the Revised Betts Questionnaire of Mental Imagery (Sheehan,1967) (see Appendix K). This task asks subjects to form images of familiar objects and people and to rate the degree of vividness which they were able to achieve. The task usually takes from 7 to 10 minutes to complete and i s one which most subjects report to be an interesting and pleasant experience. The image formation task was deemed to be sufficiently different from the processes of encoding and retrieving auditorily presented statements of self-disclosure, but at the same time appearing "intuit ively' to be related to the processes of 'impression formation1. As the actual ratings produced by the subjects were not relevant to this study, the psychometric properties of the Bett's QMI were not addressed. D. Social Memory Measures Tracking social memory i s a d i f f icu l t experimental task as i t requires standardized memory stimuli occurring in an ' in vivo' context. This study has employed an analogue format with semi-structured stimulus material. In an attempt to preserve some of the 59 complexity of social interactions, this study developed a set of statements each of which describes a single thought or event. This procedure fa l l s short of the complexity and spontaneity of naturally occurring conversational stimuli, but the sentences do offer a considerably more authentic array of social information than a memory task involving l i s t s of words. In addition, the set of pre-recorded sentences were presented within the context of an actual interact ion—a procedure aimed at increasing the contextual reality. The stimulus l i s t consists of 24 statements of self-disclosure varying in affective tone which were read onto audio tapes by the confederates. The procedures used in developing the stimulus array are presented in the section Social Recall Stimulus Development of the Preliminary Results chapter. The recall of the stimulus sentences was the major dependent variable. Subjects completed the Recall Form during the retrieval phase. The Recall Form (see Appendix L) instructed the subject to write down as many of the statements from the tape as he could remember. The instructions asked subjects to reproduce statements as accurately as they could but not to withhold statements because of doubt over exact wording. Four minutes were allowed for this recall phase. After the items had been generated , further instructions directed the subject to rate each of the sentences according to whether he would view such a self-disclosure from another person as being positive, negative or neutral. Finally, subjects were instructed to indicate how certain they were that the items they generated were in fact on the original tape. For this rating 60 subjects used a four point scale in which l=very certain, 2=somewhat certain, 3=somewhat uncertain, 4=very uncertain. This Recall Form generated several scores for analysis; total number of items recalled; number of positive, negative and neutral items recalled; percentage of total recall represented by affective items; and a mean certainty score reflecting the subject's level of confidence regarding the items he generated. Following the Recall Form subjects completed a Recognition Form (see Appendix B). The Recognition Form consisted of the original 24 sentences which were embedded randomly in a 48 sentence array which included 24 bogus sentences. Subjects were to indicate, by checking a "yes" or "no" box, whether a particular item was one of the sentences they had heard on the tape. The Recognition Form was scored for the number of items correctly identified from the tape and the number of false positive identifications. A derived recognition score was obtained by subtracting the false positive identifications 1 from the total of correctly identified items. 61 CHAPTER IV. PRELIMINARY RESULTS I. DEVELOPMENT OF THE SOCIAL RECALL STIMULUS ITEMS The goal of this phase of the study was the development of the stimulus l i s t to be used as the social information disclosed by the female confederate during the subject/confederate interaction. The stimulus l i s t was designed to be a series of behavioral or attitudinal statements which varied on a dimension of affective tone. Items were generated and included in the l i s t according to the following guidelines regarding structure, language, number, imageability, distinctiveness, and affective tone: 1. Structure The statements are to be complete sentences approximately equal in length and containing a subject, object, and predicate. Each statement w i l l express a single clear idea worded in a self-disclosing declarative form. This procedure i s similar to that used by other researchers in sentence memory (Hamilton, Katz & Leirer, 1980). 62 2. Language A l l of the statements generated w i l l use familiar language and concepts and avoid unusual words or phrasing. This guideline i s to insure quick and easy comprehension when the sentences are received aurally. The debriefing which followed the i n i t i a l ratings of the sentences for affective tone, as well as with the preliminary data collection, revealed no evidence of confusion or lack of clarity of meaning with regard to the wording or meaning of any of the statements. 3. Number The f inal array of statements w i l l be large enough to insure that a complex array of information i s presented: to insure that the number of items i s sufficiently large to achieve subset re l iab i l i t y for affective tone: and yet remain a size which i s manageable to receive. A l i s t of 24 statements, of which eight are positive in tone, eight are negative in tone, and eight are neutral in tone, was deemed appropriate for meeting the cr i ter ia of complexity, re l iab i l i ty , and manageability described above. Other researchers have used similar size arrays (Hamilton, Katz & Leirer, 1980; Hatvany, 1976; Srul l & Wyer, 1979) in sentence memory research. Preliminary work with 13 subjects yielded a mean recall of 8 items and a range of 4 to 11. This suggests that the 24 item array presents an adequate range for registering individual 63 differences without impeding the overall recal l . Additionally, the 24 item l i s t allows subsets with 8 items, assuring acceptable re l iabi l i ty for registering differences in recall due to the affective content. 4. Imageability The role of stimulus imageability in the recall of connected narrative has been postulated by several authors (Paivio, 1969; Richardson, 1981) to be highly significant. Paivio holds that material of high concreteness is capable of being encoded both semantically and in the form of non-verbal images. This capacity for dual coding i s purported to enhance the probability of recal l . While this hypothesis has been demonstrated for recall of individual words, the evidence for superiority of high imageable sentences in recall i s less clear (Richardson, 1981). Many studies employ stimuli which differ on the dimension of 'concreteness' and argue that concreteness i s an alternate measure of the 'image-arousing potential of verbal material1 (Richardson, 1981). In support of this position Paivio, Yui l le , and Madigan (1968) found a correlation of .83 between the concreteness and imageability ratings for their 925 English nouns. Richardson argues that these two dimensions are conceptually and theoretically distinct, but concedes that the image-arousing properties of a stimulus i s quite l ikely to be the primary determiner of recall 64 performance. Identification of stimulus imageability as a potentially powerful one in determining recall performance, marks this dimension as one needing control in human memory research. In the present investigation, stimulus items are sel f -statements describing thoughts, experiences, or attitudes, and therefore vary widely in their abi l i ty to evoke specific visual images. Additionally, this form of stimulus i s l ikely to be susceptible to higher degrees of individual differences in meaningfulness - a property which is a natural component of social discourse, and one which i s d i f f icu l t to control. This variabil ity was noted in the imageability raters' responses where several statements received ratings at both ends of the imageability continuum. Determination of the equivalence of imageability w i l l be reported in the Imageability Rating section below. 5. Distinctiveness Eysenck (1979) defines 'distinctiveness' with regard to information arrays, as a condition in which a statement shares relatively few features with others in episodic memory. Fisher (1981) and others have shown that item distinctiveness increases the probability of that item being recalled. The intent of this guideline was to generate items which did not differ from each other markedly with regard to this distinctiveness dimension. 65 As the l i s t of statements i s presented in a\ context of self-disclosed information pertaining to the confederate, i t can be argued that this context develops an expectancy set which increases perceived item homogeneity.- In addition, items were written, and ultimately included in the stimulus l i s t , which described relatively routine and familiar events. Selection of the f inal l i s t was made keeping in mind the coherence and compatibility of items with each other. In the f inal stimulus array, no item contradicted or departed radically from the other items presented. Preliminary analysis of the individual items recalled shows 21 of the 24 items reported, indicating no i n i t i a l evidence for favored items. However, analysis of recall patterns cannot be used as a direct check on equity of distinctiveness as one of the experimental hypotheses postulates specifically a special case of perceived distinctiveness or selective recal l . 6. Affective Tone For the purposes of this study, the f inal stimulus array would include an equal number of statements which have been shown to represent differing levels of affective tone. Independent raters w i l l determine whether each statement in the i n i t i a l item pool would be perceived as negative, positive, or neutral i f i t was disclosed to them by a young woman. The data for 66 these ratings w i l l be reported in the Affective Tone Rating section below. For the above guidelines, guidelines A, B, and C are considered in the generation of items: guidelines D and F were determined empirically by independent raters and guideline E was employed both at the item generation phase and in the f inal selection. A. ITEM GENERATION A pool of 84 statements was generated i n i t i a l l y from which the f inal 24 items were chosen. Behavioral and attitudinal statements which reflect positive, negative, and neutral attributes were generated using themes based on items from the Pleasant Events Schedule (MacPhillamy & Lewinsohn,1976). The Pleasant Events Schedule (PES) is a 104 item self-report inventory which derives scores of social and physical anhedonia from the subject's endorsement of statements reflecting pleasant and unpleasant experiences. Lewinsohn and Amenson (1975) factor analyzed the PES inventory and obtained five factors for positive mood related items and five factors for negative mood related items which together accounted for 74% of the total variance. Items in the i n i t i a l pool of statements for the present study were based directly on experiential themes from the PES scale. In addition, item generation adhered to the formula devised by Fishbein (1965) for the generation of affective attitudinal statements. In the Fishbein system, a positive expression about a positive event or a negative expression about a negative event results in a positive item (example: I enjoy quiet dinners with 67 friends). Items which express mixed valence, positive feelings about negative events or negative feelings about positive events produce negatively toned items (example: I usually try to avoid romantic involvements). B. RATING OF ITEM AFFECTIVITY The l i s t of 84 items varying in affective tone was presented to a group of judges for independent determination of affectivity. Ten individuals performed this rating, including five psychologists (individuals purported to have training and experience in discerning affective content) and five non-psychologists including a secretary, a student, two homemakers, and a lawyer. The rating procedure consisted of a card sorting process in which each of the 84 sentences was to be placed i n i t i a l l y into a positive, negative, or neutral p i le . Raters were to sort according to the following instructional set: Imagine that you are hearing these statements as self-disclosures from a young woman. You are to determine whether each statement i s a positive, negative, or neutral message for this woman to disclose about herself. After the i n i t i a l sort, raters were then instructed to sort the positive and negative piles into further degrees of intensity. The two phases of sorting resulted in each sentence being placed on a 1 to 7 continuum of affectivity ranging from extremely positive to extremely negative. 68 Phrasing of the seven sort positions was adopted from the intervals suggested by Osgood (1959) using the modifiers "extremely', "quite', and "sl ightly' for the non-neutral variations. The scores from the ten raters for the 84 items were analyzed using the LERTAP item analysis for affective scales program (Nelson, 1974). This program generated means, standard deviations and consensus percentages for each statement. C. ITEM SELECTION Clear subgroupings with high inter-rater agreement were obtained representing positive, negative, and neutral points on the scale. With regard to the sub-group differences between raters, the mean rating for the psychologist raters was X=4.04 (s.d.=1.61) and for the non-psychologist raters X=4.08 (s.d.=1.59). A t-value computed using the Direct-Difference Method yielded a t = .953 (with df = 83, t[.05] = 1.99) indicating no difference in perceived affective content between psychologists and non-psychologist raters. D. CRITERION FOR FINAL ITEM SELECTION For neutrality of affect, only a rating of four represented absence or equivalence of affective tone. For inclusion in this group a str ict criterion of (a) 9/10 judges rating the item as 4, and (b) item mean ratings to f a l l between 3.9 to 4.1 was used, and this yielded 23 acceptable statements for the neutral sample. Selection of the f inal eight items from this sample was made with regard to 69 compatibility with other items in accord with the guideline concerning distinctiveness. For inclusion in the sample of negative items, a statement (a) must be selected by a l l 10 raters as either sl ightly, quite, or extremely negative, and (b) rated as quite or extremely negative by 70% of the judges. This criterion yielded 15 items with a mean item range of 1 to 2.2. The selection criterion for positive statements was identical to that of the negative sample with (a) inclusion in one of the three positive categories by a l l judges, and (b) 70% of the judges viewing the item as either quite or extremely positive. This criterion yielded 20 items with item means ranging from 5.8 to 7. It i s clear that the raters distributions for each of the categories is non-overlapping, therefore insuring with some confidence that the items used do, in fact, represent the assigned affective domain. Table 1 shows the descriptive properties for the statements chosen in the f inal stimulus instrument. Appendix A presents the stimulus items. 70 TABLE 1 Means, Standard Deviations, and Rater Consensus Percentages for Items Selected as Negative, Positive, and Neutral for the Social Recall Stimulus Array ITEM RATER CONSENSUS VALENCE MEAN * S.D. PERCENTAGE ** NEGATIVE 2.0 .82 70 % 1.9 .74 80 1.8 .63 90 1.3 .48 100 2.2 .63 70 1.8 .63 90 2.0 .67 80 2.0 .82 70 NEUTRAL 4.0 .00 100 % 4.1 .32 90 4.0 .47 80 4.1 .32 90 4.1 .32 90 4.0 .00 100 4.1 .32 90 3.9 .32 90 POSITIVE 6.4 .52 100 % 6.3 .82 80 6.8 .42 100 6.6 .52 100 7.0 .01 100 6.0 . 47 90 6.1 .87 70 6.1 .57 90 * A rating of l=extremely negative, 7=extremely positive. ** Raters N=10. 71 E. RATING OF ITEM IMAGEABILITY The final 24 statement array was again submitted to independent raters to determine the imageability properties of the three affective subsets (positive, negative, and neutral). Raters were given the deck of 24 cards containing the stimulus statements and instructed to assign a score between 1 and 7 to each sentence by placing i t in the appropriate pile. Raters were to determine placement according to the ease with which a mental image of the statement could be formed. The intervals used for rating were adopted from the interval descriptions used in the Revised Betts QMI (Sheehan, 1967) where a value of '1' is assigned to an image which is "perfectly clear and as vivid as the actual experience" and a rating of '7' is assigned to "no image present at a l l , you only 'knowing' that you are thinking of the sentence". Ten individuals completed this rating: five students, two psychologists, one secretary, and two homemakers. Table 2 shows the means and standard deviations, and the t- values for the affective subgroup comparisons. 72 Table 2 Means, Standard Deviations, and Comparison Tests for the Imagery Ratings of the Positive, Negative, and Neutral Items Used in the Social Recall Stimulus Array ITEM . MEAN IMAGERY STANDARD VALENCE RATING * DEVIATION NEUTRAL (N=8) 2.84 .35 POSITIVE (N=8) 2.44 .61 NEGATIVE (N=8) 3.09 .65 MEAN COMPARISONS t VALUE SIGNIFICANCE ** Neutral/Positive 1.60 n.s. Neutral/Negative .96 n.s. Positive/Negative 2.09 n.s. * Raters N=10 ** t(.05,14) < 2.15 Although the comparison of the means for the positive and negative items approached significance, none of the differences between means reached stat ist ical significance at the .05 level. This finding establishes that the samples of positive, negative, and neutral items are largely equivalent with regard to the ease with which they can be visually imaged. The imaging ratings are characterized by a high 73 degree of variability in the perceived ease of image formation; this reinforces the point made earlier that significant individual differences exist with regard to the imaging process. F. ORDERING OF STIMULUS STATEMENTS The sequencing of the 24 statements was subjected to the following restrictions: 1) the sequence is to begin with a neutral item, 2) equal numbers of positive, negative, and neutral items will occur in the f i r s t and last halves of the l i s t . 3) no more than two items with the same affective designation will occur consecutively. Item selection was otherwise randomized following these restrictions. This sequencing procedure created the necessary l i s t balance for the testing of the hypothesis regarding potential Order Effect. Sequence 1 to 24 is order A and sequence 13-24, 1-12 is Order B. G. AUDIO TAPE RECORDING OF THE STIMULUS LIST Four female assistants were used in the experiment acting alternately as the confederate. Each assistant recorded her version of the 24 statements twice; once for order A and once for order B. Care was taken to insure equal pacing, clear articulation and conversational tone and expression. A four second interval between statements was 74 used to standardize the interval space. Several versions of the recording were made u n t i l the tape reflecting the above qualities were achieved. The running times for the 24 sentences, when recorded, ranged from 3 minutes 18 seconds to 3 minutes 37 seconds across assistants, indicating close approximation of pacing and stimulus duration time. II. DESCRIPTION OF SUBJECT POPULATION AND DETERMINATION OF HIGH AND LOW SOCIAL ANXIETY GROUPS Low and high social anxiety groups were determined by identifying subjects scoring i n the f i r s t and fourth quartiles on the Social Avoidance and Distress Scale as described i n the 'Subjects' section of the Methodology chapter. Two phases of subject recruitment were necessary to achieve adequate numbers of subjects i n both the high and low social anxiety categories. The quartile cut-off points on the SAD used for the determination of the low and high anxious subjects were based on the i n i t i a l intake (Fall 1983) of volunteers; this c r i t e r i o n for subject inclusion was continued for the second recruitment phase (Spring 1984) to insure consistency. The F a l l 1983 recruitment yielded 127 volunteers. Table 3 presents the means, standard deviations, and quartile ranges which the two intakes yielded separately and combined. 75 Table 3 Means, Standard Deviations, and Quartile Ranges of the Social Avoidance and Distress Scale Scores for the Volunteer Subject Population, Reported for the Fal l 1983 Intake, Spring 1984 Intake, and the Intakes Combined INTAKE N MEAN S.D. 1ST QT. 4TH QT. FALL 1983 127 7.99 5.66 0-4 12-28 SPRING 1984 92 8.04 6.52 0-3 13-28 COMBINED • 219 8.01 6.02 0-3 12-28 Table 3 indicates that the distribution of SAD scale scores in the second recruitment phase was slightly more skewed to the low end and would have placed the quartile determination points at marginally more restrictive intervals than the intervals determined by the i n i t i a l distribution. The resulting low anxious subject group had a mean score on the SAD of 2.21 (s.d.=1.15) and the high anxious group had a mean of 15.75 (s.d.=4.16). The distribution achieved in the sample was somewhat lower than the one found by Watson and Friend (1969) in their study where the mean was 11.20 for their male sample. The mean in the present sample of 8.01 (s.d.=6.02) indicates some likelihood that a less socially avoidant population self-selected for participation in 76 this study. This is understandable in the light of the demands of the study. The leaflet used for recruitment in the present study suggested the possibil ity of opposite-sex interaction, a considerably more demanding request than to only f i l l out a questionnaire without further follow-up as in the Watson and Friend study. The obtained cut-off criterion, however, i s quite similar to that used in other studies. Watson and Friend used a cut-off of 12 and above to form their highly anxious group to validate their scale. O'Banion and Arkowitz (1977) formed high and low anxious groups using the 12+ and 0 - 3 ranges on the SAD with a mean for the high anxious group of 16.03 quite similar to the obtained high anxious mean in the present study. The highly skewed distribution densely represented at the low end i s in keeping with the distributions obtained in the scale's standardization. Watson and Friend (1969) suggest that measures of social pathology (avoidance, distress) are probably not normally distributed in the general population and this skewedness may be valuable in increasing the detecting power of the scale for early signs of social withdrawal or schizoid reaction. Watson and Friend (1969) also point out that this scale was not designed to measure the opposite instance of social avoidance. The opposite of social avoidance may not imply a f f i l ia t ion but simply 'lack of avoidance'. 77 III. DETERMINATION OF SUBJECT EQUIVALENCE ACROSS EXPERIMENTAL CONDITIONS WITH REGARD TO SOCIAL AVOIDANCE AND DISTRESS SCALE SCORES, AGE, HIGH SCHOOL GRADE POINT AVERAGE, AND CONFEDERATE EFFECT To insure that the sub-samples of high and low socially anxious subjects did not differ from each other on factors that might have influenced the results, sub-samples were examined for equivalence on three subject variables: scores on the group criterion measure (SAD scores), subject's age, and subject's high school grade point average, used here as a rough measure of intellectual abi l i ty . In addition, the uniformity of experience offered by the confederates used in the study was examined for possible confounding effects. The results are reported below. A. Social Avoidance and Distress Scores Social Avoidance and Distress (SAD) scores were analyzed using two separate two-way ANOVA's (encoding condition x retrieval condition); one for the high SAD group and one for the low SAD group. Scores were compared across the four treatment combinations (social/social, social/private, private/social, private/private) in each analysis. Table 4 reports the means and standard deviations for each group in each of the four treatment combinations. Tables 5 and 6 report the ANOVA results for the high anxious and the low anxious groups respectively. 78 Table 4 Means and Standard Deviations of Social Avoidance and Distress Scale Scores for Both High and Low Anxiety Groups in A l l of the Encoding x Retrieval Treatment Combinations HIGH SOC. ANX. LOW SOC. ANX. TREATMENT M S.D. M S.D. SE/SR 15.92 5.02 2.17 1.40 SE/PR 15.33 3.49 2.25 1.14 PE/SR 16.00 4.39 2.42 1.08 PE/PR 15.75 4.14 2.00 1.04 SE=social encode, SR=social retrieve, PE=private encode, PR=private retrieve. 79 Table 5 ANOVA Results for the Social Avoidance and Distress Scale Scores of High Anxious Subjects (Encoding x Retrieval) SOURCE OF VARIATION SS DF MS F PROB. F ENCODING (B) 0.75 1 0.75 0.04 0.84 RETRIEVAL (C) 2.08 1 2.08 0.11 0.74 B x C 0.33 1 0.33 0.02 0.89 RESIDUAL 811.83 44 18.45 Table 6 ANOVA Results of Social Avoidance and Distress Scale Scores of Low Anxious Subjects (Encoding by Retrieval) SOURCE OF VARIATION SS DF MS F PROB. F ENCODING (B) 0.0 1 0.0 0.0 1.00 RETRIEVAL (C) 0.33 1 0.33 0.24 0.63 B x C 0.75 1 0.75 0.54 0.46 RESIDUAL 60.83 44 1.38 80 Tables 5 and 6 show that no differences within sub-groups assigned to the different treatment combinations was evident in either the high or low anxious groups with respect to their self-reported level of social anxiety. The range of scores for the LSA group i s much more constricted (0-3) than the range for the HSA (12-28). This reflects the nature of the scale in which low scores represent the absence of social anxiety and high scores indicate the presence of social anxiety which can be calibrated according to the intensity of the experience. Of the 48 HSA scores used in this study only nine registered at 20 or above, a range representing extreme levels of social anxiety. B. Subject's Age Two factors which could possibly influence the results of a social memory task, subject's age and intelligence, were assumed to be controlled for by the process of random assignment. In addition, as mentioned in the Subject Allocation section, some attempt was made to assure that older subjects were distributed across a l l conditions. Statist ical comparisons were made post-hoc, however, to insure lack of influence of these two variables. To examine whether groups within treatment conditions differed with regard to age in either of the social anxiety group samples, a 3-way ANOVA (Encoding condition x Retrieval condition x Anxiety Group) was performed analyzing the age variable. Table 7 presents the means and standard deviations for the two anxiety groups across treatment conditions. 81 Table 7 Means and Standard Deviations of Subject Age for Anxiety Group by Encoding by Retrieval TREATMENT COMBINATION HIGH SOC. ANX. LOW SOC. ANX. M S.D. M S.D. SE/SR 21.00 2.13 21.75 3.38 SE/PR 21.16 3.61 20.33 2.10 PE/SR 20.25 2.45 19.41 1.37 PE/PR 21.16 3.45 21.33 2.96 SE=social encode, SR=social retrieve, PE=private encode, PR=private retrieve. The results for the Anxiety Group x Encoding x Retrieval ANOVA for subject's age yielded no significant differences for main effects or factors in interaction. C. Subject's Estimated High School Grade Point Average As a check on the possible influence of intellectual differences that might have occurred between subgroups, the subject's estimated High School Grade Point Average (GPA) was used as a rough index of intellectual achievement. A 3-way ANOVA (Encoding condition x Retrieval condition x Anxiety group) was performed. Table 8 presents the mean and standard deviations for this analysis. 82 Table 8 Means and Standard Deviations for Subject's High School Grade Point Average for Anxiety Group by Encoding by Retrieval TREATMENT COMBINATION HIGH SOC. ANX. LOW SOC. ANX, M S.D. M S.D. SE/SR 2.56 .65 2.66 .49 SE/PR 2.62 .74 2.54 .45 PE/SR 2.62 .48 2.68 .44 PE/PR 2.70 .43 2.40 .39 SE=social encoding, SR=social retrieval, PE=private encoding, PR=private retrieval. No stat ist ical differences with regard to estimated GPA across conditions or anxiety group classification emerged from the ANOVA treatment. Two factors may have influenced the homogeniety of these scores. The GPA figures were the subjects recollections of their high school performance, a situation which i s vulnerable to distortion. Secondly, the subject sample was attending a two-year <x>nmunity college and i t i s l ikely that a higher proportion of ' C students would be found in this population than of a population including a l l students attending post-secondary institutions. 83 D. Confederate Effect As a check on the relative uniformity of impact which assistants had on the subsequent recall abi l i ty of their subjects, a 4-way ANOVA (Confederate x Encoding condition x Anxiety Group x Retrieval condition) was performed for the central dependent variable, total number of items recalled. Table 9 reports the means and standard deviations for each anxiety group x encoding condition x retrieval condition combination as a function of the confederate factor. 84 Table 9 Means and Standard Deviations of Recall Totals for Anxiety Group by Encoding Condition by Retrieval Condition as a Function of the Confederate Factor. HIGH SOCIAL ANXIETY SE/SR SE/PR PE/SR Confederate A 7.33 (1.52) 6.00 (1.82) 8.43 (2.37) Confederate B 5.25 (1.89) 5.80 (2.86) 6.00 (2.12) Confederate C 5.80 (2.86) — — Confederate D — — — LOW SOCIAL ANXIETY SE/SR SE/PR PE/SR Confederate A 8.33 (.57) 7.17 (1.47) 7.00 (1.82) Confederate B 8.17 (2.04) 8.75 (2.63) 7.20 (2.17) Confederate C 9.33 (1.53) — — Confederate D — 7.00 (1.41) 7.00 (2.65) SE=social encoding, SR=social retrieval, PE=private encoding, PR=private retrieval. Standard Deviations in parentheses. 85 Table 9 indicates that no consistent pattern of recall performance i s attributable to the confederate factor. This is reflected in the 4--way ANOVA execution which yielded significant differences for the Anxiety Group main effect (F=7.18,p<.01) but no significant effect for the Confederate factor as a main effect or when interacting with group or treatment factors. Due to scheduling problems and idiosyncracies of the assistants timetables equal use of the four confederates was not possible. Two confederates were used proportionately more than the other two. These results provide some remote check on confederate performance which may have systematically biased the outcome measures. It can be concluded that any significant differences found are not attributable to a systematic Confederate effect. Analysis of the sub-group differences across treatment conditions and anxiety group classification revealed no systematic influence operating with regard to the subject's SAD scores, age, high school GPA, or to differences attributable to the experimental assistants. IV. PRE AND POST MEASURES OF SELF-REPORT STATE ANXIETY Self-ratings of state anxiety were obtained from each subject prior to the experiment and on completion of the experiment, and provided additional descriptive data relevant to the subject samples. The study assumes that social anxiety i s a situationally triggered 86 experience which may generate deficits in the information processing capacities of a socially anxious individual when that individual i s in social situations. Subjects who previously endorsed social anxiety scale items indicating apprehension in social encounters were placed in a interaction with a female stranger. It was expected that the high socially anxious subjects would find the social conditions more anxiety arousing than the private conditions. In addition, i t was expected that high socially anxious subjects would find the social conditions more arousing than would the low socially anxious subjects. Analysis of the pre-post anxiety scores lends partial support to these predictions. The means and standard deviations for the anxiety ratings in the four treatment combinations, at both pre and post intervals, for both high and low socially anxious subjects are presented in Table 1 0 . 87 Table 10 Pre and Post Anxiety Ratings of High and Low Socially Anxious Subjects for Each Encoding by Retrieval Treatment Combination (Means and Standard Deviations) HIGH SOCIAL ANXIETY LOW SOCIAL ANXIETY TREATMENT MODE PRE POST PRE POST SE/SR 23.17 25.25 17.75 17.67 (4.65) (7.08) (4.33) (5.28) SE/PR 23.92 25.50 19.08 18.75 (3.50) (7.13) (4.36) (5.77) PE/SR 19.33 20.08 19.42 17.50 (4.54) (7.28) (4.72) (5.04) PE/PR 21.58 18.67 15.92 15.17 (6.01) (3.82) (3.15) (4.32) SE=social encode, SR=social retrieve, PE=private encode, PR=private retrieve. Standard Deviations are in parentheses. 88 Inspection of Table 10 indicates that the mean anxiety ratings for the HSA group are higher than the ratings for the LSA group in a l l but one of the treatment combinations. A repeated measures ANOVA confirms this overall group difference as stat ist ical ly significant (F=26.06, p<.01). The experimental experience in general appears to have been more anxiety arousing for the HSA group. The elevated anxiety level is particularly marked for the HSA group when their encoding mode was a social one; this holds for post-experimental anxiety even when the HSA subjects were retrieving privately. In addition, the repeated measures ANOVA revealed a significant Encoding Condition main effect. Anxiety ratings are significantly higher (F=10.88, p. <.01) for both the low and high socially anxious subjects when they are encoding in a social mode. This indicated that the encoding manipulation was effective in arousing anxiety in the high socially anxious group, but also, unexpectedly, in the low socially anxious group as well . This anxiety response of low socially anxious subjects to social manipulation was also found by Smith,Ingram and Brehm (1983). The interaction between the Encoding condition and the Anxiety Group classification approached significance (p.=.07) but did not reach significance at the .05 level. We can conclude that both anxiety groups experienced the social encoding mode as more stressful than the private encoding mode, and that the HSA subjects experienced a l l 89 conditions as more stressful than did the low socially anxious subjects. Social manipulation at the retrieval phase did not show a corresponding effect. A possible explanation for this result could involve the temporal sequence of events. High anxious subjects had been participating in the experiment for 30 minutes without avoiding potentially arousing experiences i t i s possible anxiety levels may have attenuated some by this time. This difference between encoding and retrieval conditions may well reflect the reality of social encounters. As discussed in the literature review, anxiety has been shown to impact more dramatically at the encoding phase compared to the retrieval phase and this i s reflected in the predictions formulated for the present study. In support of this interpretation i t should be noted that subjects in the Private Encode/Social Retrieve condition encountered the social manipulation for the f i rs t time at the retrieval phase and yet no significant change in anxiety level i s apparent. Differences between pre and post anxiety ratings failed to reach significance when collapsed across Anxiety Group and Condition, or in any of the interactions with Anxiety Group or Conditions. It i s plausible that the impact on the HSA group was masked by an already elevated anxiety level prior to the experiment. The nature of the study as described in the recruitment leaflet and on the Consent Form suggested that heterosexual interactions were l ikely in the experiment. This description might well have created anticipatory 90 anxiety in subjects who admittedly find social contact anxiety-provoking. Support for this notion of anticipatory anxiety i s shown by comparing the pre-test scores in this study with results obtained elsewhere. The mean pre-test score for the high socially anxious group on the abbreviated version of the STAI-S inventory ranks at the 80th percentile for college students as reported by Spielberger (1983) in his normative data for the scale. Similar scores have been reported for clients entering treatment programs for remediation of inter-personal anxiety (Kanter & Goldfried, 1979). The anxiety ratings obtained from the subjects in the present study appear to be markedly elevated prior to the i n i t i a l experimental phase creating a ceil ing effect on the range of possible change in anxiety which the STAI i s able to register. One f inal comment with regard to the measurement of anxiety seems appropriate: the experience of anxiety i s such a complex, changing and interactive experience that attempting to measure subtle changes in such a linear fashion i s open to question. For example, high anxious subjects might well experience a sense of rel ief when they see themselves somehow coping in a dreaded social situation. In this example unchanged or even lowered anxiety ratings in the later experimental phase might be apparent, demonstrating the fluctuating reactions from sustained exposure to stressful experiences. In summary, pre and post measures of self-report state anxiety indicated that the HSA subjects were significantly more anxious than the LSA subjects in a l l of the treatment conditions. This finding 91 confirms the difference between experimental groups as determined by the Social Avoidance and Distress Scale. Additionally, the social encoding experience was identified as clearly the most anxiety-provoking condition for the HSA individuals and to a lesser extent for the LSA subjects as well . V. CODING AND JUDGING OF SUBJECT GENERATED RECALL ITEMS The Free Recall Items generated by each subject during the Retrieval phase of the experimental sequence were the only measures taken which could not be objectively scored or computed. These items required some evaluation and were reviewed and evaluated by three independent judges (one male and two female students). The criterion used for the acceptance or rejection of a generated recall production was l iberal , not requiring verbatim wording, but needing to contain "the thrust or essence of the message". This i s in keeping with previous findings that a listener's memory for sentences may be inaccurate at the level of word-for-word recal l , but accurate at the level of semantic paraphrasing (Hamilton,Katz & Leirer, 1980; Sachs,1967). As example of the criterion, the item "I usually try to avoid romantic involvements" was credited for the production "didn't want to get involved in any romantic situations", but not for the production "I haven't had a serious relationship". Prior to scoring, the judges discussed 20 items similar in content to the stimulus l i s t with regard to how this subjective criterion would be applied to these examples. Strong consensus was apparent. Each judge was to indicate for each item generated in the study "yes" or "no" 92 whether i t was an acceptable approximation of the original item. For f inal inclusion, items with "yes" responses from at least two of the three judges were accepted as reasonable replications of the original stimulus item. Table 11 shows the number of items generated and the consensus pattern reported by anxiety group classification. Table 11 Evaluation of Judges'Item Acceptability Rating Reported by Anxiety Group for Each Treatment Combination Treatment Acceptance Pattern Percent Combination 3/0 2/1 1/2 0/3 Acceptable HSA SE/SR 67 5 2 14 82% HSA SE/PR 68 3 5 20 73% HSA PE/SR 87 2 3 10 87% HSA PE/PR 121 5 5 19 84% LSA SE/SR 101 1 2 13 87% LSA SE/PR 85 7 3 11 87% LSA PE/SR 80 5 4 13 83% LSA PE/PR • 108 5 5 19 83% HSA=High Social Anxiety, LSA=Low Social Anxiety. SE=social encoding. SR=social retrieval, PE=private encoding, PR=private retrieval. 3/0=three yes votes for acceptance and 0 no votes; 0/3 i s the reverse. Total Number of Items Generated: HSA=436 LSA=462 Total Number of Items Acceptable: HSA=357 LSA=392 93 Unanimous consensus for items either accepted or rejected was 93% showing strong conceptual agreement. Of the total number of items generated (898) 83% were judged to be acceptable and this percentage is reflected consistent across groups within treatment conditions. Following the determination of item acceptability, the number of neutral, positive, and negative items, and the percentage of affective recall could be computed using the affective ratings determined in the Social Recall instrument standardization. Scores from a l l subjects were entered into a computer data f i l e and analyzed using the SPSS format. VI. ANALYSIS OF ITEM SERIAL ORDER EFFECT The f i rs t data treatment was a test of serial order effect. Lack of significance for this factor would allow the data matrix to be collapsed to a 2 x 2 x 2 design with n = 12 for each c e l l , a procedure valuable in gaining predictive power. Order effect was included in the design as a control for potential but unanticipated sequence effects. Table 12 reports the ANOVA results for differences in total number of items recalled for item order by encoding condition by anxiety group by retrieval condition. 94 Table 12 The Effects of Order of Items on the Recall Performance of High and Low Socially Anxious Subjects by Treatment Condition SOURCE OF VARIATION SS DF MS F PROB. F Anx. Group (A) 12.04 1 12.04 1.92 .17 Encoding (B) 57.04 1 57.04 9.10 .00* Retrieval (C) 28.16 1 28.16 4.49 .04* Item Order (D) 1.50 1 1.50 .24 .63 A x B 48.17 1 48.17 7.69 .01* A x C 3.37 1 3.37 .54 .46 A x D 1.04 1 1.04 .16 .68 B x C 57.04 1 57.04 9.11 .00* B x D 1.04 1 1.04 .16 .68 C x D .17 1 .17 .03 .87 A x B x D 1.50 1 1.50 .24 .626 B x C x D 5.04 1 5.04 .80 .372 A x C x D 2.04 1 2.04 .33 .570 A x B x C .00 1 .00 .00 1.000 A x B x C x D 2.67 1 2.67 .43 .516 WITHIN 501.00 80 6.26 *p < .05 Inspection of this table shows clearly that stimulus item sequence one i s not significantly more memorable than sequence two either as a main effect or when jointly interacting with any of the experimental conditions. The lack of significant differences with regard to item order allows for the collapsing of the 4-way factorial design to a 3 -way design. 95 CHAPTER V. EXPERIMENTAL RESULTS The present study uti l ized a completely crossed factorial design with multiple measures on memory retrieval, the dependent variable, requiring a multivariate stat ist ical treatment. As a l l of the dependent measures are responses taken from the same subjects, i t i s assumed, as suggested by Kerlinger and Pedhazur (1973), that the dependent measures are related in some unknown fashion. Multivariate analysis of variance allows for the testing of significant differences between means after the variance shared by the measures has been removed, thereby minimizing the possibil ity of spurious F ratios. Since some of the dependent measures are l ikely dependent upon each other (subscores of total scores) a single MANOVA including a l l dependent measures i s inappropriate. This analysis, therefore, involves two separate MANOVA analyses: one for the hypotheses regarding overall retention under the various treatment modes, and one for the hypothesis regarding differences in recall with respect to the content of the stimulus material. Hypotheses one and two predict differences in the amount of information remembered by socially anxious subjects under varied conditions, and hypothesis three predicts differences in the emotional content of the material remembered. 96 Hypothesis one predicts that memory deficits occur in socially anxious subjects as a result of situational factors, and therefore no global memory impairment w i l l be found when comparing high and low socially anxious subjects across the various treatment conditions. This prediction would be supported by the presence of an interaction effect between the anxiety group variable and either the encoding or the retrieval condition variables, and the absence of a main effect for anxiety group. Hypothesis two predicts that the situational factors occurring at the encoding phase w i l l be the c r i t i ca l ones precipitating memory di f f icult ies for social anxious subjects. This hypothesis would be supported by the presence of an encoding condition by anxiety group interaction. Hypothesis three predicts that the HSA group w i l l recall more negative information regarding others than w i l l the L S A group. This prediction would find support in a significant encoding by anxiety group interaction with the MANOVA executed for affective measures, and would be supported by the same interaction in the ANOVA for percentage of affective items recalled. Discussion of the results w i l l begin with an examination of the hypothesis regarding quantity of recall as assessed in MANOVA I , and then address the hypothesis regarding the recall of items according to affective tone. 97 MANOVA I: Analysis of Recall and Recognition Measures Three dependent measures were analysed f o r differences i n the number of sentences remembered. Table 13 presents the r e s u l t s of the MANOVA for the following memory measures: t o t a l number of items r e c a l l e d , corrected recognition scores, and the number of f a l s e p o s i t i v e i d e n t i f i c a t i o n s i n the recognition procedure. Means and standard deviations f o r the three dependent measures i n each of the anxiety group by encoding condition by r e t r i e v a l condition combinations are reported i n Appendix M. Table 13 MANOVA I: Analysis of Anxiety Group by Encoding Condition by Ret r i e v a l Condition E f f e c t s on Three Dependent Measures, Total R e c a l l , Total Recognition and Total False P o s i t i v e s SOURCE OF WILES APPROX. ERROR PROB. VARIANCE VALUE F DF F Anx. Gp. (A) .97 .99 86 .40 Encoding (B) .90 3.19 86 .03* R e t r i e v a l (C) .89 3.42 86 .02* A x B .91 2.79 86 .04* A x C .99 .35 86 .79 B x C .85 5.12 86 .00** A x B x C .99 .36 86 .78 * p r o b a b i l i t y < .05 ** p r o b a b i l i t y < .01 98 The significant differences indicated in Table 13 lend clear support to the hypotheses predicting memory deficit due to situational factors. No evidence of a global t ra i t deficit in memory for high socially anxious subjects emerged from this analysis. Hypothesis one, regarding the presence of a deficit situationally, i s confirmed in the presence of the Anxiety Group x Encoding interaction. Socially anxious males perform as well as non-anxious males on memory tasks involving the retention of person-relevant descriptions, when the circumstances of receiving and retrieving the information do not provoke social anxiety. Support for hypothesis two emerges from this analysis as well . Significant main effects were obtained for the Encoding factor indicating that the presence of the coactor during stimulus exposure significantly reduced the abi l i ty of subjects to recall the stimulus later. This would seem to be true for both anxiety groups, however, these findings must be tempered in the light of the significant interaction effects involving the encoding factor. The interaction with anxiety level would suggest that the impact on memory of the Encoding condition changes according to the subject's level of social anxiety. The interaction of the Encoding factor with the Retrieval factor suggests that the presence or absence of the coactor at the time of retrieval modifies the impact which the Encoding condition exerts by i t se l f . Further analysis i s necessary to identify (a) where specifically within the condition combinations the differences 99 occur, and (b) which of the multiple dependent measures account for the significant differences obtained. Following the significant MANOVA a discriminant analysis was done as a post-hoc procedure. As there were three dependent measures and two levels of the group variable the number of possible functions per factor (as determined by the lesser of the K- l computations) was one. A Bartlett test of significance indicated that each of the functions was stat ist ical ly significant. Table 14 presents the standardized discriminant function weights for each of the dependent measures in the MANOVA for each of the main effects or interaction effects that emerged as stat ist ical ly significant. Table 14 Standardized Discriminant Coefficients for the Statist ical ly Significant Main and Interaction Effects Reported in MANOVA I SIGNIFICANT MANOVA EFFECTS Recall DEPENDENT MEASURES Recognition False Positives Encoding (B) -1.01 - .08 - .07 Retrieval (C) .95 - .59 .38 Anxiety Gp. x Encoding (A x B) .88 - .08 .19 Encoding x Retrieval (B x C) 1.06 - .18 - .70 100 The relative weights for each measure as reported in Table 14 indicate that the total number of items recalled i s clearly the most significant contributor to the differences obtained between treatment combinations. Total recall accounts almost totally for the significance found in the Encoding factor as well as in the Encoding x Anxiety interaction. For the Retrieval condition the recognition score adds some strength to the discrimination, and in the Encode x Retrieval interaction the number of false positives i s a secondary contributor. Recall emerges as a much more powerful predictor than recognition, a finding which has implications for the identification of the site of memory deficit in the information processing chain. This topic w i l l be addressed in more detail in the Discussion section. Location of the cel ls within the significant interactions which account for the greatest difference can be achieved by computing means in discriminant space for each c e l l , plotting these means, and computing estimates of the ce l l contrasts. Table 15 shows the mean standardized discriminant scores for the Encode/Anxiety and Encode/Retrieval interactions; Figure 2 shows the plotting of these scores in standard deviations, and Table 16 presents the ce l l contrasts for these significant interactions. 101 Table 15 Computed Mean Standardized Discriminant Scores for the Statist ical ly Significant Interactions Reported in MANOVA I Using the Discriminant Coefficients from Table 15 Anxiety x Encoding Group Interaction HSA/SOC = .76 HSA/PRIV =-.16 LSA/SOC =-.44 LSA/PRIV =-.15 Encoding x Retrieval Interaction SOC/SOC = .10 SOC/PRIV = .44 PRIV/SOC = .38 PRIV/PRIV =-.91 102 Figure 2 Plotting of Discriminant Means for the Encoding by Anxiety Group Interaction and the Encoding by Retrieval Interaction Table 16 Cell Contrasts and Significance Estimates for the Anxiety Group x Encoding Interaction and the Encoding x Retrieval Interaction Using the Computed Standard Discriminant Means ANXIETY GROUP x ENCODING INTERACTION 95% Contrast Estimate of Confidence Significance Contrast Interval (p < .05) HSA SO/HSA PR .91 +.62 to +1.20 * HSA SO/LSA SO 1.20 +.91 to +1.49 * HSA SO/LSA PR .92 +.63 to +1.21 * HSA PR/LSA SO .29 .00 to + .58 n.s. HSA PR/LSA PR .01 - .28 to + .30 n.s LSA SO/LSA PR .28 - .01 to + .57 n.s HSA=high social anxiety, LSA=low social anxiety,S0=social encoding, PR=private encoding. ENCODING x RETRIEVAL INTERACTION Contrast 95% Estimate of Confidence Contrast Interval Significance (p < .05) SO SO/SO PR SO SO/PR SO SO SO/PR PR SO PR/PR SO SO PR/PR PR PR SO/PR PR - .34 - .28 1.01 .06 1.35 1.29 - .05 to - .63 +.01 to - .57 +.72 to +1.30 - .06 to +.35 +1.06 to +1.64 +1.00 to +1.58 n.s. * n.s. SO=SOCIAL, PR=PRIVATE Encoding condition is left of the slash, is to the right of the slash. retrieval condition 104 Figure 2 presents the plotting of the means for the Encode/Anxiety interaction; i t i s clear that the mean for the high SA subjects in the Social Encode condition i s notably distinct from the other three means. This distinction i s confirmed stat ist ical ly in the contrast table which shows that the only significant contrasts at the .05 level involve contrasts with HSA/SOC. This finding i s a strong endorsement of the central experimental hypotheses. The retention of social information by socially anxious males when encoding in a social situation i s demonstrated to be significantly poorer than their retention in a private encoding situation, and significantly poorer than the retention of low anxious males encoding in either a social or private condition. The significant main effects for the Retrieval factor is dependent on the interpretation of the Encoding/Retrieval interaction. As reported in Table 16 and depicted in Figure 2 above the Private Encode/Private Retrieve condition i s distinct from the other ce l ls . This finding is consistent with the experimental assumption that the Private Encode/Private Retrieve combination would be the least distracting condition. What i s of interest in this finding, however, is the indication that this relationship is true for both high and low anxious groups. The difference between means for the Social Encode/Social Retrieve and the Social Encode/Private Retrieve also registers as marginally significant stat ist ical ly . 105 MANOVA II: Analysis of Affective Content Recall Hypothesis three posits that socially anxious males selectively remember social information depending on i t s affective tone. Specifically i t was predicted that socially anxious males would recall more affective information and of the affective information recalled, more negative items than positive. The measures related to the third hypothesis include the percentage of positive items recalled, the percentage of negative items recalled, the percentage of neutral items recalled, and the percentage of the total recall score represented by affective items (both positive and negative). The latter score is a composite of the previous three scores and was analyzed separately using a three-way ANOVA (Encode condition x Anxiety Group x Retrieval condition). The positive, negative and neutral percentage scores were analyzed simultaneously in MANOVA II. Table 17 presents the results of this three-way MANOVA (Encode condition x Anxiety Group x Retrieval Condition) for the three affective dependent measures. Appendix N contains the means and standard deviations for the three dependent measures used in MANOVA II reported for anxiety group by encoding condition by retrieval condition. 106 Table 17 MANOVA II: Analysis of Anxiety Group x Encoding Condition x Retrieval Condition Effects on the Three Affective Dependent Measures (Percentage of Positive, Negative, and Neutral Items Recalled) SOURCE OF WILKS APPROX. ERROR PROB. OF VARIANCE VALUE F DF F Anx. Gp. (A) .95 1.46 86 .23 ns Encoding (B) .92 2.36 86 .08 ns Retrieval (C) .98 .70 86 .56 ns A x B .98 .63 86 .60 ns A x C .97 .78 86 .51 ns B x C .94 1.93 86 .13 ns A x B x C .96 1.18 86 . 32 ns Contrary to the expectations as reflected in hypothesis three, no differences with regard to the recall of affectively-toned items were found in any of the treatment combinations for either group. The reduction of total item recall as a result of the Encoding treatment x Anxiety Group interaction, as reported in MANOVA I, apparently occurred with a reduction of recalled items equally across a l l of the affective-tone sub-categories. The hypotheses predicting a greater retention of affectively-toned items and a selective preference of negative items in the high socially anxious group were not born out. No differences in the ratio of positive, negative and neutral items was found in either group as a result of the manipulation of 107 situational social anxiety at the encoding phase or at the retrieval phase. Analysis of Total Affective Recall Percentage A further attempt to identify differences in recall regarding affective material was made on a post-hoc basis. For this analysis a Total Affective Recall percentage score was formed by combining the percentage of positive items recalled from the total l i s t of items with the percentage of negative items recalled. This single measure was then analyzed using a three-way ANOVA (Anxiety Group x Encoding Condition x Retrieval Condition). The results of this ANOVA are reported in Table 18. Table 18 3-Way ANOVA for the Percentage of Affective Items (Positive plus Negative) Recalled by High and Low Socially Anxious Subjects as a Function of Encoding and Retrieval Conditions SOURCE OF VARIANCE SS DF MS F PROB. F Anx. Gp. (A) 810.84 1 810.84 3.71 .05* Encoding (B) 1211.26 1 1211.26 5.54 .02* Retrieval (C) 133.01 1 133.01 .60 .43 A x B 68.34 1 68.34 .31 .57 A x C 412.51 1 412.51 1.89 .17 B x C 1020.51 1 1020.51 4.67 .03* A x B x C 615.09 1 615.09 2.81 .09 RESIDUAL 19209.41 88 218.28 * pj< .05 108 Table 18 reveals significant overall differences in mean percentage of affective item recall for two of the three variables identified; Anxiety Group classification and Encoding condition. In addition, an interaction effect for Encoding x Retrieval modes was significant. The main effect for Anxiety Group i s surprising and contrary to the predictions of Hypothesis three. This result indicates that across a l l treatment conditions HSA subjects recall a lower, as opposed to the predicted higher, percentage of affective material (Mean=58.49%) than do LSA subjects (Mean=64.31%) whose score approximates quite closely the actual percentage of affective items in the original stimulus l i s t (66.66%). The HSA subjects appear to have selectively focussed on the neutral self-disclosures i n i t i a l l y or to have retrieved the neutral items in their search more effectively regardless of whether they heard the information publicly or privately or whether they were requested to recall that information alone or in the presence of the self-disclosing person. This finding, while contrary to the manner in which socially anxious individuals process self-relevant information, suggests that shy individuals may process information about others in a manner quite different from the way that they process information about themselves. Table 18 also reveals an overall effect for the Encoding Condition. If one does not consider the effect of retrieval for the moment, i t 109 can be seen that the recall of affective items i s significantly poorer when the information i s received in a social context. This relationship is consistent with the findings regarding recall in general as reported in MANOVA I and suggests that the presence of the female self-discloser not only inhibits recall overall, but also interferes with retention of affectively-toned messages specifically, and this latter relationship appears to apply to both anxiety groups. If method of retrieval i s considered, however, the superiority of one encoding condition over the other reverses. Table 19 reports the mean percentages for the significant Encoding x Retrieval (B x C) interaction. A clear indication of interaction effects i s obtained by calculating estimates of the interaction using ce l l means, marginal means, and the grand mean. Estimates of this interaction were computed and graphed in Figure 3. Table 19 Means for the Percentage of Affective Recall Reported for the Significant Encoding x Retrieval Interaction ENCODING CONDITION social (Bl) private (B2) RETRIEVAL CONDITION social (Cl) private (C2) 53.42 67.04 62.29 62.82 ROW MEANS 57.85 64.93 COLUMN MEANS 60.23 62.55 110 Figure 3 Graphing of the Computed Estimates for the Encoding x Retrieval Interaction MEAN PERCENTAGE AFFECTIVE RECALL CORRECTED FOR ROW, COLUMN AND GRAND MEAN EFFECTS social (Bl) private (B2) ENCODING CONDITION social retrieval (Cl) private retrieval (C2) The joint effects of encoding and retrieval manipulatons on the amount of affective items recalled i s presented in Figure 3. The relationship depicted indicates that the superiority of one level of encoding over the other i s confined to treatment combinations in which the manipulation at encoding and the manipulation at retrieval are consistent. The social encode/social retrieve combination registers the lowest level of affective item recal l ; this finding i s consistent with the results from the MANOVA I analysis. What i s puzzling here, however, i s the indication that the private condition at encoding intereacts jointly with the private condition at retrieval to disrupt recall of affective material. This finding i s 111 in contrast to the previous analyses and is not clearly interpretable. Additionally, the indication here that mixed combinations (combinations in which a social manipulation occurs either at encoding or at retrieval) interact to faci l i tate recall of affective items is in contrast to previous findings from MANOVA I which indicated a detrimental effect with the presence of the female confederate at encoding and a negligible effect of her presence at retrieval. The results from this interaction do not f i t consistently into a single theoretical framework and some aspects are not supported by any of the current formulations regarding the impact of anxiety on memory processes. At present no clear explanation i s obvious and conclusions drawn from these findings are seen as tentative. Thus, sunmarizing the results of the affective measures, a clear main effect for anxiety group indicates that HSA individuals recall significantly fewer emotionally-toned items. An overall difference indicating decreased recall of affective items with social encoding was indicated, but the significant interaction between the encoding and the retrieval conditions presents an ambiguous pattern of joint effects which is not clearly interpretable. IV. SUMMARY OF RESULTS In summary, hypothesis one predicting the presence of a deficit depending on situational factors, and the absence of a global t ra i t deficit in memory for socially anxious subjects i s supported. 112 Hypothesis two which predicted a significant impact for encoding manipulations but not for retrieval conditions when interacting with anxiety level is also supported. Based on the interaction effect we see that retrieval also plays a role. It may well be that the impact of retrieval on socially anxious subjects does not differ from the impact which retrieval conditions have on low anxious subjects, whereas encoding conditions appear to be selectively impairing high anxious subjects. In contrast to the prediction in hypothesis three of greater recall of affective material, particularly higher percentages of negative self-disclosures, by highly socially anxious subjects - the opposite was found. High SA males recalled significantly less affective material in a l l situations, social and private, than did low anxious subjects. This clear difference suggests that the hypervigilance to affective messages previously demonstrated in socially anxious populations occur only with self-referenced messages and information about others i s responded to in a uniquely selective manner. The social encoding phase, as in the analysis of overall retention, i s implicated as well in the memory disturbance for affective items specifically. More specific indications of treatment effects which contribute to differences in abi l i ty to recall affective material are ambiguous and uninterpretable. 113 CHAPTER VI. DISCUSSION OF RESULTS This study has demonstrated that socially anxious men experience problems with memory when they are in situations that arouse their anxiety. The findings are supportive of theorists who have postulated cognitive dysfunctions as components of social anxiety (Clark & Arkowitz, 1975; Goldfried & Sobocinski, 1975; Kanter & Goldfried, 1979; Leary, 1983). The results do not support the predictions from a Spence-Spence learning theory perspective which predicts enhanced memory as a result of anxiety-provoking conditions, but rather, contribute to the momentum of the cognitive perspective on emotional disorders which identifies internal processes as c r i t i ca l factors in the origin and maintenance of emotional and behavioral d i f f icu l t ies . MEMORY DEFICIT The memory dysfunction identified in the present study stands in contrast to previous findings in memory performance using socially anxious subjects. O'Banion and Arkowitz (1978) and Smith,Ingram and Brehm (1983) found no memory deficits , but did find enhanced memory for negative self-relevant information. The discrepancy in findings can clearly be understood by considering the important methodological differences (see Chapter 2) and in particular the difference in memory stimulus content used in this study as compared to the previous studies. A major departure from the O'Banion and Arkowitz study and the Smith, Ingram and Brehm study i s the focus here on 114 memory for information about other people as recalled by individuals who are made anxious by contact with other people. The present study has expanded the understanding of how highly socially anxious individuals process their experience, and i t suggests that processing information about others differs from the way HSA males process information about themselves. It i s plausible that the disorientation fe l t in social settings which triggers the self-referent hypervigilant posture, occurs because of an inadequate and impoverished representation of the external events. Such fragmented and barren social schemata leave the socially anxious person with limited information on which to base his social behavior. This view identifies an individual's memory for people and events as central to the appropriateness and effectivenesss of the on-going social behavior, and a determinant of the dysfunctional information sampling strategies which have been demonstrated in socially anxious populations. ENCODING OR RETRIEVAL Situational factors have emerged in this study as c r i t i c a l determinants of memory performance. This finding supports theorists who view dysfunctional behaviors associated with social anxiety as stenming from specific contextual circumstances and not from some consistent and long-term underlying dysfunctional personality characteristics. Support for a situational view i s consistent with the conceptualization of social anxiety as a state-specific experience (Leary, 1983; Mischel, 1973; Zimbardo, 1977) and has 115 direct implications for the appropriate focus of therapeutic interventions. Socially anxious men recalled less information about another person when they received that information in a social interaction with the self-discloser. This deficit i s not apparent when the HSA males received the same information in private. Additionally, no deficit i s apparent for LSA subjects in either the social or private information exposure conditions. The manipulation of social context at the time of information exposure resulted in a selective deficit for the high anxious males only. Manipulation of social context at the retrieval stages resulted in no corresponding defic i t . The lack of impact registering at the retrieval phase requires some qualification. Differences between the social encoding experience and the social retrieval experience could have potentially affected the outcome. Two points with regard to these differences require elaboration. One difference between the social manipulations at encoding and at retrieval has to do with the task demands. At encoding the subject has a brief interaction with a female stranger; at retrieval, there is less interaction but the subject i s required to report the remembered self-disclosures to the female stranger directly. Both of these task demands are interpersonally taxing, however, i t may be that one of them is more anxiety-provoking than the other. In the present study measuring this anxiety-provoking differential was not possible. Although i t appeared to the author possible to construct coherent rationales for the greater impact of either encoding or 116 retrieval phase anxiety, the consensus of current opinion (Eysenck, 1977; Greenberg & Safran, 1980; Mayes, 1983; Mueller, 1976) tends to see the early encoding phase as containing the processes most vulnerable to anxiety driven intrusions. The sequence of the encoding and the retrieval experiences in the overall experimental sequence i s a second major difference which, by definition, could not be controlled. It i s possible that the detrimental impact of social encounters occurs most intensely at the i n i t i a l contact phase which acts to overwhelm socially anxious individuals and intrude on their cognitive functioning. With time these individuals may recover, regain some composure, and adjust to the stressful conditions. If this factor i s in fact prominent, then the anxiety-provoking potential of the encoding phase i s l ikely to be much more powerful than the anxiety-provoking potential of the retrieval manipulation. Although this sequencing issue does not permit a test of equivalence at encoding and at retrieval, a hypothetical presence of greater stress at the encoding phase would l ikely be an accurate reflection of the social reality for socially anxious subjects. For this reason lack of experimental equivalence does not alter the general thrust of the interpretation of the present findings. The role of retrieval which emerges from the present study suggests that i t i s not as significant a phase with regard to the intrusive impact of social context as i s the encoding phase. The pattern of results implicates the early phase of information processing as the c r i t i ca l point contributing to later memory 117 defic i t . The results here support an attentional perspective which argues that increased self-preoccupation during social interaction wi l l reduce an individual's capacity to sample the available information in the environment. Additionally, however, an encoding perspective, which posits recall d i f f icul t ies due to incomplete and shallow i n i t i a l information encoding i s consistent with the present findings. ATTENTION VS DEPTH OF PROCESSING The encoding phase has been identified as having a major contribution to the memory dysfunction found in HSA men. Where in the encoding process does the deficit occur? As Hall (1982) points out no taxonomy of encoding processes exists, but two major aspects of this activity have received extensive consideration: the sampling processes and the transformational processes. As discussed earl ier, information sampling events are postulated by Wine (1976) to be dysfunctional in anxious populations. In addition to the transformational processes, Greenberg and Safran (1981) also include attention as an encoding process, asserting that attention i s guided by expectancy and i s therefore an active process in an individual's construction and reconstruction of reality. Alternatively, Craik and Lockhart (1972) have hypothesized cognitive events which organize, symbolize and transform information at input as being c r i t i ca l in determining the availabil ity of that information at output. These processes involve conscious association and elaboration of new information which deepens and enriches the memory 118 trace. With regard to the present data i t i s meaningful to ask whether the recall d i f f icul t ies experienced by HSA subjects are a function of reduced sampling of available information due to restricted attention deployment or are they functions of a failure to do something with the new information to insure that i t w i l l be available for retrieval at a later time? RECALL VS RECOGNITION Examination of differences in performance for recall tasks as compared to recognition tasks suggests that the deficit experienced by HSA individuals is not an attentional one. Although disagreement exists as to the processes underlying recall and recognition activit ies (Klatzky 1975); there i s considerable evidence to suggest that recognition performance tasks most often result in better memory than recall tasks. A dual-processing model (Kintsch 1971) would view recognition as a relatively pure measure of storage. In contrast, a depth of processing perspective views recognition as involving the processes of reconstruction and scanning, as does the process of recal l , but involves a less complex reconstruction of the episodic trace than i s required in recal l . Although Craik and Lockhart (1972) i n i t i a l l y argued that information encoded at shallow levels was less durable and therefore would be lost over time, their position has evolved with the refinement of their theory. Lockhart, Craik and Jacoby (1976) suggest the possibil ity that traces are equally durable and not lost from the system due to shallow encoding. They argue that shallowly encoded 119 traces, however, may be impossible to access as they are not semantically distinctive. Thus, shallow traces may be accessed when strong cues are available, as in recognition, but not when reconstruction is required with relatively l i t t l e stimulus available as in recall tests. Recognition requires relatively passive memory involving the presentation of the actual stimulus material as a cue for a decision regarding familiarity. Accurate recognition memory as found in the present study would suggest that the stimulus material did register and was stored in Long Term Memory. The material was necessarily attended to for this storage to occur. In the present study, as reported in the MANOVA I results, recall scores differed for high and low social anxiety groups, but recognition scores were not significant predictors in any of the group by treatment combinations. The mean percentage of correctly recognized items was 82.25% for the HSA and 86.50% for the LSA with no significant difference in any of the treatment variations. These percentages reflect a high degree of information retention for both groups when the retrieval requirements are relatively passive. Although recognition performance i s dependent on the nature and quality of the distractor items, very similar recognition performance has been obtained in other sentence recognition tasks (Shepard, 1967). These recognition data suggest that attention to the stimulus material does not appear to have been affected by the anxiety manipulations. It i s presumed that the process of environmental sampling involves attention deployment. 120 The recognition data indicate that attentional deployment was, apparently, uniformly on-task for both high and low anxiety groups in a l l conditions. The recall results have indicated that anxiety manipulations during the early exposure stage did result in a selective defic i t , and that this deficit was apparently not due to an attentional lapse. An alternative to the attentional hypothesis i s Mueller's (1976) encoding hypothesis which argues that information received i s not transformed or symbolized as completely for anxious individuals. Mueller adopts a depth of processing framework in conceptualizing anxiety-provoked memory deficits . From this perspective the indication is that HSA subjects receiving personal information from others directly, have diff iculty processing that information in meaningful and elaborate ways. Shy men would appear to attend only to superficial aspects of newly experienced material and f a i l to organize this experience around more personally or socially relevant themes. This description includes the process of attention deployment and assumes that the error occurs, not in the i n i t i a l scanning process, but in the subsequent articulation or highlighting of less personally relevant stimuli. The separation of attentional processes from the transformational encoding processes which occurs when contrasting Wine's model to Mueller's model may ,however, be an a r t i f i c i a l and spurious division. Depth of processing tasks as they have typically been conducted (Craik & Lockhart, 1972; Rogers, Kuiper & Kirker, 1977) require 121 subjects to consciously process information ( i .e . "Are these descriptions true in describing yourself?"). This conscious processing involves sustained inward attention for task completion. Although the present study does not address this issue directly, i t may be of value to speculate about the relationship between attentional processes and the 'depth of processing' model. Viewing an attentional deficit as including intrusions on internally focussed attention which disrupt the encoding processes in addition to the intrusions on externally focussed attention which result in disruptions of environmental sampling, may provide a refinement of attentional theory which would add to the understanding of depth of processing deficits . One's personal experience of attention deployment readily confirms that attention can be a rapidly fluctuating process which guides our momentary consciousness as the focus darts from point to point. It i s also apparent that the latency and redundancy of real l i f e events allow this to occur without significant loss.of information from any one source. A conversation with a friend might consist of considerable attention to the tone and content of the verbal message with brief departures to notice gestures, clothing, a passing aquaintance, or to the covert dialogue occurring simultaneously. The deployment can occur so quickly and smoothly as to give the impression of attending to two or more sources of information simultaneously, but the research on unattended information indicates that this is an inaccurate impression. It could be that the switch of attention to covert thoughts and images accompanying an experience may be a c r i t i ca l factor in the elaborate encoding of that 122 experience. There is evidence that 'deep' processing takes longer to complete (Craik , 1973) and that level of processing i s a function of imagery processes (Morris & Stevens, 1974) which presumably provide a richer array of salient attributes for the encoded stimuli. What i s proposed here is a model of flexible encoding whereby attention fluctuates rhythmically between sampling external information and reflecting on internal processing of that material. Any disruption of attention occurring either externally or internally could be detrimental to the completeness of the encoded message. In this context, socially anxious subjects would tend to process information while in a social interaction in a more peripheral mode; capable of adequate external attending but demonstrating a lack of sustained internal attention either from the intrusion of sel f -focussed cognitions or, perhaps, as an avoidance strategy. This dysfunctional process then would result in more barren representations of reality representations that are not easily accessible without considerable stimulus material to assist in the retrieval search. This speculation, distinguishing internal and external modes of attention, i s also compatible with the suggestion that the fate of information which i s not deeply or elaborately processed is not a failure to move from Short Term Memory (STM) to Long Term Memory (LTM) , but rather exists in LTM as "available" information but not "accessible" information to use the distinction of Tulving and Pearlstone (1966). Craik, Lockhart and Jacoby (1976) have incorporated this accessibility distinction into their evolving levels of processing model, a refinement which also accounts for the success of repetition and rote strategies in establishing LTM, a 123 point of concern for c r i t ics of Craik and Lockhart (Nelson,1977). Why do highly socially anxious individuals process information more peripherally? One explanation would suggest that the higher rate of self-pre-occupied runiinations compete with the processes necessary for deep processing of the new information, with the effect that the encoding of the new information i s sporadic and incomplete. A second possibility might involve the individual's motivation. Schmidt (1983) showed that expectancy of recall at a later date significantly affected the retention level of sentences. It i s consistent with the emerging picture of self-schemata for HSA individuals to postulate that these individuals may have a pessimistic outlook regarding the possibil ity of developing relationships and, in fact, believe that there w i l l be l i t t l e chance of having to recall the other-related personal information in the future. This expectancy posture leads to a more passive attentional style with minimal sustained inner focus and consequently results in ill-formed cognitive representations. IMPACT OF SOCIAL CONTEXT Encoding information socially regardless of the social context of retrieval produces a deficit in socially anxious subjects only. Retrieval of information socially regardless of the social context at input produces no corresponding memory def ic i t . The interaction between encoding context and retrieval context, however, created deficits in a l l of the combinations involving some social manipulation. This result reflects the fact that the Private 124 Encode/Private Retrieve combination resulted in the highest retention scores in both the high and low social anxiety groups. The social manipulations employed in this study had an impact on low anxious subjects as well as high anxious subjects. This finding reinforces the assumption underlying the experimental design that social context is a more powerful and taxing experience than the totally private conditions. Within this indication of impact, however, the high anxious group departed from the impact fe l t by the low anxious group in a marked way when they received the information in the social context. It appears that meeting people and hearing about them in a personal context impacts on everyone to some degree; however, socially anxious subjects show a marked reduction in their functioning in that social encounter. AFFECTIVE MESSAGES Analysis of the recall performance with regard to the affective context of the material revealed patterns in contrast to the experimental predictions. High socially anxious subjects were found to remember significantly fewer emotionally toned self-disclosures than did the low anxiety group. This finding does not support the prediction that socially anxious individuals are hypervigilant to negative information, and again suggests that socially anxious subjects may process other-referent information in quite different ways than they do sel f -referent information. 125 The emotional tone of stimulus material can be seen as a feature to be processed at a deeper level. As contrasted with the phonological features of language, for instance, emotional messages inherently involve aspects of meaning. As Zajonc (1980) points out, consistent findings from the literature using the Semantic Differential to map cognitive space indicate that meaning i s highly saturated with affect. The selective focussing on neutral information demonstrated in socially anxious males further refines the nature of the memory deficit identified in the analysis of overall retention. It appears that socially anxious men selectively attend to neutral features of information about others and selectively avoid affectively-ladened information about others, in a l l conditions in which they receive and retrieve that information. In contrast to the results from "MANOVA I regarding total item recal l , which indicated a memory deficit for HSA individuals situationally dependent on the social context at the information input stage; the pattern for affective information reveals a consistent inabil i ty to recall affectively-toned information independent of the encoding or retrieval context. Shy men may exhibit a chronic predisposition to avoid emotionally ladened messages regarding others even when experiencing those messages in the relatively non-threatening contexts of totally private input and retrieval. The consistent preference for neutral material about others over emotionally-toned messages indicates a stable predisposition operating in addition to the demonstrated situational effect. 126 It would appear that HSA males process fewer affective items on a consistent basis and that a deficit in overall information retention may occur under certain taxing social circumstances. From a therapeutic intervention perspective, the present study indicates that more stable aspects of perceptual style need to be addressed when treating social anxiety, in addition to the regulation of situational anxiety and remediation of dysfunctional behaviors which are tied to contextual cues. These results indicate that for socially anxious males at least, one's self-schema may not be the organizing framework used to process data about others as suggested by Markus (1980). Several schema theorists (Landau & Goldfried, 1981; Neisser, 1976; Taylor & Crocker, 1981) have described schemata as representing an efficient way to organize, encode, and retrieve new information. With regard to remembering new information, Hastie and Kumar (1979) have demonstrated that information which is compatible with existing well-formed schemata is more l ikely to be recalled. As the se l f -schema for most people is particularly well-formed i t can provide a quick and elaborate structure for managing information about others. "When we evaluate an object or event we are describing not so much what i s in the object or in the event, but something that i s in ourselves." (Zajonc,1980) Markus (1980) asserts that the self-schema is the single most well-defined cognitive network a person possesses and i t i s through this self-schema that we organize our perceptions of others. The indication from the present study suggests that HSA males may be at a particular disadvantage in their social functioning i f they, in fact, are not able to,avail themselves of the useful 127 functions which their own self-schema could provide in forming impressions of others. High socially anxious subjects who have been shown to be hypervigilant to emotional material about themselves appear to emphasize more neutral features when viewing others. As in the analysis of overall retention, the recognition data would suggest that such a selective mechanism is not occurring at the environmental sampling level, due to regulation of external attention, but, more l ikely involves the elaborative processes occurring subsequent to the sampling process. The tendency to selectively focus on neutral material i s consistent with divergent perspectives which hold that anxious individuals tend to avoid stimuli which heightens their anxiety (Freud, 1936; Wolpe, 1958). In this context emotionally-toned information would be seen as anxiety provoking and thus leading to either avoidance behavior or ego defense manuveurs, depending on which theoretical approach i s endorsed. Thus, shy men may acknowledge emotional input (environmental sampling phase), but avoid more elaborate consideration of that input (transformational phase) in order to modulate their own levels of anxiety. This view i s consistent with the postulation by Zajonc (1980) of a pre-attentive process involving affective encoding. Applied to the emerging picture of socially anxious individuals in this research, i t would seem that the i n i t i a l perception of affective meaning regarding others could trigger a negative feeling (possibly at some preconscious level) which then 128 directs attention away from the affective material and results in selective focussing on the more neutral information available. An alternative explanation to the avoidance hypothesis which could account for the selective attention to neutral information in HSA individuals is the notion that shy males go to considerable lengths to be non-judgemental, fa i r and unobtrusive in their affairs with others. This possibility would hold that shy men upon hearing emotional information about another person would not be l ikely to encorporate this more subjective information into their i n i t i a l impression formation due to this desire to allow others privacy around sensitive areas and to remain fa i r . Zimbardo (1977) has described this aspect of shyness as one of the positive features of shyness in the eyes of others and in the minds of the shy individuals themselves. While the intent of this posited unobtrusive interpersonal posture appears to be one of considerateness, the consequences nevertheless would be interference in the completeness of the encoded message. Regardless of the explanations for this demonstrated information processing tendency of HSA subjects, the tendency i s l ikely to be a socially dysfunctional one. A shy male meets a woman for the f i r s t time and experiences some neutral as well as some emotionally-toned information about her; at a second encounter this individual demonstrates a reduced abi l i ty to recall information about this new acquaintance. This scenario contains the seeds of social failure. Without an accurate and 129 available framework for this relationship to build on, i t w i l l become increasingly d i f f icu l t to deepen the contact, allow for the development of intimacy, and to begin to develop a common history. Memory, and particularly social memory, i s c r i t i ca l in maintaining and enhancing social relations. The selective loss of memory becomes an alienating process and perpetuates the self-defeating cycles identified in socially dysfunctional populations. MEMORY DEFICIT, PERFORMANCE DEFICIT, OR COGNITIVE DISTORTION Discussion of memory involves making assumptions about inferred inner processes based on some observable performance. Before examining the possible implications a memory deficit accompanying high social anxiety would have, i t would be useful to address two alternative explanations to the recall differences obtained which could challenge the validity of the memory interpretation: (1) the possibility of a performance deficit in which reporting accessed memories is inhibited, and (2) the possibil ity of the recall deficit reflecting distortions in thought rather than an inabil i ty to access material. Goulet and Massei (1969) have shown that anxious college students performing a paired-associate learning task may withhold responses to the stimuli unti l they experience a greater level of certainty regarding the correctness of their response, than do low anxious subjects. This posture would be consistent with social anxiety studies which have indicated a high fear of negative evaluation component which would elevate the need to be correct. It i s 130 possible that the HSA subjects in this study were somehow reluctant to report the stimulus material which they accessed due to fear of negative evaluation, social inhibition, embarrassment over having to report personal information about others, or some other rationale for non-compliance. Although this explanation cannot be dismissed absolutely, considerable evidence suggests that i t i s an improbable one. To begin with, examination of the certainty ratings which each subject completed for each of their generated recollections reveals no difference in the mean certainty rating between the low and high anxiety groups in any of the Encoding x Retrieval combinations (Anxiety Group main effect F=.724, p=.39; Anxiety Group x Encoding F=.004, p=.952; Anxiety Group x Retrieval F=.092, p=.762; Anxiety Group x Encoding x Retrieval F=1.197, p=.270). No tendency i s evident here for high anxious subjects to offer responses reflecting higher certainty levels. Secondly, a performance deficit explanation in which greater certainty is predicted for HSA subjects would expect a higher percentage of correct responding for this group—the opposite trend was found, although i t did not reach stat ist ical significance. Inspection of the ratios of correctly recalled items to the total number of items generated produced an accuracy rate of 81% for the HSA subjects and 85% for the LSA subjects. Withholding less certain responses is not reflected in these figures where the HSA subjects actually generated marginally more inaccurate responses. 131 A performance deficit explanation based on the assumption that HSA individuals are less comfortable communicating the recalled personal material to a female stranger, would predict a lower production rate in the social retrieval mode for the HSA subjects than in the private retrieval mode. This interpretation i s not supported by the results. Although the private encode/private retrieve condition was clearly the superior mode for memory retention for both anxiety groups (HSA M=12.50, LSA M=11.42), i t i s interesting to note that the second best performance mode for the HSA group was the private encode/social retrieve condition (HSA M=8.50) producing slightly higher recall than the social encode/private retrieve (HSA M=8.08). These findings are contrary to the performance deficit expectations which see the social retrieval context as a response inhibiting one. The recall performance of the HSA group in the private encode/social retrieve condition was equal to that of the LSA group (M=8.50). The absence of recall differences between anxiety groups for the private encode/social retrieve condition coupled with the absence of markedly inhibited recall performance in the private encode/social retrieve mode when compared to the other treatment conditions for the HSA group, argues against the performance deficit explanation and highlights the negative impact of the social encoding experience. The above data arguing against a performance deficit interpretation are consistent with impressionistic data obtained during the retrieval task and the debriefing phase. Confederates reported that subjects frequently exhibited the "aha" experience as they recognized items on the recognition form which they had failed to recal l . 132 During the debriefing, subjects were asked i f they had held back any responses or felt reluctant to offer items which had occurred to them. No subject indicated that they had withheld any recollections, however, several subjects indicated that reporting some of their recollections was awkward and embarrassing at times. It appears that the motivation to not offend the confederate by verbalizing personal material about her (which would lead a shy person to inhibit the reporting of certain responses) was possibly cancelled out by a motivation to not offend the confederate by failing to remember important information she had shared. A different alternative explanation to the memory deficit position could argue that the stimulus material was equally accessed by both anxiety groups, but this information was significantly distorted in the reporting process by the HSA group so that items did not meet the criterion used for judging a response to be correct; an error of commission rather than omission. This explanation i s conceptually a more difficult one than the performance deficit alternative as the process of memory cannot be completely isolated from the dysfunctional mediations which exaggerate, distort, or omit detail on a selective basis. In fact, i t may be that the cognitive distortions (such as those described as occurring in depression by Beck [1971]) may have at their basis an impoverished memory system, and i t may be that distortions in meaning are really occurring at the memory trace encode phase. As memory and thought are not mutually exclusive processes this alternative is not possible to address directly. Based on the data presented above (addressing the performance explanation) which indicate that the percentage of accurate 133 responses was high for the HSA group (81%) and not significantly different from the accuracy rate for the LSA group (85%), the dominant role of cognitive distortion i n the demonstrated r e c a l l d e f i c i t has been ruled out i n the present analysis. As memory can be seen as a reconstructive process, i t i s r e a l i s t i c to assume, however, that on-going cognitive distortions interweave and p a r a l l e l the memory processes, which are seen as central from the perspective of the present study. CLINICAL IMPLICATIONS The image of shy men functioning i n new social situations which emerges from this examination of social memory contributes to the growing understanding of the dysfunctions associated with social anxiety. The following scenerio may be typical: A social anxious male i s introduced to a female student. Immediately the familiar signs of anxiety surface: the heart i s pounding faster, breathing becomes quicker and more shallow, the stomach tightens and suggests that he might throw-up at any minute. There i s a slight dizziness which interrupts concentration and a sense of pre-occupation with nothing i n particular which destroys his a b i l i t y to attend to this new acquaintance. The self-dialogue begins: "Does she notice my nervousness?", "Will I blow i t again?". There i s a tremendous urge to just get up and get out of there. His eyes dart quickly, taking i n her face, her clothes, her hands, and momentarily her eyes. He hears himself engaging i n superficial small talk i n response to her comments; this makes him smile nervously and evaluate his own performance: "You phoney!" As the f i r s t few minutes end, the anxiety l i f t s s l i g h t l y . They have gotten by the f i r s t few painful moments and the conversation i s a b i t more specific, a b i t deeper now. She i s t e l l i n g 134 him what she has been occupied with recently: She has grown apart from a close friend, she feels moodier these days, she' l l have to get some orange juice that always cheers her up, that reminds her she needs to stop at the store on the way home. These messages are heard, but they register on a more meaningful level only sporadically. Responding to the meaning of her conments he is hampered by intrusions of thoughts about himself, about how he must be perceived by her. He searches his awareness for appropriate self-disclosures to match hers in order to keep the momentum of the conversation intact, but he comes up empty. He experiences that sinking feeling again, and a subtle fatalism that te l l s him once more this painful experience isn' t going to go anywhere his motivation drops noticeably and there i s a sense of withdrawing. As the wave of personal and not so personal detail from his partner washes over him, he begins to construct an impression of his new acquaintance. Even in his somewhat disengaged state her conments which imply happy or sad experiences pique his anxiety and make him feel vulnerable he tr ies not to think about them much. He focusses on the safer neutral ground. Even though his new acquaintance was responsive, and conmunicated at a level suggesting interest and trust, this socially anxious fellow departs depressed about his performance and depressed about his social future. The next day a chance encounter allows these two people to talk with each other again. Unable to access much of their conversation from before and unable to volunteer much about his own l i f e , he feels the encounter to be awkward and distant. This coolness after a real attempt on the part of the other to make contact i s experienced as disinterest, even rejection. The grounds for growing intimacy are lost in his blurred and barren recollection of their encounter only a day before. There i s almost a sense of unrealness about their previous meeting as they depart, and this individual i s left with a partly wistful and a largely sel f -accusatory frame-of-mind. 135 The above scenerio describes how the memory dysfunctions apparent in socially anxious populations contribute to these individuals' social distress. The identification of memory processes as central components in the experience of social anxiety has important general as well as specific treatment implications. THE ROLE OF MEMORY IN THE TREATMENT OF SOCIAL ANXIETY It could be argued that many of the previous cognitive deficits identified as features of social anxiety (negative self-schemata, attributions, self-statements) are a l l contingent on this more fundamental process of memory. The content of current cognitions is dependent on current perceptual input combined with information previously deposited in memory. Irrational beliefs which may guide selective perception, the content of on-going internal dialogue, and the expectancy of failure a l l rely on previous experience. If i t were possible to miraculously banish previous memory from our lives then the dysfunctional expectations and thoughts would have no substance or origin. For this reason i t becomes increasingly important to examine memory d i f f icul t ies accompanying social anxiety. It may be that factoring out a memory component from other cognitive dysfunctions would eliminate what i s now seen as a variety of independent dysfunctional processes. Focussing on the sampling, encoding, and retrieval of social information may prove to be a more f ru i t fu l approach than addressing more peripheral effects of these central processes. 136 Although the focus of the present research was a descriptive one, aimed at articulating more ful ly the cognitive processes underlying social anxiety, memory enhancement interventions for the treatment of social anxiety are clearly suggested in the present conceptualization. A cognitive-behavioral approach to social anxiety would view anxiety as a by-product of dysfunctional cognitions (irrational beliefs, negative internal dialogue, distortions in construal). This relationship has been demonstrated by researchers who have reduced anxiety through cognitive restructuring (Kanter & Goldfried,1979; Meichanbaum,1979). Once this process has begun, a self-perpetuating cycle may exist whereby faulty cognition leads to anxiety and heightened anxiety, in turn, interferes with cognitive processing. The present results have demonstrated that the arousal of social anxiety in socially anxious subjects does inhibit memory. The second aspect of the proposed cycle, in which deficient memory processing leads to anxiety, i s a consistent implication of the present research, but has not been addressed directly in this study. An interesting direction for future research would involve a demonstration of this postulate of social anxiety resulting from faulty memory process. This faulty schema/social anxiety cycle i s seen to contribute to the development and perpetuation of social anxiety in three possible ways. F i rst , as mapped out in Chapter Two, memory i s identified as a central component of interpersonal competency. The inabil i ty to recall personal information about others impedes the development of 137 intimacy by creating the impression of inattention and disinterest, and by failure to encorporate important information into the context of the growing relationship. The second manner in which memory impairment may contribute to social anxiety involves the perpetuation of self-defeating behavior cycles. In this context behavior i s seen to be schema-driven. Behavior becomes inappropriate and ineffective when the guiding schemata are significantly unrepresentative of the social reality. Socially anxious individuals are seen to sample their environment adequately, but f a i l to construct associatively-rich structures which organize their experiences adequately. The behavior of socially anxious persons is then based on what incomplete information i s accessible, and their behavior i s frequently off-target. The f i r s t two ways in which memory i s proposed to impact on the experience of social anxiety are largely related to interpersonal performance deficits . In contrast, a third possibil ity i s cast in more intrapsychic terms. This notion holds that, in large part, i t is not one's experiences themselves which may be anxiety-provoking, but rather the individual's construal of those experiences. This position is a central focus of cognitive restructuring interventions. In addition to the behavioral deficits discussed above, the inappropriate social schemata resulting from selective and shallow memory processing may exacerbate the experience of anxiety. In the above conceptualizations i t i s assumed that the awkwardness and paralysis described by socially anxious clients i s not primarily 138 a s k i l l s def ic i t , but arises from the misinterpretation of their experiences. As i s frequently acknowledged in a communications framework, the inattention to or misreading of facial expression, gesture, voice tone, or message content can significantly distort the conmunication. What is suggested in this speculation i s a distortion in memory structures occurring from the failure to encode social information elaborately which, in turn, leads to sel f -defeating behaviors and anxiety-provoking cognitions. Memory di f f icult ies contribute to ill-formed and unrepresentative schemata through omissions which impoverish one's access to the complexities of social experience. Additionally, selective attention to only parts of the experience during the encoding of those memories serves to distort the integrity of the original event. With regard to HSA clients, addressing memory function in social encounters may be a valuable way to enhance communication s k i l l s , and to begin to understand the dysfunctional cognitions which contribute to the self-defeating behaviors. MEMORY ENHANCEMENT INTERVENTION: POSSIBLE NEW DIRECTIONS FOR THE TREATMENT OF SOCIAL ANXIETY Treatment strategies which address memory deficit or schema distortion in social anxiety are not presently part of treatment programs designed for this client group. Memory interventions can be logically extrapolated, however, from the results of this study. The following discussion presents some speculations regarding treatment possibi l i t ies which focus on memory dysfunction as a source of social anxiety. 139 The results from this study indicate that memory deficits are occurring in the recall process, but recognition of experienced material was relatively intact. The findings suggest that the data for more elaborate social schemata are available in long-term memory, and although not presently accessible, could possibly be made accessible through appropriate c l in ica l strategies. The results here are encouraging from a therapeutic perspective. These results imply that HSA clients have registered the necessary information for understanding their world in a rich and complex way, but these individuals have failed to actively ut i l i ze that information to form schemata which accurately reflect their reality. If the HSA client acts as i f his constricted representations of social encounters are accurate, his behavior w i l l be inappropriate, awkward, and ineffective, and his prospects for satisfying interpersonal contact w i l l be discouraging. Interventions aimed at unearthing the unaccessed material and using that material to embellish, vivify and refine the existing barren representations are indicated. Guided imagery techniques (Bandler and Grinder, 1977; Singer, 1978), Evocative Reflection (Rice,1974), Focussing (Gendlin, 1980), or hypnosis may serve as specific manuveurs in helping HSA clients to enrichen their constrictive social schemata and enhance the fullness and complexity of their internal representations of social experiences. The present data suggest that ful ler representation of experiences are available on a recognition basis but not when freely recalled. Therapeutic interventions which can provide stimulus cues to hook the dormant memories may be useful in luring those memories into awareness. The use of concrete objects 140 (photos,mementos, ticket stubs), other sensory cues (a particular perfume, a piece of music) or a return to the context of the social encounter may provide the catalyst for contacting new material to be processed during counselling sessions. Consistent with this notion of stimulus cued recall and memory enhancement is the notion of state-dependent memory (Bower,1981). Inducing a previously experienced feeling of panic, depression, or humiliation in a controlled counselling environment may provide a mood stimulus cue which w i l l faci l i tate the socially anxious client in accessing important memories. Interventions which are aimed at gaining increasing control of attentional processes also are indicated. In this category would be Gestalt exercises which are designed to increase awareness of attentional deployment. John Stevens (1971) has identified several exercises in his book Awareness which are geared to exploring, experimenting with, and ultimately controlling one's conscious awareness. Stevens identifies internal, external, and fantasy zones as possible areas of focus, and he encourages an individual to develop a capability to f luidly and flexibly 'shuttle' between these zones at w i l l . Prototypical attentional training procedures found in the test anxiety literature (Sarason,1972) may be appropriately adapted to socially anxious populations. With socially anxious individuals, practice at intentionally maintaining an external focus would be beneficial in not only increasing contact with the environment, but also in decreasing self-rumination and thereby reducing anxiety. 141 Exercises designed by Meichenbaum (1977) and others which faci l i tate an individual's awareness of on-going internal dialogue would be advantageous to highly social anxious clients as well. This intervention could be followed by training to deliberately move away from such self-ruminations. Secondarily, focussing on the negative quality of these ruminations would draw attention to the selective and self-defeating schemata which guide the socially anxious cl ient 's perceptions and actions; a c r i t i ca l step in the eventual restructuring of these dysfunctional processes. With regard to specific memory strategies, i t may be possible to develop a specific memory enhancement program for socially anxious clients which would be designed to increase elaborative rehearsal and encourage richer and deeper encoding. Just as the stage mneumonist has honed certain mneumonic strategies to enhance his memory to a remarkable degree, so could the socially anxious individual learn relevant cognitive manuveurs to enhance memory of social information. This intervention secondarily encourages more outward attention coupled with a deliberate inward manipulation and, i f successfully accomplished, i s incompatible with extensive attention to self-dialogue. A program of social memory enhancement could involve a series of experiences which contain gradually increasing memory demands. Exercises involving retention of information about others found in print, audio or videotape mediums could be gradually accelerated to include practice in encoding and retrieving information in controlled 142 social situations and ultimately rea l - l i fe social encounters. In addition to the socially anxious cl ient 's performance, the feelings and thoughts arising from these memory exercises would provide valuable material for counselling sessions. A close examination of how the client selectively processes new information about others may reveal systematic distortions or omissions operating in social encounters which lead to self-defeating interpersonal behaviors. Evidence from the present study indicates that the socially anxious client is l ikely to access a much more constricted range of information about others than do non-anxious individuals, and in particular, to omit the more meaning-laden affective messages. Demonstrating this memory bias in a counselling context and discussing the interpersonal impact that i t may have are seen as valuable therapeutic objectives. In addition to the observable improvement in social memory s k i l l s , a memory enhancement intervention is l ikely to have positive spin-off value in other areas as well, such as modifying the cl ient 's success expectancy, increasing the cl ient 's self-esteem, and motivating the client to seek out social encounters. A third area of intervention suggested by the findings of this study involves treatments with an affective emphasis. There i s some indication from the present study that highly socially anxious subjects possess a relatively stable pre-disposition to avoid affective content in messages from others. Treatment strategies which would faci l i tate an individual's comfort with affective material are indicated. A variety of intervention strategies are available which would address this therapeutic goal, including: 143 Gestalt techniques aimed at expanding contact with body experiences which are seen as central to feeling states; individual psychodynamic procedures; or specifically tailored group experiences in which the amount of self-disclosure which the client is encouraged to share i s gradually increased. DIRECTIONS FOR FURTHER RESEARCH Several unanswered questions have surfaced, in the course of the present research, which warrant further investigation. To begin, a deficit in recall of information about others was demonstrated in this study, in apparent contrast to previous research which has identified only a hypervigilance to negative self-referent information. An important missing component here has to do with the memory performance of high socially anxious individuals with regard to non-evaluative autobiographical information. How well do these individuals recall the incidental details of their own lives? Findings in this area would contribute significantly to the understanding of self-schema formation in socially anxious populations. Additionally, demonstrating experimentally that dysfunctional memory can cause anxiety in socially anxious populations would confirm the postulated cognitive deficit/anxiety cycle. A design which would treat memory as the independent variable with dependent measures taken on dimensions of anxiety i s suggested. This demonstrated relationship would add significant strength to the memory/social anxiety relationship which emerged in the present research. 144 Finally, investigation of individual differences within the classification of high socially anxious individuals would be useful in refining the target population for which social memory i s particularly troublesome. This direction of research i s in keeping with the current trends in therapy research toward achieving a better match between particular clients and particular interventions. The classification of social anxiety contains sufficient multiplicity of cause and multiplicity of experiences to warrant careful inspection for within group variation. Examination of correlates associated with socially anxious subjects exhibiting high retention and low retention of socially received information could reveal important refinements which are relevant to treatment. CONCLUSION The awkward social behavior, low self-esteem, and excruciatingly painful feelings of embarrassment and anxiety which are characteristic of socially anxious individuals may, in part, be understood as a result of the dysfunctional ways in which these individuals encode and retrieve their social experiences. The present study has demonstrated that high socially anxious males recall significantly less information about others i f they experience that information in a social situation. Of particular significance is the indication from this study that emotionally-toned messages are recalled significantly less than are neutral messages. Additionally, the indication here suggests that the demonstrated memory deficit i s l ikely to be a failure to encode experiences 145 elaborately, which results in poorly-formed and impoverished schemata for organizing and understanding social experience. The identification of a memory deficit occurring in the encoding phase for HSA clients offers a new focus for therapeutic attention with this client population. This new focus also suggests specific intervention possibilities which could facilitate socially anxious clients in contacting and utilizing the inaccessible memories which create and perpetuate their interpersonal distress. 146 FOOTNOTES 1 Although there i s some controversy as to the most efficient way to correct for inflated recognition performance (Brown,1976) this method has been used frequently as a speedy correction for recognition data. The implication underlying this correction that memory i s an all-or-none phenomenon i s not optimal, however, the procedure was deemed adequate in the present study as recognition was not the central measure of retention and the signal detection corrections were judged to be too time intensive. Additionally, i t i s the difference in recognition percentages between subject groups which i s of interest in this study rather than an absolute index of recognition efficiency. 147 REFERENCES Ansbacher, H.L. Adler's place today in the psychology of memory. Journal of Personality, 1947, 3, 197-207. Argyle, M., Trower, P, & Bryant, B. Explorations in the treatment of personality disorders and neuroses by social s k i l l s training. Brit ish Journal of Medical• Psychology, 1974, 47, 63-72. Argyle, M. & Kendon, A. The experimental analysis of social performance. In L. Berkowitz (Ed.), Advances in  experimental social psychology (Vol. 3). New York: Academic Press, 1967. Arkowitz, H., Lichtenstein, E., McGovern, K, & Hines, P. The behavioral assessment of social competence in males. Behavior Therapy, 1975, 6, 3-13. Arkowitz, H., Hinton, R., Per l , J . , & Himadi, W. Treatment strategies for dating anxiety based on real l i f e practice. Counselling Psychologist, 1978, 7, 41-46. Bandler, R. & Grinder, J . The structure of magic J_. Palo Alto, Cal if : Science and Behavior Books, 1975. Bandura, A. Principles of behavior modification. New York: Holt, Rinehart & Winston, 1969. Borkovec, T.D., Stone, N.M., O'Brien, G.T., & Kaloupek, D.G. Identification and measurement of a c l in ica l ly relevant target behavior for analogue outcome research. Behavior  Therapy, 1974, 5, 503-513. Breck, B. & Smith, S. Selective recall of se l f -descriptive traits by socially anxious and nonanxious females. Social Behavior and Personality, 1983, 11(2), 71-76. Brower, P.E., & Mueller, J.H. Serial position effects in inmediate and f inal recall as a function of test anxiety and sex. Bulletin of the Psychonomic  Society, 1978, 12, 61-63. Brown, J . An analysis of recognition and recall and of problems in their comparison. In R. Brown (Ed.), Recall  and Recognition. New York: John Wiley & Sons, 1976T Buros, O.K. (Ed.) The eighth mental measurements yearbook. Hyde Park, N.J.: Gryphon Press, 1978. Byrne, D., McDonald, R.D., & Mikawa, J . Approach and avoidance af f i l ia t ion motives. Journal of Personality, 1963, 31, 21-37. 148 Cacioppo, J.T., Glass, C R . , 8c Merluzzi, T.V. Self-statements and self-evaluations: a cognitive-response analysis of heterosocial anxiety. Cognitive Therapy  and Research, 1979, 3(3), 249-262. Caldwell, A., Calhoun, K., Humphreys, L., & Cheney, T. Treatment of socially anxious women by a s k i l l s training program. Journal of Behavior Therapy and  Experimental Psychiatry, 1978,~1F, 315-320. Cattell R., Personality and Mood by Questionnaire. San Francisco: Jossey Bass, 1973. Christensen, A., Arkowitz, H. & Anderson, J . Beaux for cupid's errors: practice dating and feedback for college dating inhibitions. Behavioral Research & Therapy, 1975, 13, 321-331. Clark, J . & Arkowitz, H. Social anxiety and self-evaluation of interpersonal performance. Psychological Reports, 1975, 36, 211-221. Craighead, W.E., Kimball, W., & Rehak, P. Mood changes, physiological responses, and self-statements during social rejection imagery. Journal of Consulting and Cl inical Psychology. 1979, 47, 385-396. Craik, F.I.M. A 'levels of analysis' view of memory. In P. Pliner, L. Krames, and T. Alloway (Eds.), Communication  and Affect. London, Academic Press, 1973. Craik, F.I.M. & Lockhart, R.S. Levels of processing: a framework for memory research. Journal of Verbal  Learning and Verbal Behavior, 1979, 11, 671-684. Curran, J .P . An evaluation of a s k i l l s training program and a systematic desensitization program in reducing dating anxiety. Behavior Research and Therapy, 1975, 13, 65-68. Sk i l l s training as an approach to the treatment of heterosexual-social anxiety: a review. Psych  Bulletin, 1977, 84, 140-157. Curran, J . P . , Gilbert, F., & L i t t l e , L. A comparison between behavioral training and sensitivity training approaches to heterosexual dating anxiety. Journal of Counseling  Psychology, 1976, 23, 190-196. Curran, J . P . , & Gilbert, F. A test of the relative effectiveness of a systematic desensitization program and an interpersonal s k i l l s training program with date anxious subjects. Behavior Therapy, 1975, 6, 510-521. 149 Diggory, J.C. Self-evaluation, concepts and studies. New York: Wiley, 1966. Dutta, S. & Kanungo, R. Affect and Memory: A Reformulation. Toronto: Pergamon Press, 1975. Easterbrook, J.A. The effect of emotion on cue uti l ization and the organization of behavior. Psychological  Review, 1959, 66, 183-201. Erdelyi, M.H., & Goldberg, E.B. Let's not sweep repression under the rug: toward a cognitive psychology of repression. In J.F. Kilstrom & F.J. Evans (Eds.) Functional disorders of memory. Hil lsdale, New Jersey: Erlbaum, 1979"! Erikson, E. Identity: Youth and Cr is is . New York: Norton, 1968. Eysenck, M. Human memory; theory, research and individual  differences. Toronto: Pergamon Press, 1977. Fishbein, M. A consideration of beliefs and their role in attitude measurement. In M. Fishbein (Ed.) Readings in  attitude theory and measurement. New York: John Wiley and Sons Inc., 1967. Fisher, R. Interaction between encoding distinctiveness and test conditions. Journal of Experimental Psychology:  Human Learning and Memory, 1981, 7, (4), 306-310. Freedman, D. An ethological approach to the genetical study of human behavior. In S.G. Vandenburg (Ed.), Methods  and goals in human behavior genetics. New York: Academic Press, 1965. Freud, A. The ego and mechanisms of defense. New York: International University Press, 1946. Gendlin, E. Focusing. New York: Everest House, 1978. Girodo, M. , Dotzenroth, S . , & Stein, S. Causal attribution bias in shy males: implications for self-esteem and self-confidence. Cognitive Therapy and Research, 1981, 5, 325-338. Glasgow, R., & Arkowitz, H. The behavioral assessment of male and female social competence in dyadic heterosexual interactions. Behavior Therapy, 1975, 6, 488-498. 150 Glass, C.R., Gottman, J .M. , & Shmurak, S.H. Response acquisition and cognitive self-statement modification approaches to dating s k i l l s training. Journal of  Counseling Psychology, 1976, 23, 520-526. Glass, C.R., & Merluzzi, T.V. Cognitive assessment of social evaluative anxiety. In T.V. Merluzzi, CR. Glass, & M. Genest , Cognitive assessment. New York: Guilford, 1981. Galassi, J . P . , & Galassi, M.D. Modification of heterosexual s k i l l s deficits . In A. Bellack and M. Hersen (Eds.), Research and practice in social s k i l l s training. New York: Plenum Press, 1979. Goldfried, M., Linehan, M., & Smith, J . Reduction of test anxiety through cognitive restructuring. Journal of  Consulting and Cl inical Psychology, 1978, 46, 32-39. Goldfried , M. & Sobocinski, D. Effects of irrational beliefs on emotional arousal. Journal of Consulting  and Cl inical Psychology, 1975, 43(4), 504-510. Goldfried, M., & Robins, C. Self-schemas, cognitive bias and the processing of therapeutic experiences. In P.C. Kendall (Ed.) Advances in cognitive behavioral research  and therapy (Vol. II). New York: Academic Press, 1982. Gottesman, I. Differential inheritance of the psychoneuroses. Eugenics Quarterly, 1962, 9, 223-227. Goulet, L.R., & Mazzei, J . Verbal learning and confluence thresholds as a function of test anxiety, intelligence, and stimulus similarity. Journal of  Experimental Research in Personality, 1969, 3, 247-252. Greenberg L., & Safran, J . Encoding and cognitive therapy: changing what clients attend to. Psychotherapy:  Theory, Research and Practice, 1981, 18, 163-169. Gross, T.F., & Mastenbrook, M. Examination of the effects of state anxiety on problem-solving efficiency under high and low memory conditions. Journal of Educational  Psychology, 1980, 72(5), 605-609. Hal l , J.F. An invitation to learning and memory. Boston: Allyn and Bacon, 1982. Hamilton, D., Katz, L., & Leirer, V. Cognitive representation of personality impressions: organizational processes in f i rs t impression formation. Journal of Personality and  Social Psychology, 1980, 39, 1950-1963. 151 Hastie, R., & Carlston, D. Theoretical issues in person memory. In R. Hastie, T. Ostrom, E. Ebbesen, R. Wyer, D. Hamilton, & D. Carlston (Eds.), Person memory: the  cognitive basis of social perception. Hillsdale, New Jersey: Lawrence Erlbaum, 1980. Hastie, R. & Kumar, P. Person memory: personality traits as organizing principles in memory for behaviors. Journal of Personality and Social. Psychology, 1979, 37, 25-38. Hatvany, N.G. Shyness, emotional arousal, and information  processing. Unpublished masters thesis, Stanford University, 1976. Hersen, W., & Bellack, A .J . Assessment of social s k i l l s . In A.R. Ciminero, K.R. Calhoun, & H.E. Adams (Eds.), Handbook of behavioral assessment. New York: John Wiley and sons" ,1977. Higgins, E.T., Herman, C P . , & Zanna, M.P. (Eds.) Social  cognition: the Ontario symposium. Hillsdale, New Jersey: Erlbaum, 1981. Johnson, R.E. Memory based rehearsal. In G.Bower (Ed.), The  psychology of learning and motivation (Vol. 14). New York: Academic Press, 1980. Kanter, N. & Goldfried, M. Relative effectiveness of rational restructuring and self-control desensitization in the reduction of interpersonal anxiety. Behavior  Therapy, 1979 ,10, 472-490. Kerlinger, F.N., & Pedhazur, E.J. Multiple regression in behavioral research. New York: Holt, Rinehart and Winston, Inc., T973. Kihlstrom, J.K. On Personality and memory. In N. Cantor, & J.F. Kihlstrom (Eds.), Personality, cognition and social interaction. Hil lsdale, New Jersey: Erlbaum, T58T. Kintsch, W. The representation of meaning in memory. Hillsdale, New Jersey: Erlbaum," 1974. Klatzky, R. Human memory: structures and processes(2nd, ed.) San Francisco: W.H. Freeman, 1980. Kraft, T. Social anxiety model of alcoholism. Perceptual  Motor S k i l l s , 1971, 33, 797-798. Kupke, T.E., Hobbs, S.A., & Cheney, T.H. Selection of heterosocial s k i l l s . Behavior Therapy, 1979, 10, 327-335. 152 Landau, R., & Goldfried, M. The assessment of schemata: a unifying framework for cognitive, behavioral, and traditional assessment. In P.C. Kendal & S.D.Hollon (Eds.) Assessment strategies for cognitive-behavioral interventions. New York: Academic Press, 1981. " Lazarus, R., & Opton, E. The study of psychological stress: a summary of theoretical formulations and experimental findings. In CD. Spielberger (Ed.) Anxiety and behavior. New York: Academic Press, 1966. Lawlis, G.F. Response styles of a patient population on the fear survey schedule. Behavior Research and Therapy, 1971, 9, 95-102. Leary, M. Understanding social anxiety: social personality and c l in ica l perspectives. Beverly H i l l s , Ca l i f . : Sage Library "of Social Research, 1983. Ledwidge, B. Cognitive behavior modification: a rejoinder to Locke and to Meichenbaum. Cognitive Therapy and Research, 1979, 3(2), 133-140. Levitt, E.E. The psychology of anxiety (2nd ed.) Hillsdale, New Jersey: Erlbaum, 1980. Lewinsohn, P.M., & Amenson, C.S. Some relations between pleasant and unpleasant mood-related events and depression. Journal of Abnormal Psychology, 1978, 87, 644-654. Lingle, J .H. , Dukerich, J .M. , & Ostrom, T.M. Accessing information in memory based impression judgements: incongruity versus negativity in retrieval selectivity. Journal of Personality and Social Psychology, 1983, 44, 262-272. " " Lockhart, R., Craik, F., & Jacoby, L. Depth of processing, recognition and recal l . In R. Brown (Ed.) Recall and Recognition. New York: John Wiley and Sons, 1976. Locke, E. Behavior modification i s not cognitive - and other myths: a reply to Ledwidge. Cognitive Therapy and  Research, 1979, 3(2), 119-125. MacPhillamy, D.J., & Lewinsohn, P.M. Manual for the pleasant events schedule. University of "Oregon, 1976. Mandler, G. Mind and emotion. New York: John Wiley and sons, 1975"! Markus, H. Self-schemata and processing information about the self . Journal of Personality and Social  Psychology. 1977, 35, 63-78. 153 Markus, H. and Smith, J . The self thought and memory. In D.M. Ulegner & R.R. Vallacher (Eds.) The self in social psychology. New York: Oxford, 1980. Marshall, P.H., & Werder, P.R. The effects of the elimination of rehearsal on primacy and recency. Journal of Verbal Learning and Verbal Behavior, 1972, 11, 649-653. ' " " " ~ Martinson, W., & Zerface, J . Comparison of individual counselling and a social program with non-daters. Journal of Counseling Psychology, 1970, 17, 36-40. Mayes, A. (Ed.). Memory in animals and humans: some  comparisons and their theoretical implications. Wokingham, Birk. : Van Nostrand, ReinhoYt, 1983. Meichanbaura, D. Cognitive behavior modification: the need for a fairer assessment. Cognitive Therapy and  Research, 1979, 3 (2 ) , 127-132. Cognitive modification of test anxious college students. Journal of Consulting and Cl inical  Psychology, 1972, 39, 370-380T Meichanbaum, D., Gilmore, B., & Fedoravicious, A. Group insight v.s. group desensitization in treating speech anxiety. Journal of Consulting and Cl inical  Psychology, 1971, 36, 410-42T: Mischel, W. Toward a cognitive social learning reconceptualization of personality. Psychologica1  Review, 1973, 80, 252-283. Personality and assessment. New York: John Wiley and Sons, 1968. Mischel, W., Ebbesen, E., & Zeiss, A. Determinants of selective memory about the self . Journal of Consulting and Cl inical Psychology, 1976, 44, 92-103. Morris, P.E., & Stevens, R. Linking images and free recal l . Journal of Verbal Learning and Verbal Behavior, 1974, 13, 310-3T5. Mueller, J.H. Anxiety and cue ut i l izat ion in human learning and memory. In M. Zuckerman, & CD. Spielberger (Eds.) Emotions and anxiety: new concepts, methods, and  applications. Hil lsdale, New Jersey: Erlbaum, 1976. Nasby, W., & Yando, R. Selective encoding and retrieval of affectively valent information: two cognitive consequences of children's mood states. Journal of  Personality and Social Psychology, 1982, 43, 1244-1253. 154 Neisser, U. Cognition and Reality. San Francisco: W.H. Freeman, 1976. Nelson, CR. Guide to LERTAP use and interpretation. University of Otago, New Zealand: Department of Education, 1974. Nelson, D., Reed, V.S., & McEvoy, C L . Doubts about depth. Journal of Experimental Psychology: Human Learning and  Memory, 1979, 5, 24-44. O'Banion, K., & Arkowitz, H. Social Anxiety and selective memory for affective information about the self . Social Behavior and Personality, 1977, 5(2), 321-328. Osgood, C., & Suci, G. The measurement of meaning. Urbana: University of I l l ino is Press, 1957. Ostrom, T.M., Pryor, J .B . , & Simpson, D.D. The organization of social information. In E.T. Higgins, C P . Herman, & M.P. Zanna (Eds.). Social cognition: the Ontario  Symposium (Vol. 1). Hil lsdale, New Jersey: Erlbaum, 1981. Paivio, A. Mental imagery in associative learning and memory. Psychological Review, 1969, 76, 241-263. Paivio, A., Yui l le , J . C , & Madigan, S.A. Concreteness, imagery, and meaningfulness values for 925 nouns. Journal of Experimental Psychological Monograph  Supplement, 76, 1968. Phi l l ips , G. Help for shy people. Englewood C l i f f s , New Jersey: Spectrum, 1980. Pilkonis, P. Shyness, public and private, and i t s relationship to other measures of social behavior. Journal of Personality, 1977, 45, 585-595. The behavioral consequences of shyness. Journal of  Personality, 1977, 45, 596-611. Rehm, L.P. , & Marston, A.R. Reduction of social anxiety through modification of self-reinforcement. Journal of  Consulting and Cl in ical Psychology, 1968, 32, 565-574. Rice, L.N. The evocative function of the therapist. In D.A. Wexler & L.N. Rice ( Eds.) Innovations in c l ient - centered therapy. New York: John Wiley and Sons, 197?: Richardson, F . C , & Tasto, D.L. Development and factor analysis of a social anxiety inventory. Behavior  Therapy, 1976, 7, 453-462. 155 Richardson, J.T.E. Mental imagery and human memory. London: MacMillen Press Ltd. , 1980. Rogers, CR. Client centered therapy. Boston, Houghton Mi f f l in , 1951. Rogers, T.B., Kuiper, N.A., & Kirker, W.S. Self reference and ther encoding of personal information. Journal of  Personality and Social Psychology, 1977, 35, 677-688. Sachs, J .S. Recognition memory for syntactic and semantic aspects of connected discourse. Perception and  Psyche-physics, 1967 , 2 , 437-442. Saltz, E. Manifest anxiety: have we misread the data?. Psychological Review, 1970, 77, 568-573. Sarason, I.G. Experimental approaches to test anxiety: attention and the uses of information. In CD. Spielberger (Ed.) Anxiety: current trends in theory  and research (Vol 2). New York: Academic Press, 1972. Schlenker , B.E., & Leary, M.R. Social anxiety and self presentation: a conceptual model. Psychological  Bulletin, 1982, 92, 641-649. Schmidt, Steven. People expecting recall and recognition employ different encoding processes. Memory and  Cognition, 1983, 11,172-180. Sears, D.O. Social anxiety, opinion structure and opinion change. Journal of Personality and Social Psychology, 1967, 7, 142-151. Sheehan, P.W. A shortened form of Bett's questionnaire upon mental imagery. Journal of Cl inical Psychology, 1967, 23, 386-389. Sheppard, R.N. Recognition memory for words, sentences, and pictures. Journal of Verbal Leaning and Verbal  Behavior, 1967, 6, 156^163. Shields, J . Monozygotic twins: brought up apart and brought  up together. London: Oxford University Press, 1962. Siegel, J . , & Loftus, E. Impact of anxiety and l i f e stress upon eyewitness testimony. Bulletin of the  Psychonomic Society, 1978, 12(6), 479-480. Singer, J . , & Pope, K. The power of human imagination. New York: Plenum Press, 1978. 156 Spielberger, C., State-trait anxiety inventory: a comprehensive bibliography. Palo Alto, California: Consulting Psychologists Press Inc., 1983. Spielberger, C., Goresuch, R., & Lushene, R. Manual for the  state-trait anxiety inventory (self-evaluation  questionnaire). Palo Alto, California: Consulting Psychologists Press Inc., 1970. Spielberger, C., Goresuch, R., Lushene, R., Vagg, P. & Jacobs, G. Manual for the state-trait anxiety  inventory:STAI (form Y). Palo Alto, California: Consulting Psychologists Press Inc., 1983. Smith, R.E., & Sarason, I. Social anxiety and the evaluation of negative interpersonal feedback. Journal of  Consulting and Cl inical Psychology, 1975, 43, 429. Smith, T.W., Ingram, R.E., & Brehm, S.S. Social anxiety, anxious self-preoccupation and recall of self-relevant information. Journal of Personality and Social  Psychology, 1983, 44, 1276-1283. Spence, K.W. A theory of emotionally based drive (d) and i t s relation to performance in simple learning situations. American Psychologist, 1958, 13, 131-141. Anxiety (drive) level and performance in eyelid conditioning. Psychological Bulletin, 1964, 61, 129-139. Spence, K.W., Farber, I.E., McFann, H.H. The relation of anxiety (drive) level to performance in competitional and non-competitional paired-associates learning. Journal of Experimental Psychology, 1956, 52, 296-305. Stevens, J.O. Awareness: exploring, experimenting,  experiencing. Moab, Utah: Real People Press, 1971. Taylor, S.E. , & Crocker, J . Schematic bases of social information processing. In E.T. Higgins, C P . Herman and M.P. Zanna (Eds.) Social cognition: the Ontario  symposium. ( Vol. 1). Hil lsdale, New Jersey: Erlbaum, 1981. Thorpe, G. Desensitization, behavioral rehearsal, self instructional training and placebo effects on assertive-refusal behavior. European Journal of  Behavior Analysis and Modification, 1975, 1_, 30-44. Trexler, L.D., & Karst, T.O. RET, placebo and no-treatment effects on public speaking anxiety. Journal of  Abnormal Psychology, 1972, 79, 60-67. 157 Trower, P. Situational analysis of the components and processes of behavior of socially ski l led and unskilled patients. Journal of Consulting and Cl inical  Psychology, 1980, 48 (1), 327-339. Tulving, E., & Pearlstone, Z. Availabil ity versus accessibility of information in memory for words. Journal of Verbal Learning and Verbal Behavior, 1966, 5, 381-391. Twentyman, C., & McFall, R.M. Behavioral training of social s k i l l s in shy males. Journal of Consulting and  Cl inical Psychology, 1975 , 43 , 384-39"5. Watson, D., & Friend, R. Measurement of social-evaluative anxiety. Journal of Consulting and Cl inical  Psychology, 1969, 33, 448-457. Weiner, B. Theories of motivation: from mechanism to  cognition. Chicago: Markham, 1972. Wine, J.D. Test anxiety and direction of attention. Psychological Bulletin, 1971, 76, 92-104. Evaluation anxiety: a cognitive-attentional construct. In H.W. Krohne & L.C. Laux (Eds.) Achievement,  stress, and anxiety. Washington, D.C: Hemisphere, 1980 Wolfe, J . L . , & Fodor, I.G. Modifying assertive behavior in women: a comparison of three approaches. Behavior  Therapy, 1977, 8, 567-574. Wolpe, J . Psychotherapy by reciprocal inhibition. Stanford: Stanford University Press, 1958. Zajonc, R.B. Feeling and thinking: preferences need to interferences. American Psychologist, 1980, 35, 151-175. Zigler, E., & Phi l l ips , L. Social competence and the process-negative distinction in psychopathology. Journal of Abnormal and Social Psychology, 1962, 65, 215-222. Zimbardo, P. Shyness, what i t i s , what to do about i t . London: Addison-Wesley Publishing Co., 1977. Zubrycki, CR. ,& Borkowski, J.G. Effects of anxiety on storage and retrieval processes in short-term memory. Psychological Reports, 1973,33, 315-320. 158 APPENDIX A SOCIAL RECALL STIMULUS LIST 1. My family moved to this area when I was three. (o) 2. Writing letters to friends i s more trouble than i t i s worth. (-) 3. I love having long discussions with friends. (+) 4. Saturday is the day I do my vacuuming. (o) 5. I'm always wil l ing to admit when I've made a mistake. (+) 6. I feel wonderful when I am praised by someone. (+) 7. When I'm riding a bus, I find that my mind wanders. (o) 8. People around me seem to be quite c r i t i ca l of me lately. (-) 9. I usually try to avoid romantic involvements. (-) 10. I get a lot of pleasure from playing with children. (+) 11. I noticed a new building going up on 6th Street. (o) 12. I rather that people left me alone. (-) 13. I must remember to get some milk on the way home. (o) 14. I feel l ike I am growing apart from a close friend of mine. (-) 15. I get tremendous satisfaction from doing a job well . (+) 16. I am thinking about sell ing my car. (o) 17. I enjoy quiet dinners with friends. (+) 18. I frequently get quite moody. (-) 19. I used to prefer apple juice to orange juice. (o) 20. I am afraid that I won't do my job properly. (-) 21. I have always loved to have my back massaged. (+) 22. I have yet to meet a man that I was really attracted to. (-) 23. People often feel comfortable confiding in me. (+) 24. I sometimes feel hungry when i t isn ' t mealtime. (o) Order A: 1 through 24 Order B: 13 through 24, 1 through 12 159 APPENDIX B RECOGNITION FORM 1. Working in the garden i s a real pleasure f o r me. 2. My family moved to this area when I was three. 3. I am thinking of s e l l i n g my car. 4. Keeping a diary gives me a l o t of s a t i s f a c t i o n . 5. I enjoy seeing travel photos of foreign countries. 6. I seem to get sick frequently. 7. Writing l e t t e r s to friends i s more trouble than i t i s worth. 8. The beauty of sunsets i s largely overrated. 9. I am afraid that I won't do my job properly. 10. Sometimes I go to the movies by myself. 11. Saturday i s the day I do my vacuuming. 12. People often f e e l comfortable confiding in me. 13. I l i k e playing with soft kittens or puppies. 14. When I fee l l i k e getting away I li k e to read. 15. The taste of food i s a real pleasure to me. 16. I feel refreshed after a good nights' £;locp. 17. I enjoy quiet dinners with friends. 18. I often i n s i s t on haveng things my own way. 19. I sometimes f e e l hungry when i t i s n ' t mealtime. 20. My father i s i n s t a l l i n g a new garage door. 21. I have never bought a lottery t i c k e t . 22. I'm always w i l l i n g to admit when I've made a mistake. 23. I don't know why some people are so interested in music. 24. I need to pick up my dry cleaning today. 25. I f e e l wonderful when I am praised by someone. 26. I have switched brands of toothpaste recently. 27. I l i v e in an apartment. 28. At work I have to be near some obnoxious people. 29. I usually try to avoid romantic involvements. 30. I find cooking a drudgery. 31. People around me seem to be quite c r i t i c a l of me lacely. 32. I have yet to meet a man that I was r e a l l y attracted to. 160 APPENDIX B continued 33. I must remember to get some milk on the way home. 34. I seem to f i n d myself g o s s i p i n g f r e q u e n t l y . 35. I get a l o t of p l e a s u r e out o f p l a y i n g with c h i l d r e n . 36. P a r t i e s u s u a l l y bore me. 37. I n o t i c e d a new b u i l d i n g going up on 6th s t r e e t . 38. I •m not sure when c l a s s e s b e g i n . 39. I f e e l l i k e I am growing a p a r t from a c l o s e f r i e n d of mine. 40. I f r e q u e n t l y get q u i t e moody. 41. I look forward to a r e l a x i n g bath i n the evening. 42. I used to p r e f e r apple j u i c e to orange j u i c e . 43. I have o f t e n enjoyed f l i r t i n g with a man. 44. I get tremendous s a t i s f a c t i o n out of doing a job w e l l . 45. I r a t h e r that people l e f t me a l o n e . 46. I enjoy making t h i n g s f o r my b r o t h e r ' s k i d s . 47. I have always loved to have my back massaged. 48. I study i n the l i b r a r y a t l e a s t twice a week. 161 APPENDIX D Guidelines for the Confederate Interaction During the experiment you w i l l be introduced to a male subject whom you have never met, and the two of you w i l l be asked to chat together for a few minutes as i f you were getting to know each other for the f i r s t time. Following your chat you w i l l play the tape on which you have recorded the stimulus sentences, and following the tape you w i l l open the door to signal the experimenter that this phase i s done. Your role in this interaction i s a particularly important one in which we w i l l try to achieve an experience which closely approximate the real l i f e experience of talking to a stranger for the f i r s t time. As you are part of the experiment, you may possibly feel more in control than does the male subject, which could cause you to feel more confident, more at ease, or more assertive than you normally would. It i s important to be aware of these feelings and control them by adopting the role of a person who has found themselves in a new and unfamiliar interaction. It i s essential that you approach each subject in a consistent manner and that the subject believe that you are just another volunteer student who has been thrust into this situation as they have. Respond as 'naturally' as you can for this kind of situation. Do not be overly faci l i tat ive i f awkwardness occurs, nor aloof. Experience such awkwardness and respond as you would under similar but natural conditions. Although you w i l l find yourself l ik ing or disl iking certain individuals more than others, try to act consistently with a l l subjects. You w i l l be seated approximately 3 to 4 feet apart, be aware of your body posture, eye contact, and non-verbal behaviors (smiling, gesturing) in your attempt to be consistent. Your behavior w i l l be checked randomly to make sure that you are maintaining consistency. During the interaction you may talk about any subject that spontaneously emerges with a few exceptions. Do not talk about the experiment i tse l f or the room you are in . In addition, do not talk about subjects directly related to the content of the stimulus items on the tape. Topics around the areas of college l i f e , career plans, major f ie ld of study, or present job situation are suggested for i n i t i a l conversation starters. After approximately 5 minutes of conversation you then say: "I think he wants us to l isten to this tape I made earl ier ." Begin the tape and l isten to i t together; when the tape i s finished, stop the tape player and open the door to signal the experimenter that you have finished. 163 APPENDIX E BACKGROUND INFORMATION NAME AGE MARITAL STATUS: s i n g l e married separated/divorced FIRST LANGUAGE APPROXIMATE HIGH SCHOOL GRADE POINT AVERAGE (A=4.0, B=3.0, C=2.0, D=1.0) SEMESTER AT COLLEGE: 1st 2nd 3rd 4th NUMBER OF BROTHERS SISTERS . PRESENTLY EMPLOYED ? YES NO LAST MOVIE SEEN__ LAST BOOK READ LAST MUSIC EVENT ATTENDED 164 APPENDIX F CONSENT FORM This study, Impression Formation i n Shyness, conducted by L. Greenberg and E. Biggs, i s interested i n examining the processes by which shy and not-shy men form impressions of the women they meet. The study w i l l involve subjects l i s t e n i n g t o s e l f - d e s c r i p t i v e information e i t h e r from a person d i r e c t l y or on audiotape. Following t h i s , subjects w i l l be asked questions regarding t h e i r impressions. The e n t i r e procedure w i l l l a s t approximately 20 minutes. Each subject's responses w i l l be c a r e f u l l y c o n t r o l l e d to assure c o n f i d e n t i a l i t y and anonymity. Individual scores w i l l be coded and extracted f o r use as group data, and the o r i g i n a l information w i l l be destroyed. Any questions you have regarding t h i s study w i l l be answered f u l l y before the study begins, and opportunity to ask questions following p a r t i c i p a t i o n i s provided. I acknowledge that I have received a copy of t h i s consent form and I agree to p a r t i c i p a t e i n t h i s study with f u l l understanding that I may terminate and withdraw from the experiment at any time without being accountable f o r the reasons f o r such an action or without t h i s action influencing my student standing or future treatment i n any way. Signed 165 APPENDIX G: Private Encoding INSTRUCTIONS WE ARE INTERESTED IN HOW MEN FORM IMPRESSIONS OF THE WOMEN THEY MEET. ON THE TAPE IN FRONT OF YOU IS A RECORDING MADE BY A FEMALE STUDENT. SHE HAS WRITTEN DOWN AND RECORDED SEVERAL THINGS ABOUT HERSELF WHICH SHE IS WILLING TO SHARE. WHEN YOU ARE READY, PLEASE TURN ON THE RECORDER BY PRESSING THE 'PLAY' BUTTON AND LISTEN TO HER RECORDING CAREFULLY. LISTEN TO THE TAPE ONLY ONCE AND DO NOT STOP THE TAPE ONCE IT HAS BEGUN. THE RECORDING LASTS APPROXIMATELY 3 1/2 MINUTES AND A VOICE AT THE END WILL SIGNAL THE TIME TO STOP THE RECORDER. (PRESS THE BUTTON MARKED 'STOP') WHEN YOU HAVE FINISHED LISTENING TO THE TAPE AND HAVE STOPPED THE RECORDER, OPEN YOUR DOOR TO SIGNAL THE EXPERIMENTER THAT YOU HAVE FINISHED. Thank You. 1 6 6 APPENDIX H: Private Retrieval INSTRUCTIONS EARLIER YOU LISTENED TO A TAPE ON WHICH A FEMALE STUDENT RECORDED SEVERAL THINGS ABOUT HERSELF. IN THIS PHASE OF THE STUDY YOU ARE ASKED TO RECALL AS MANY OF THOSE STATEMENTS AS YOU CAN, AND TO WRITE THESE STATEMENTS DOWN ON THE FORM PROVIDED. TRY TO REMEMBER THE WORDING AS ACCURATELY AS YOU CAN, BUT WRITE DOWN ALL OF THE THINGS YOU REMEMBER EVEN IF THE WORDING DOESN'T SEEM EXACT. WHEN YOU HAVE LISTED AS MANY ITEMS AS YOU CAN REMEMBER IN THE 4 MINUTES ALLOTTED, PROCEED TO THE NEXT PAGE OF THE FORM WHICH DESCRIBES HOW EACH OF THE RECALLED STATEMENTS IS TO BE RATED. WHEN YOU HAVE FINISHED WITH THE STATEMENT RATING, OPEN THE DOOR TO SIGNAL THE EXPERIMENTER THAT YOU HAVE FINISHED. Thank You. 167 A P P E N D I X I T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F T/F SOCIAL BEHAVIO* SCALE I feel relaxed even In unfamiliar social situations. I try to avoid situations which force M to be very sociable. It Is easy for me to relax when I an with strangers. I have no particular desire to avoid people. I often find social occasions upsetting. I usually feel calm and comfortable at social occasions. I a* usually at ease when talking to someone of the opposite sex. I try to avoid talking to people unless I know then well. If the chance cones to Meet new people. I often take i t . I often feel nervous or tense in casual get-togethers in which both sexes are present. I am usually nervous with people unless I know them well. I usually feel relaxed when I am with a group of people. I often want to get away from people. I usually feel uncomfortable when I am in a group of people I don't know. I usually feel relaxed when I meet someone for the fir s t time. Being introduced to people mikes me tense and nervous. Even though a room is ful l of strangers, I may enter it anyway. I would avoid walking up and joining a large group of people. When my superiors want to talk with me. I talk willingly. I often feel on edge when I am with a group of people. I tend to withdraw from people. I don't mind talking to people at parties or social gatherings. I am seldom at ease in a large group of people. I often think up excuses in order to avoid social engagements. I sometimes take the responsibility for introducing people to each other. I try to avoid formal social occasions. I usually go to whatever social engagement I have. I find it easy to relax with other people. 168 APPENDIX J SELF—EVALUATION QUESTIONNAIRE Name Date DIRECTIONS: A number of statements which people have used to desc r ibe themselves are given below. Read each statement and then b lacken i n the appropr ia te c i r c l e to the r i g h t of the statement to i n d i c a t e how you f e e l r i g h t now, that i s , at t h i s  moment. There are no r i g h t or wrong answers. Do not spend too much time on any one statement but g ive the answer which seems to d e s c r i b e your present f e e l i n g s b e s t . S o NOT A' SOMEV >ERATEI -! s c Ci > <; a < X P CO O CO O 1. I f e e l © ® ® ?. I f e e l ® ® ® 3. I ® ® ® 4. I f e e l ® ® ® 5. I f e e l ® ® ® 6. I f e e l ® ® ® 7. I f e e l ® ® ® 8. I f e e l ® ® ® 9. I ® ® ® 10. I ® ® ® 169 APPENDIX J continued SELF-EVALUATION QUESTIONNAIRE Name ; Date_ DIRECTIONS: A number of statements which people have used to descr ibe themselves are given below. Read each statement and then b lacken i n the appropr ia te c i r c l e to the r i g h t of the statement to i n d i c a t e how you f e l t when you were i n the task  s i t u a t i o n . There are no r i g h t or wrong answers. Do not spend too much time on any one statement but g ive the answer which seems to desc r ibe your f e e l i n g s at that time b e s t . 3 NOT A: SOMEV )DERATEI 1 >< 3 c - > p a > << CO O X s 1. I ® ® ® 2 . I ® ® ® 3 . I ® ® 4. I ® ® ® 5. I ® ®- ® 6. I ® ® ® 7. I ® ® ® 8. I ® ® ® 9 . I ® © ® 10. I ® ® ® X2 170 APPENDIX K THE BETTS QUI VIVIDNESS OF IMAGERY SCALE The aim of t h i s t e s t Is to determine the v i v i d n e s s of your Imagery. The Items of the t e s t w i l l b r i n g c e r t a i n linages to your mind. You are to rate the vividness o f each Image by reference to the accom-panying r a t i n g s c a l e , which i s shown at the bottom of the page. For example, i f your image Is 'vague and dim' you give I t a r a t i n g of 5. Record your answer In the brackets provided a f t e r each Item. Just w r i t e the appropriate number a f t e r each item. Before you turn to the Items on the next page, f a m i l i a r i z e y o u r s e l f with the d i f f e r e n t c a t -egories on the- r a t i n g s c a l e . Throughout the t e s t , r e f e r to the r a t i n g s c a l e when Judging the vividness of each image. A copy of the s c a l e w i l l be printed on the other page. Complete each item before moving to the next. Try to do each item separately Independent of how you may have done the other items. The Image aroused by an Item of t h i s t e s t may be: P e r f e c t l y c l e a r and as v i v i d as the a c t u a l experience Rating 1 Very c l e a r and comparable i n v i v i d n e s s to the a c t u a l experience Sating 2 Moderately c l e a r and v i v i d Rating 3 Not c l e a r or v i v i d , but recognizable Rating 4 Vague and dim Rating 5 So vague and dim as to be hardly dlscernable Rating 6 Mo image present at a l l , you only 'knowing' that you are thinking of the object Rating 7 An example of an item on the t e s t would be one which asked you to consider an image which comes to your mind's eye of a red apple. I f you check your v i s u a l image and i t was moderately c l e a r and v i v i d you would check the r a t i n g s c a l e and mark '3' i n the brackets as fo l l o w s : A red apple (3) 171 APPENDIX K continued Think of some r e l a t i v e or f r i e n d whoa you frequently see, considering c a r e f u l l y the p i c t u r e that r i s e s before your mind's eye. C l a s s i f y the images suggested by each of the following questions as i n d i c a t e d by the degrees of clearness and vi v i d n e s s s p e c i f i e d on the Rating Scale. Item 1. The exact contour of face, head, shoulders and body ( ) 2. C h a r a c t e r i s t i c poses of head, a t t i t u d e s of body, e t c . ( ) 3. The p r e c i s e c a r r i a g e , length of step, etc i n walking ( ) 4. The d i f f e r e n t colors worn i n some f a m i l i a r clothes ( ) Think of seeing the following, considering c a r e f u l l y the p i c t u r e which comes before the mind's eye; and c l a s s i f y the images suggested by the following questions as in d i c a t e d by the degree of clearness and vividness s p e c i f i e d on the Rating Scale. 5. The sun as i t sinks below the horizon ( ) 6. A glass of milk ( ) 7. The door to your house ( ) 8. The l i g h t s of the c i t y at night as seen from above ( ) Rating Scale P e r f e c t l y c l e a r and as v i v i d as the a c t u a l experience .Rating 1 Very c l e a r and comparable v i v i d n e s s to the act u a l experience Rating 2 Moderately c l e a r and v i v i d Rating 3 Not c l e a r or v i v i d , but recognizable Rating 4 Vague and dim Rating 5 So vague and dim as to be hardly d i s c e r n i b l e Rating 6 No image present at a l l , you only 'knowing' that you are thinking of the object Rating 7 172 APPENDIX L RECALL FORM DIRECTIONS: L i s t i n the spaces provided below as many of the statements as you can remember which your p a r t n e r had recorded on the tape about h e r s e l f . Be as accurate as you can, but don't omit statements i f you f e e l the wording isn»t exact. You w i l l have 4 minutes to r e c a l l as many of the statements as you can. C T STOP HERE FOR FURTHER INSTRUCTIONS 1 7 3 APPENDIX L continued PART TWO DIRECTIONS: Now go back to the items which you have remembered on the f i r s t page and, using the s c a l e below, i n d i c a t e f o r each item how c e r t a i n you are that t h i s was one of the statements which your partner made on the tape. VERY SOMEWHAT SOMEWHAT VERY CERTAIN CERTAIN UNCERTAIN UNCERTAIN Place the appropriate number i n the space under column "C f o r each item on page one. GO ON TO THE NEXT PAGE 174 APPENDIX L continued PART THREE DIRECTIONS: Now go back to the f i r s t page again and i n d i c a t e i n column "T" whether you see each statement  to be a P o s i t i v e , Negative, or Neutral statement f o r your  partner to have d i s c l o s e d about h e r s e l f , f o r P o s i t i v e mark (+) (pleasant, d e s i r a b l e ) f o r Negative mark (-) (unpleasant, undesirable) f o r Neutral mark (0) ( n e i t h e r p o s i t i v e or negative, o r e q u a l l y s o ) WHEN YOU HAVE FINISHED, RETURN THIS FORM TO THE EXPERIMENTER. THANK YOU FOR YOUR PARTICIPATION ! 175 APPENDIX M Means and Standard Deviations for the Three Dependent Measures, Total Recall, Total Recognition and Total False Positives for Anxiety Group by Encoding Condition by Retrieval Condition HIGH ANXIOUS GROUP RECALL RECOGNITION FALSE PCS. mean s.d. mean s.d. mean s.d. SE/SR 6.00 2.80 21.75 3.57 2.08 3.50 SE/PR 5.92 2.19 20.83 3.15 2.33 3.55 PE/SR 7.42 2.50 22.17 1.53 .75 .96 PE/PR 10.42 3.48 21.42 3.63 .83 1.75 LOW ANXIOUS GROUP RECALL RECOGNITION FALSE POS. mean s .d. mean s.d. mean s.d. SE/SR 8.50 1.62 22.50 1.51 1.33 2.27 SE/PR 7.67 1.92 21.58 1.31 .75 1.21 PE/SR 7.08 1.97 22.17 1.80 .75 1.06 PE/PR 9.33 2.39 21.83 3.30 1.83 2.82 S=S0CIAL, P=PRIVATE, E=ENCODING, R=RETRIEVAL 176 APPENDIX N Means and Standard Deviations for the Mean Percentage Positive, Negative and Neutral Recall for Anxiety Group by Encoding Condition by Retrieval Condition HIGH ANXIOUS GROUP MEAN % POSITIVE NEGATIVE NEUTRAL mean s.d. mean s.d. mean s.d. SE/SR 25.58 17.88 20.91 20.51 53.83 19.65 SE/PR 27.67 13.35 37.33 24.42 35.83 19.10 PE/SR 32.33 16.74 31.50 15.87 35.83 16.05 PE/PR 30.75 13.75 27.08 12.64 42.00 7.90 LOW ANXIOUS GROUP MEAN % POSITIVE NEGATIVE NEUTRAL mean s.d. mean s.d. mean s.d. SE/SR 27.00 12.70 32.92 11.05 39.92 16.74 SE/PR 23.75 9.96 35.75 10.44 40.25 11.40 PE/SR 34.00 12.92 36.25 16.24 29.41 12.40 PE/PR 31.91 13.78 34.92 9.87 32.83 10.36 S=SOCIAL, P=PRIVATE, E=ENCODING, R=RETRIEVAL 177 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items