UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Anger management in elementary school: descriptive exploration of a group intervention Kalmakoff, Sandra Mary 1994

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

Item Metadata

Download

Media
831-ubc_1994-0464.pdf [ 7.93MB ]
Metadata
JSON: 831-1.0054036.json
JSON-LD: 831-1.0054036-ld.json
RDF/XML (Pretty): 831-1.0054036-rdf.xml
RDF/JSON: 831-1.0054036-rdf.json
Turtle: 831-1.0054036-turtle.txt
N-Triples: 831-1.0054036-rdf-ntriples.txt
Original Record: 831-1.0054036-source.json
Full Text
831-1.0054036-fulltext.txt
Citation
831-1.0054036.ris

Full Text

ANGER MANAGEMENT IN ELEMENTARY SCHOOL: DESCRIPTIVE EXPLORATION OF A GROUP INTERVENTION by SANDRA MARY KALMAKOFF B.A., University of Saskatchewan, 1972 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS 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 June, 1994 (§) Sandra Mary Kalmakoff In presenting this thesis in partial fulfillment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make i t freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It 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 Vancouver, Canada (J ABSTRACT Factors which may be c r i t i c a l to the success of children's anger management programs remain largely uninvestigated. Accordingly an exploratory study was undertaken to describe some of these factors, personal and situational, as observed during the implementation of a school program on anger management. Fifty-eight students (28 males and 30 females) in grades 5, 6, and 7 (ages 10.75 to 14.0 years) participated in a 6-lesson guidance unit on anger management. The unit consisted of a didactic component, covering cognitive-behavioural anger management techniques, and an art component, promoting exploration and expression of anger-related affect through painting. The two components occurred in a different sequence within the program for each half of the sample. Subjects provided data at the beginning, middle and end of the program on anger expression s t y l i s t i c s (Pediatric Anger Expression Scale), state and t r a i t anger and anxiety (Pediatric Anger Scale and Pediatric Anxiety Scale), and knowledge of curriculum content (Anger Management Questionnaire). Subjects also completed the Children's Nowicki-Strickland Internal-External control scale. Independent raters assigned ratings for degree of anger and degree of defensiveness to each of the three paintings completed by each subject within the art component. Data analysis assessed the relationship between scores on the repeated measures and paintings and the following factors: the nature and sequence of program components and a c t i v i t i e s ; and subject age, a b i l i t y level, gender, and locus of control. In the second painting, subjects undergoing the art component f i r s t (art-first) expressed more anger (p < .05) than those doing the didactic component f i r s t (didactic-first), while on the third painting the latter were rated as more defensive (p < .01) than the former. The sample as a whole (p < .01) and both the a r t - f i r s t (p < .01) and didactic-first (p < .05) groups expressed more anger in the second painting than in either of the other two. Gender differences were found in state anger and knowledge of curriculum content at mid-program, and in degree of defensiveness on the f i r s t painting ( a l l p < .05). External locus of control was associated at low levels with t r a i t anxiety at mid-program and at the end, and with state anxiety at the end ( a l l p < .05). Across the sample, t r a i t anxiety was higher pre-program than at either the mid-point or the end (p < .05). State anger and state anxiety were substantially correlated (p < .01) on a l l three measurement occasions, although the range of the former measure was truncated due to floor effects. The results highlight the potential importance of the following factors: content and sequence of program components, nature of the relationship between instructor and students, and levels of state anger and state anxiety. Practical and research implications of the findings are discussed. TABLE OF CONTENTS ABSTRACT i i TABLE OF CONTENTS iv LIST OF TABLES v i i i LIST OF FIGURES X ACKNOWLEDGEMENTS xi DEDICATION x i i CHAPTER I: INTRODUCTION 1 CHAPTER II: LITERATURE REVIEW 4 Anger Management versus Aggression Management 4 The Role of Affect 5 The Role of Conation 6 Accounting for Affect and Conation 8 Art Therapy in Anger Management 8 Anger Management and Individual Differences 9 Gender 10 Level of Maturity 11 Locus of Control 13 Summary of Research Shortcomings 14 Significance of the Study 15 Purpose of the Study 16 Research Questions 17 CHAPTER III: METHODS 19 Subjects 19 The Program 22 Research Design 2 3 Data Collection and Instruments 25 Self-Report Measures 25 Repeated Measures 25 Locus of Control Measure 27 Administration and Scoring of Self-Report Measures 27 Other Measures 28 Measure of Ability 28 Ratings of Subjects* Art Work 29 Analysis of Data 30 Preliminary Analyses 33 Analyses Addressing Research Questions 34 CHAPTER IV: RESULTS 40 Analyses Addressing Research Questions 40 Research Question 1: Nature of Program Component 41 Research Question 2: Order of Components 41 Research Question 3 : Anger Expression in the Art Component 42 Research Question 4: Age 46 Research Question 5: Ability 46 Research Question 6: Gender 47 Research Question 7: Locus of Control 48 Research Question 8: Other 48 Change Scores on Repeated Measures 48 State Anger Frequency Distributions 51 State and Trait Correlations 57 Ancillary Analyses 64 Frequency Distributions and Histograms 64 Correlation Coefficients 64 Test-retest Consistency 64 Interrater Reliability 66 Representativeness of Available Scores on Abil i t y Measure 66 CHAPTER V: DISCUSSION 67 Summary of Results 67 The Major Findings 70 Changes in Anger Expression in the Paintings.... 70 Component Order and Content 76 High Scoring Subjects on State Anger 80 Correlations between State and Trait Variables..86 Gender 89 Change Scores on Repeated Measures 92 Age as a Factor 93 Ability Level 97 Locus of Control 99 Limitations of the Study 101 Research Limitations 102 Program Limitations 110 Summary and Conclusion I l l REFERENCES 117 APPENDICES 123 Appendix A: Letter to Parents of Subjects 123 Appendix B: Lesson Plans for the Anger Management Program 124 Appendix C: Learning Outcomes of the Anger Management Program 159 Appendix D: Congruence of the Anger Management Program with the Intermediate Health and Life Skills Curriculum 160 Appendix E: Reliability and Validity of the Pediatric Anger Expression Scale 162 Appendix F: The State-Trait Anxiety Inventory for Children 166 Appendix G: Anger Management Questionnaire 167 Appendix H: Construction of the Anger Management Questionnaire 169 Appendix I: Reliability and Validity of the Children's Nowicki-Strickland Internal-External Control Scale 170 Appendix J: Administration of the Self-Report Measures 172 Appendix K: Scoring and Interpretation of Questionnaire Data 175 Appendix L: Development of the Rating Scales for Subjects' Paintings 187 Appendix M: Instructions Read to Raters of Subjects' Paintings 188 Appendix N: Paintings of High Scorers on State Anger 191 LIST OF TABLES Table 1: Subject Characteristics by Group 21 Table 2 : Anger Management Program Implementation Schedule 24 Table 3: Variables: Names and Abbreviations 31 Table 4: Differences between Paintings 1, 2, and 3 on Degree of Anger: Total Sample 43 Table 5: Differences between Paintings 1, 2, and 3 on Degree of Anger: Didactic-first Group 44 Table 6: Differences between Paintings 1, 2, and 3 on Degree of Anger: A r t - f i r s t Group 45 Table 7 : Changes in Sample Mean Scores on Repeated Measures 50 Table 8: Chi-square Results for State Anger Subgroups by Scores on Repeated Measures 54 Table 9: Chi-square Results for State Anger Subgroups by Degree of Anger 55 Table 10: Chi-square Results for State Anger Subgroups by Degree of Def ensiveness 56 Table 11: Intercorrelations between State and Trait Anxiety and Anger 58 Table 12 : Intercorrelations between the Anger Expression Stylistics 60 Table 13 : Intercorrelations between State and Trait Anxiety and the Anger Expression S t y l i s t i c s . . . 61 Table 14 : Intercorrelations between State and Trait Anger and the Anger Expression S t y l i s t i c s 62 Table 15: Rel i a b i l i t y Coefficients for Repeated Measures 65 Table 16: Variable Pairs with Increasing Correlations Between Time 1 and Time 3 88 Table 17 : Crosstabulations of Program Group by Degree of Def ensiveness 96 Table B-1: Li s t of Adjectives for Use with Introductory Exercise in Didactic Component - Lesson 1....125 Table J-1: Anomalies in Subject Responses on PAES, PANG, and PANX 179 LIST OF FIGURES Figure 1: Histogram showing frequency distribution of sample scores on state anger at Time 1 52 Figure B-1: Anger Buttons worksheet 126 Figure B-2 : Diagram of Anger Mountain 13 2 Figure B-3: My Self-Talk worksheet 133 Figure B-4: Script for role play from Didactic Component - Lesson 3 141 Figure B-5: First "I" Messages worksheet 142 Figure B-6: Second "I" Messages worksheet 144 Figure B-7: Text for guided visualization in Art Component - Lesson 2 156 ACKNOWLEDGEMENTS I wish to thank the members of my committee for their guidance, their unstinting encouragement and their unshakable faith that I would complete this project. In particular I thank Dr. John Allan for his warm enthusiasm for my work and his pragmatic advice, and Dr. Roy Travis, who helped me unravel the mysteries of st a t i s t i c s while leading me on many absorbing and edifying diversions. I also thank Dr. DuFay Der for his assistance as a member of the examining committee. In addition, I wish to acknowledge the students who served as subjects, and their teachers, without whose participation the study would not have been possible. I am indebted to Joanne Majcher and Nancy Sarrat-Cave for their important contribution as the raters of the paintings. Thanks are also due Georgina Marshall, Karen Reiss, Dan Vie, and Candace Wagman who helped in the development of the ratings scales for the paintings. Finally, I am deeply grateful for the emotional, intellectual and physical support provided by my friends and family, in particular Janet Patterson, Adrianne Ross, Joe Tannenbaum and especially Lezlie Wagman. DEDICATION For Lezlie, and for Jacob CHAPTER I INTRODUCTION Violence and aggression are apparently on the increase both in North American society as a whole and within that microcosm of society, the school. One response to this social c r i s i s has been the development of school-based programs for conflict resolution or violence prevention. Anger management forms an important component of most such programs. In recent years a substantial body of research has appeared on a variety of approaches to teaching anger management. Much of the literature focuses on adult anger management training, but an increasing number of such interventions are being developed for use with children. The most common approach to children's anger management u t i l i z e s cognitive-behavioural techniques to teach social problem-solving (e.g., Kazdin, Esveldt-Dawson, French, & Unis, 1987; Lochman, Curry, Burch, & Lampron, 1984), or self-instructional s k i l l s (Meichenbaum & Goodman, 1971; Camp & Bash, 1981), or a combination (Goldstein & Glick, 1987; Kendall & Braswell, 1982). Such programs have been designed for use with various populations of children, ranging from psychiatric i n -patients (Kazdin et a l . , 1987; Saylor, Benson, & Einhaus, 1985), and hyperactive children (Meichenbaum, 1981), to referred or identified groups of school children (Kendall & Braswell, 1982; Lochman et a l . , 1984), and general school populations (Schneider, 1974). Despite the growth of this f i e l d , taken as a whole, the research on children's anger management reveals a number of limitations. These include confusion between the concepts of anger and aggression, disregard of the affective and conative patterns that underlie anger, and neglect of the role of individual differences in anger management. The current study was designed to begin to address such issues through an exploratory investigation of some of the many factors at play during the implementation of one anger management program for children. The particular program used in the study focused on anger rather than aggression, emphasizing cognitive and affective processes in anger management. The program added an art therapy component to certain of the standard cognitive techniques, as a way of acknowledging and working with some of the affective content underlying children's anger. Chapter II presents a review of the literature on anger management, focussing on approaches used with children in group settings, and examining in particular the areas of limitation in the body of research, as noted above. Arising directly from the literature review are detailed statements of the significance and purpose of the current study and the specific questions i t addressed. Chapter III reports the research methods used in the current study, including descriptions of subject selection and characteristics, program content, data collection procedures, instrumentation and data analysis techniques. Chapter IV presents the results of the data collection and analyses. Chapter V consists of a discussion and interpretation of the results, and includes recommendations for future research. CHAPTER II LITERATURE REVIEW This chapter consists of an overview of the research and descriptive literature on anger management methods and programs, focussing primarily on those used with children in group settings. The literature review occurs within the context of a c r i t i c a l framework consisting of three major sections. The f i r s t section examines the definition of anger management^ looking in particular at the role that affect and conation may play in anger and anger management. The second section considers art therapy as a possible complement to the standard techniques of group anger management training. The third section discusses several individual differences which may affect the outcome of anger management programs. In each of these sections, areas of limitation in the literature are noted. The significance and purpose of the current study, which derive from the foregoing discussion, are presented in the following sections, and the chapter concludes with a statement of the specific research questions addressed by the study. Anger Management versus Aggression Management As A v e r i l l (1983) noted with respect to studies of anger in general, most anger management interventions for children make no distinction between anger and aggression. The goal of such programs is generally to have children control the unacceptable behaviour that i s often associated with feelings of anger, that i s , to control aggression. Usually l i t t l e consideration is given to children's emotional experience of anger, or to helping them learn to understand, accept and manage a l l aspects of their anger, including affect, volition, cognition and behaviour. The Role of Affect Wickless and Kirsch (1988) provided evidence that anger is often associated with feelings of sadness or anxiety. While anger management approaches that emphasize aggression control may indeed produce changes in behaviour, i t i s not known whether they also f a c i l i t a t e change in associated constellations of affect. The results of one study (Saylor et a l . , 1985) of the use of a cognitive-behavioural program with a c l i n i c a l population suggested that sometimes the affect associated with aggression may be so intense as to impede even behavioural change: The subjects " f e l t too angry...to care about using the coping techniques" (p. 13). Some programs do attempt to address aspects of anger in addition to behaviour or behaviour-related cognitions. Fishman-Alschuler and Alschuler (1984) recommended that students be taught to recognize their own internal and external cues that they are becoming angry. Omizo, Hershberger, and Omizo (1988), included sessions on "awareness of feelings" and "reactions to anger" (p. 243) in their program that otherwise focussed on behavioural responses. Elardo and Caldwell's (1979) social development program contained lessons designed to enhance children's awareness of their own and others' thoughts and feelings. The anger management component of the Committee for Children's (1989) Second Step curriculum clearly distinguished between anger and aggression, and encouraged children to become aware of the physical sensations associated with anger arousal, prior to applying stress reduction techniques to them. However, even when a program includes some such "awareness" component, i t usually comprises a small fraction of the total program. In addition, the emphasis on changing behaviour, which often takes a moralistic tone, tends to obscure any intent there may be in the program to validate children's feelings of anger and related emotions, or to help them deal with these feelings in a self-accepting manner. The Role of Conation Cognitive-behavioural approaches to anger management have tended to disregard not only the role of emotion, but also that of conation. Chaplin's (1968) Dictionary of Psychology defines conation as "striving; acting; willing. The conative aspect of the personality i s that which i s characterized by purposive behaviour and the impulse to act" (p. 105) . Stratton and Hayes (1988) note that "the conative domain was one of the three domains of the human psyche outlined by Galen, in the 2nd century BC; the other two being the affective domain and the cognitive domain" (p.37). William James (1890/1981) characterized volition as "primarily a relation...between our Self and our own states of mind" (p. 1172), in which the "essential phenomenon" is "effort of attention" to "the idea of the wise action" (p. 1167). Piaget's theory of intelligence assumed that rational thought or reason is dependent upon both affect and the w i l l (Travis & Cote, 1989). In spite of conation's long history as a construct, vol i t i o n a l influences on human behaviour have largely been ignored in modern psychological research, although studies in areas such as delay of gratification, self-efficacy, personal causation and choice behaviour "hint that vol i t i o n a l factors might be involved in these psychological phenomena" (Howard & Conway, 1986, p. 1243). The significance of conation in anger management is twofold. First, individuals' strivings to change their circumstances or to get what they want often lead them to actions that arouse anger and create situations requiring i t s management. Second, managing anger often necessitates impulse control, delay of gratification, the making of choices, and a w i l l to stay with a situation and see i t through constructively. An individual's patterns of striving and volition, therefore, may be assumed to influence both his or her experience of anger (including i t s intensity, duration and expression) and the strategies available to the individual for coping with i t . Conation may also be assumed to play a role in the learning of anger management techniques, where a great deal of "effort of attention" may be required to change well-established habits of impulsive reaction. Accountincf for Affect and Conation Whether a program is aimed at anger management or aggression management, then, becomes more than a question of definition. The distinction may also reflect two different conceptions of the nature of the psyche. An exclusive focus on the cognitive-behavioural control of aggression belies a uni- or bi-dimensional view of psychological functioning. A more complete, ho l i s t i c perspective is required to account for the affective and conative influences underlying angry behaviour. Art Therapy in Anger Management An approach to working with emotions such as anger that stands in contrast to cognitive-behaviourism i s the use of art therapy. A major difference is that art therapy starts with and tends to focus exclusively on the underlying psychic issues. In Jungian terms, problem-solving, social s k i l l s , and anger management techniques f a l l within the purview of the practical, logical Ego, the centre of consciousness (Allan, 1988); art, on the other hand, comes from the Self - that intuitive aspect of the unconscious which promotes psychological growth and development, and which speaks in the language of pictures and images (Allan & Bertoia, 1992). A child's art work "reflects back to the child that which is coming from deep inside" (Vogli-Phelps, 1985, p. 35). Art work can provide opportunities for the symbolic expression of d i f f i c u l t and dangerous emotions such as anger in the contained safety of the art media (Mills, 1985). Rubin (1984) asserted that "genuinely experiencing the expression of any affect and finding that i t can be controlled and need not be destructive make for a very useful learning experience" (p. 93). Fearful of unleashing an uncontrollable, destructive power, a child may often repress or censor the thoughts, feelings and strivings that underlie his or her anger (for example, thoughts of revenge, feelings of fear, self-hatred or hopelessness, strivings for control). Art therapy offers an avenue for reaching, exploring, and transforming the buried truth of children's anger. The literature on art therapy includes Shennum's (1987) study which found that an art and dance therapy program reduced the levels of acting-out behaviour and emotional unresponsiveness of disturbed children in residential treatment. However, there are no reports of investigations of the use of art therapy activities within programs specifically aimed at anger management. Anger Management and Individual Differences In his review of the prevailing issues in anger management intervention studies, Novaco (1985) noted that "questions concerning the match between client and treatment" (p. 221) have not been adequately addressed. Not much i s known about the role of individual differences in diffe r e n t i a l treatment outcomes of anger management programs. Three individual differences which may well play major roles are gender, level of maturity, and the personality characteristic denoted by the "locus of control" construct (Rotter, 1966). Gender In numerous studies boys have been reported to be more aggressive than g i r l s (Maccoby & Jacklin, 1974a). While studies of children's anger are rare, there i s some evidence suggesting a gender difference in affect as well as behaviour. Franko, Powers, Zuroff and Moskowitz (1985) soli c i t e d descriptions of affect-producing events and found that boys described more events that e l i c i t e d anger than did g i r l s . In addition to such differences in anger-related affect and behaviour, other gender differences may come into play in the learning context provided by a children's anger management program. For example, in pre-adolescence, g i r l s tend to display superior verbal ab i l i t y , while boys have a better grasp of spatial relationships (Maccoby Se Jacklin, 1974a). This trend suggests that g i r l s of this age may respond well to the aspects of an anger management program that have to do with manipulating language, while boys might do better in a different learning modality, such as that provided, for example, by art activities. There is a strong possibility, then, that boys and g i r l s w i l l have different needs when i t comes to learning to manage their anger, and dissimilar responses to any anger management program which is insensitive to these differences. Because the majority of research on children's anger management programs has been conducted using male subjects, to date there is l i t t l e evidence regarding the impact of gender differences on children's experience of anger or their response to such programs. Level of Maturity Another area of individual difference in which there has been very l i t t l e research with respect to anger management is level of maturity, as represented by age. Even the broader f i e l d of developmental differences in the experience and expression of anger is relatively unexplored. One study (Rotenberg, 1985) found that in a sample of children from four different grade levels ( f i r s t , third, f i f t h and seventh) who were asked to describe incidents in which they were angry, a l l ages reported similar causes and intensities of anger. Age differences were found, however, in the motives children ascribed to their anger. Responses to the question "Were you angry in order to (a) get back at someone, (b) try and keep from looking bad, (c) make the other person see my point of view and how I was thinking, (d) make the other person change their (his/her) ways, and (e) just to be mean," (p. 103) showed "an increase with age in the constructive motive of see my point of view" (p. 105). Rotenberg also found age differences in the behavioural manifestations of anger reported by his subjects: Physical aggression decreased with age; and verbal insult, indirect retaliation, and non-retaliation increased with age. Most anger management programs include a major cognitive component. Children at different levels of cognitive development may respond very differently to the cognitive aspects of anger management instruction. Younger children may simply be unable to grasp the concepts, or may have d i f f i c u l t y , for example, learning self-instructional techniques (Meichenbaum & Goodman, 1971), or understanding the connection between a hypothetical example of anger and their own lives. A few studies have attempted to link children's understanding of emotions with stages of cognitive development (e.g., Carroll & Steward, 1984; Cowan, 1982). In a series of studies, Nannis (1988) collected data suggesting that the type of reasoning children use in discussing feelings "roughly parallels" (p. 43) that used when they reason about physical problems. But Nannis cautions that "the development of emotional understanding is not a uniform process" (p. 44) that necessarily occurs in orderly stages. She also notes that "highly charged emotional issues [may] either undermine more advanced cognitive structures or eclipse them" (p. 45). It seems certain that there is a strong relationship between children's developmental level and many aspects of the experience and expression of anger, but the precise nature of that relationship has yet to be f u l l y investigated. Locus of Control A number of "personality" characteristics may be hypothesized to play a role in how the individual child responds to an anger management intervention. One such individual difference is that denoted by the construct of locus of control (Rotter, 1966) . It seems lik e l y that children with a greater sense of internality, believing that they themselves can have an effect upon how they handle their anger, w i l l be more predisposed to learn to do so than those with a greater sense of externality, who may perceive their angry behaviour as being caused by forces external to themselves. There i s some evidence of a relationship between locus of control and affective states. For example, Natale (1978) demonstrated associations between externality and depression, and between internality and elation among undergraduates. Ollendick (1979) found low but significant correlations between high anxiety and external locus of control in fourth-grade children. However, few studies have investigated the relationship between anger and locus of control. Winefield (1981) reported a negative association between readiness to express frustrated anger towards the outer environment and externality, but the correlation was extremely small (r = -.17, E < -05), and the sample consisted entirely of chronically or acutely i l l adults. In a sample of 12- and 13-year-old Nigerian boys, Maqsud (1980) found that teachers rated the externals higher than the internals on antisocial behaviour. Perhaps externally oriented children feel angrier than internals; certainly anger management programs are often aimed at children whose behaviour i s frequently antisocial. The concept of locus of control and that of conation are logically related. It is reasonable to assume that the degree of striving or volition exerted by an individual with internal locus of control w i l l be greater than that of a person with external locus of control, because the former believes that such exertion can make a difference, while the latter does not. Strickland's (1973) finding that externally oriented children chose an immediate reward, while internals chose greater, delayed gratification, lends support to this hypothesis and to the presumed relationship between conation and locus of control. Summary of Research Shortcomings There appear to be gaps in the literature on children's anger management in several areas. Most strikingly, scant attention has been given so far to the role of affect and conation in anger and anger management, and to the role of individual differences in children's differential responses to anger management interventions. Significance of the Study Most studies dealing with anger management programs have attempted to provide evidence of narrowly defined intervention effects. Two considerations directed the current study away from such an evaluative approach, toward a more descriptive exploration of some of the factors that may play a crucial role in the success or failure of children's anger management programs. Fir s t , a survey of the literature on anger and anger management reveals an enormous range of definitions of anger, perceptions of i t s components, and factors assumed to be c r i t i c a l in i t s regulation. As the foregoing discussion has indicated, many of these constructs and variables have not been extensively researched, especially among nonclinical populations of children. These conditions point clearly to the need for further exploration, rather than an evaluation of particular program effects. Second, most of the studies done to date have emphasized one element of the phenomenon of anger, usually the cognitive or the behavioural component. Few have attempted to describe anger in i t s f u l l complexity, as an everyday experience with many facets (an exception i s A v e r i l l , 1983). This complexity, together with the related richness of the processes unfolding in any program aimed at altering the expression of anger, c a l l for the application of exploratory, descriptive research that can begin to address and portray some of that complexity. Accordingly, the current research was not designed to explain the factors that might cause an anger management program to work well (or otherwise), or to predict the success or failure of specific techniques for teaching anger management. Rather, the study was developed to build upon what has already been done in cognitive-behavioural anger management research and add to existing knowledge about children's anger management by providing a description of the interplay between variations in personal qualities (as represented by individual differences in gender, age, and locus of control) and the situational structures of a particular anger management program (i.e., the deliberately manipulated opportunities for anger expression and exploration contained therein). The trends and patterns identified as important in this example of a children's anger management intervention are expected to be useful both as a guide for practitioners and as a source of ideas for further research. Purpose of the Study The specific purpose of the study was to investigate, in a descriptive and exploratory manner, the interplay of personal and situational variability within the context of an elementary school guidance unit on anger management. The anger management program providing this context consisted of two different components: a cognitive-behavioural module, presented in a standard didactic manner (hereafter referred to as the didactic component); and an art therapy module (hereafter referred to as the art component). These two components or situations occurred in a different sequence within the program for each half of the total subject group, which allowed for observation of how the factors of personal variability - level of maturity (as represented by age), gender, and locus of control -interacted with each component and with each situational sequence. Measures taken on several variables related to anger expression and coping strategies furnished the basis for such observation, providing evidence of subject behaviour and thought within the context of the sequence of situations. (These measures, referred to in the research questions l i s t e d below as the anger-related measures, are identified and described in Chapter III.) Research Questions The study addressed the following specific questions: 1. Is there a relationship between subject responses on the anger-related measures and the nature of the program component (art or didactic) most recently experienced? 2. Is there a relationship between the order in which subjects experienced the sequenced situations (the didactic and art components) and subject responses on the anger-related measures? 3. Is there a relationship between subjects' anger expression within the art component and the sequence of art activities? 4. Is there a relationship between subject age and subject responses on the anger-related measures? 5. Is there a relationship between subject a b i l i t y level (as distinct from age) and subject responses on the anger-related measures? 6. Is there a relationship between subject gender and subject responses on the anger-related measures? 7. Is there a relationship between subjects' locus of control and their responses on the anger-related measures? 8. What other regularities in pattern of interaction between situation and personal qualities are discernable in this context? CHAPTER III METHODS This chapter reports the sample characteristics, research design and instruments, data collection procedures and analytic methods used within this study. The f i r s t section provides a description of the subjects who participated in the study, and of the groups they comprised. The second section describes the anger management program that was presented to them. The third section explains the research design, and the fourth section reviews the data collection procedures and the specific instruments u t i l i z e d in the study. Within this section, the instruments are described in two subsections: The f i r s t specifies the self-report measures, including the repeated measures and the locus of control scale, and refers to the procedures used for administering and scoring them; and the second reports the other measures used in the study, including the measure of a b i l i t y and the ratings of subjects' art work. The last section outlines the methods used for s t a t i s t i c a l analysis of the data. Subjects The 58 subjects were students in a middle class, ethnically diverse suburban elementary school. They were drawn from two classes in the school, one a " s p l i t " Grade 5/6 class, and the other a Grade 7 class. Written parental consent for the students' participation in the study was requested for a l l 53 members of these classes (see Appendix A), with consent being received for 62. Results for 3 subjects who were each absent from two or more consecutive lessons of the anger management program were not included in the analysis. In addition, results were not included for one subject who consistently gave the same response to every item on each questionnaire - a case of suspected "malicious compliance" (L. D. Travis, personal communication, 1993). For the purposes of this study the students were divided into four groups, each consisting of about half a class. Both research considerations and issues of teacher concern were brought to bear in the assignment of individual students to groups. The alphabetical class l i s t for the Grade 7 class was divided in half to form the two Grade 7 groups (Groups 1 and 4). One student was moved from one of these groups to the other to separate a pair of siblings. The youngest seven Grade 6 students were assigned to the group with the Grade 5 students to form a youngest group (Group 2). One Grade 5 ESL (English as a Second Language) student who was of Grade 6 age was placed in the Grade 6 group (Group 3). At the beginning of the experimental period, the subjects ranged in age from 10 years, 9 months to 14 years, 0 months. The f i n a l sample included 30 females and 28 males. The grades, number of males and females, number of subjects, and mean age for each group are shown in Table 1. Subnect Characteristics by Group Group 1 2 3 4 Grade(s) 7 5 (n = 7) 6 (n = 7) 5 (n = 1) 6 (n = 14) 7 Gender composition F = 7 M = 8 F = 10 M = 4 F = 8 M = 7 F = 5 M = 9 n 15 14 15 14 Mean age (months) 154.3 134.9 140.8 155.1 Note. F = female; M = maie. The Program Each group met separately with the researcher/ experimenter for 40 minutes once a week during the 6-week intervention period to participate in a six-lesson guidance unit on anger management (hereafter referred to as the anger management program). The program consisted of two parts, a cognitive-behavioural component employing didactic instructional techniques (the didactic component), and an art therapy component (the art component) which promoted student self-expression and self-exploration through painting. Each component contained three of the six anger management lessons. The anger management program covered the following topics: the nature of anger, distinguishing between angry feelings and angry behaviour, physiological indicators of anger, self-instructional and other coping strategies, and assertive statements. In the didactic component of the program these topics were addressed through direct teaching, pencil-and-paper tasks, discussion, and role play. In the art component they were explored through painting a c t i v i t i e s , and through individual and group discussion of the art work produced. (See Appendix B for detailed lesson plans.) The anger management program was developed by the researcher specifically for the purposes of this study. Program content was based primarily upon the anger management lessons found in Peer Conflict Resolution through Creative Necfotiation (Kalmakoff & Shaw, 1987) . Other sources included Crary (1984), Kriedler (1984), Palormares (1975), Schmidt and Friedman (1983), and United Nations Education (1986), The anger management program was deemed by school personnel to be congruent in i t s objectives for student learning (see Appendix C) with those of the d i s t r i c t -mandated Health and Life Skills Curriculum (see Appendix D). The program was introduced and taught to the students as a component of this curriculum. Research Design The study used a repeated measures design, with a l l subjects participating in both conditions - both the didactic and the art component - according to the schedule shown in Table 2. Anger Manacrement Program Implementation Schedule Group Grade 5/6 Grade 6 Grade 7 Grade 7 Lesson (n=14) (n=15) (n=15) (n=14) 1 D A D A 2 D A D A 3 D A D A 4 A D A D 5 A D A D 6 A D A D Note. D = Didactic component; A = Art component. Data Collection and Instruments Self-Report Measures Repeated Measures Prior to the beginning of the program (pretest. Time 1), after completion of the f i r s t component, that i s , between Lessons 3 and 4 (intermediate test. Time 2), and at the end of the program (posttest. Time 3), measures were taken on the following variables, seen to be salient within the context of an anger management program: anger expression, state anger, t r a i t anger, state anxiety, t r a i t anxiety, and knowledge of anger coping strategies. Anger expression. This variable was measured using the Pediatric Anger Expression Scale (PAES) (Jacobs, Phelps, & Rohrs, 1989), a self-report t r a i t measure which assesses anger expression s t y l i s t i c s . The study used a 15-item, three-factor variant of this scale (G. A. Jacobs, personal communication, January 15, 1990). The three factors (anger expression s t y l i s t i c s ) are: anger-out (open expression of anger in uncontrolled or aggressive behaviour), anger-suppression (holding in or denial of anger), and anger-reflection/control (control of angry reactions and direct attempts to resolve the conflict). The factor subscales consist of five items each (e.g., "I show my anger" on the anger-out subscale; "I hold my anger in" on the anger-suppression subscale; and "I control my temper" on the anger-reflection/control subscale), (See Appendix E for r e l i a b i l i t y and validity information on this scale.) State and t r a i t anger, and state and t r a i t anxiety. These variables were measured using the Pediatric Anger Scale (PANG) and the Pediatric Anxiety Scale (PANX) (Jacobs, cited in Jacobs et a l . , 1989). These scales are self-report inventories, 19 and 2 0 items respectively, which measure anger and anxiety as emotional states and personality t r a i t s . The state subscales require children to assess the intensity of emotion experienced at the moment of responding to each item (e.g., "I am mad" for state anger; "I feel worried" for state anxiety). The t r a i t subscales require children to report the frequency with which they have experienced the personality characteristic specified in each item (e.g., "I have a bad temper" for t r a i t anger; "I worry too much" for t r a i t anxiety). Jacobs et a l . (1989) noted that these scales are "similar in format and conception to the State-Trait Anxiety Inventory for Children (Spielberger, Edwards, Lushene, Montuori and Platzek, 1973)" (p. 60). (See Appendix F.) Knowledge of anger coping strategies. This variable was measured using the Anger Management Questionnaire (AMQ) (see Appendix G), a scale constructed by the researcher to assess subjects' knowledge of the content of the didactic component of the program. This measure is a 20-item true/false questionnaire. (See Appendix H for details on the construction of this questionnaire.) Locus of Control Measure The Children's Nowicki-Strickland Internal-External Control Scale (CNSIE) (Nowicki & Strickland, 1973) was administered with the pretests. This scale is a paper-and-pencil measure consisting of 40 questions which subjects answer by marking either the yes or the no placed next to the question (e.g., "Are some kids just born lucky?"; "Do you believe that i f somebody studies hard enough he or she can pass any subject?"). "The items describe reinforcement situations across interpersonal and motivational areas such as a f f i l i a t i o n , achievement and dependency" (Nowicki & Strickland, 1973, p. 149). (Appendix I reports r e l i a b i l i t y and v a l i d i t y data for this scale.) Administration and Scoring of Self-Report Measures The repeated measures and the CNSIE were administered by the subjects' classroom teachers according to the protocol described in Appendix J. Due to student absences, for a few of the subjects the measures were taken at somewhat different times than for the rest of the sample. See Appendix J for details. The procedures used for scoring the PAES, the PANG, the PANX, and the CNSIE are specified in Appendix K. This appendix also reports anomalies found in the raw data results from these self-report measures, and describes how these anomalies were interpreted for scoring purposes. Chapter V includes a discussion of the anomalous data. Other Measures Measure of Abil i t y Considering chronological age as a factor in anger management necessitated also including a b i l i t y as a factor, in order to be able to determine whether any age effects resulted from a b i l i t y differentials or from some other age-related factor. To provide a measure of ability, subjects' scores on the Test of Cognitive Ski l l s (TCS) (CTB/McGraw-Hill, 1981) were obtained from their teachers. The TCS i s a series of ab i l i t y tests designed to assess students' level of academic aptitude. It includes a Sequences test that measures the student's a b i l i t y to see principles implicit in patterns or sequences, an Analogies test measuring a b i l i t y to see concrete and abstract relationships, a Memory test, and a Verbal Reasoning test that measures the a b i l i t y to reason logically. Percentile rank scores for each test and weighted composite percentile ranks were supplied to the researcher. Only the composite scores were used in the current analysis. The TCS was used for the purposes of this study because i t was the most recently administered test of i t s kind. The Grade 6 and 7 subjects had been tested a year and a half prior to the experimental period, and the Grade 5 subjects had been tested two and a half years before. A f u l l data batch for this measure was not acquired due to administrative sensitivities. TCS scores for only 39 of the 58 subjects were obtained. Ratincfs of Subjects' Art Work In the art component of the anger management program, subjects produced one painting each during each of the three lessons. Six subjects were absent from the f i r s t painting lesson (1 or 2 per group), and 6 from the third lesson. Prior to administration of the posttest questionnaires, the experimenter retaught the third lesson to 3 of the latter subjects ( a l l from Group 2). There was thus a total of 9 absences from lessons in the art component, and a total of 165 paintings produced. Photographic slides were taken of each painting. Two independent raters viewed the slides and rated each painting for degree of anger and degree of defensiveness• The raters viewed the entire set of slides twice, the f i r s t time to answer the question "Does the depictor give evidence of an angry state; i f so to what degree?" by rating each painting for degree of anger on a 3-point ordinal scale (none. moderate. extreme); the second time, the raters answered the question "Does the depictor give evidence of defensiveness. that i s , an attempt to disguise an angry state or inhibit i t s expression?" by rating each painting on the same 3-point scale. (Appendix L describes the development of these scales and the raters' qualifications.) Subjects were asked to provide a t i t l e for each painting, and in many cases, the t i t l e was displayed in the painting as an integral part of the image. There i s some evidence to suggest that the words that accompany such art work can increase the accuracy of prediction of the artist's emotional state (J. Allan, personal communication, 1993). For these reasons, the raters were asked to include the t i t l e in their consideration of each painting, and the t i t l e s were read aloud to the raters as the slides were shown. Prior to viewing the slides, the raters were shown a preliminary set of pictures containing two or three examples of paintings from each rating category. The raters viewed the entire set of slides in the same - random - order both times. The raters were blind to the specific nature of the study (except insofar as i t dealt with anger in children), including the conditions under which the paintings were produced. (See Appendix M for details of the instructions given to the raters.) Analysis of Data This section describes, f i r s t , the set of preliminary s t a t i s t i c a l analyses applied to the data collected, and second, the specific analyses conducted in order to address each of the research questions. Table 3 l i s t s the names and abbreviations of a l l the variables used in these analyses. Variables: Names and Abbreviations Variable Score on locus of control measure (CNSIE) lESCORE Score on a b i l i t y measure (TCS) TCS State anxiety, Time 1 (pretest) STANXl State anxiety. Time 2 (intermediate test) STANX2 State anxiety. Time 3 (posttest) STANX3 State anger. Time 1 STANGl State anger. Time 2 STANG2 State anger. Time 3 STANG3 Trait anxiety. Time 1 TRANXl Trait anxiety. Time 2 TRANX2 Trait anxiety. Time 3 TRANX3 Trait anger. Time 1 TRANGl Trait anger. Time 2 TRANG2 Trait anger. Time 3 TRANG3 Anger--reflection/control. Time 1 ANGRXCl Anger--reflection/control. Time 2 ANGRXC2 Anger--reflection/control. Time 3 ANGRXC3 Anger--out. Time 1 ANGOUTl Anger--out. Time 2 ANG0UT2 Anger--out. Time 3 ANG0UT3 (table continues) Abbreviation Table 3 (continued) Variable Abbreviation Anger-suppression, Time 1 ANGSUPPl Anger-suppression, Time 2 ANGSUPP2 Anger-suppression, Time 3 ANGSUPP3 Anger Management Questionnaire, Time 1 AMQl Anger Management Questionnaire, Time 2 AMQ2 Anger Management Questionnaire, Time 3 AMQ 3 Degree of anger. Painting 1 ANGPICl Degree of anger. Painting 2 ANGPIC2 Degree of anger. Painting 3 ANGPIC3 Degree of defensiveness. Painting 1 DEFPICl Degree of defensiveness. Painting 2 DEFPIC2 Degree of defensiveness. Painting 3 DEFPIC3 Preliminary analyses Frequency distributions and histograms. These were obtained for each variable, f i r s t across the total sample, and then by program group. Correlation coefficients. Pearson product-moment correlations were calculated for each variable with each of the others. The resulting matrix of correlation coefficients was intended to indicate areas of possible relationship among the variables. Significance levels were calculated using the two-tailed test. Test-retest consistency. A r e l i a b i l i t y analysis was performed on each of the repeated measures scales (STANX, STANG, TRANX, TRANG, ANGRXC, ANGOUT, ANGSUPP, and AMQ) using Cronbach's Alpha. Interrater r e l i a b i l i t y . The consistency between the ratings given to the paintings by the two independent raters was measured using Pearson product-moment correlations. Three correlation coefficients were obtained: one for the ratings on degree of anger, another for the ratings on degree of defensiveness^ and a third overall measure of interrater consistency using both sets of ratings (that i s , the ratings for anger and defensiveness taken together). Analysis of representativeness of available scores on ab i l i t y measure. As a means of determining whether the available scores on the Test of Cognitive S k i l l s (TCS) were representative of the entire sample, the sample was divided into two groups - subjects having TCS scores, and those without - and t tests were conducted to assess significant differences between these two groups on a l l variables. Significance levels were calculated using the two-tailed test. Analyses Addressing Research Questions The research questions appear below as subsection heads. In each case, anger-related measures refers to (a) the repeated measures: the Pediatric Anger Expression Scale, the Pediatric Anger Scale, the Pediatric Anxiety Scale, and the Anger Management Questionnaire; and (b) the ratings assigned to the paintings for degree of anger and degree of defensiveness. Following each heading the data analyses pertaining to that question are reported. Research Question 1; Is there a relationship between subject responses on the anger-related measures and the nature of the program component (art or didactic) most recently experienced? At Time 2, half the subjects had completed only the didactic component of the anger management program, while the other half had completed only the art component. Consequently, in order to investigate the possible differential impact of the two different program components, chi-square analysis was used to compare Time 2 results from the repeated measures for subjects undergoing the didactic component of the program f i r s t (Groups 1 and 2, hereafter referred to as didactic-first) with those for subjects who experienced the art component f i r s t (Groups 3 and 4, a r t - f i r s t ) . At Time 3 each of the program-order groups (i.e., d i d a c t i c - f i r s t and art-first) had completed the component that came second for that group (art and didactic respectively), allowing for a within-group comparison of the data produced immediately following each component. Thus each of the program-order group•s scores on the repeated measures were used to conduct t tests for significant differences between the group mean at Time 2 and that at Time 3. Because the a r t - f i r s t group was exposed to the didactic component between Time 2 and Time 3, i t was expected that their Time 3 scores on the Anger Management Questionnaire would be higher than their Time 2 scores, so for this variable the one-tailed test of significance was used. For a l l others, the two-tailed test was used. (Since the paintings were produced during one of the program components, not after, i t was not appropriate to include the painting ratings in the analyses addressing this question.) Research Question 2 : Is there a relationship between the order in which subjects experienced the sequenced situations (the didactic and art components) and subject responses on the anger-related measures? It was conjectured that because the two program components constituted two disparate approaches to exploring anger, undergoing one or the other component f i r s t might predispose students to a different experience of the second. Because half the sample went through the didactic component f i r s t and then the art component, and the other half experienced the components in the reverse order, i t was possible to investigate whether order of exposure to the two different program components had any bearing on the observed results. Chi-square analysis was used to compare repeated measures results for the didactic-first group with those for the a r t - f i r s t group at Time 1, Time 2, and Time 3. The groups' ratings of the paintings for anger and defensiveness were also compared using chi-square analysis. Research Question 3 ; Is there a relationship between subjects' anger expression within the art component and the sequence of art activities? Anger expression within the art component was assessed through the ratings assigned to subjects' paintings for degree of anger and degree of defensiveness. T tests were conducted to determine i f there were significant differences between the ratings given to the f i r s t , second and third paintings. The i n i t i a l analysis included the ratings for the entire sample. Subsequently the sample was divided by program-order group and the same tests were performed on each group separately. It was predicted that the degree of anger expressed in the paintings would increase over the course of the art component, and that the degree of defensiveness would decrease. Thus the one-tailed test of significance was used. Research Question 4 ; Is there a relationship between subject age and subject responses on the anger-related measures? The age-graded formation of the four program groups (Groups 1, 2, 3, and 4) allowed for this question to be approached through a series of crosstabulations of group number by repeated measures results, with chi-square analysis. Chi-square analysis was also used to assess the distributions, across the groups, of the ratings assigned to the paintings for degree of anger and degree of defensiveness. Research Question 5: Is there a relationship between subject a b i l i t y level and subject responses on the anger-related measures? The available scores on the measure of a b i l i t y (TCS) were crosstabulated with a l l results from the repeated measures and with the painting ratings, and subjected to chi-square analysis. In addition, the distribution of the available TCS scores of the didactic-f i r s t group were compared with that of the available scores of the a r t - f i r s t group to determine whether any differences between the program-order groups could be attributed to a b i l i t y . Research Question 6: Is there a relationship between subject gender and subject responses on the anger-related measures? T tests were conducted to assess significant differences between group means for males and females on a l l the repeated measures. Significance levels were calculated using the two-tailed test. The relationship between gender and scores on the repeated measures was also investigated using chi-square analysis. In addition, chi-square analysis was u t i l i z e d to assess any relationship between gender and the ratings of the paintings for anger and defensiveness. Research Question 7; Is there a relationship between subjects' locus of control and their responses on the anger-related measures? Only the subjects with the eight most extreme locus of control scores were included in this exploratory analysis, those with the four most external scores being compared with those with the four most internal scores. Given that most of the locus of control scores clustered close to the median, this approach was taken in order to highlight possible relationships between locus of control and the other measures. These small lESCORE subgroups were crosstabulated with the scores on a l l repeated measures and with the paintings ratings, and the results were analyzed using chi-square tests. Research Question 8 ; What other regularities in pattern of interaction between situation and personal qualities are discernable in this context? Three different analyses were performed to address this question. In the f i r s t , sample change scores on the repeated measures were examined: T tests were conducted to assess significant differences between the sample mean score on each measure at Time 1, Time 2, and Time 3. It was predicted that the sample mean score on the Anger Management Questionnaire would increase between Time 1 and Time 3, so the one-tailed test of significance was used for this measure. Otherwise the two-tailed test was used. Because the other analyses addressing this question arose out of the results of the previous analyses, the former are reported in Chapter IV: Results. CHAPTER IV RESULTS This chapter presents, f i r s t , the results of the analyses undertaken to answer the research questions, which are l i s t e d at the conclusion of Chapter II and again in the last section of Chapter III; and second, the results of the preliminary analyses that were applied to the data collected and which furnished results of an ancillary nature. In the f i r s t section of this chapter, the eight research questions are reiterated and each is followed in turn by a description of the results of the pertinent analyses. In the second section, each ancillary analysis is noted, and i t s results are described. Analyses Addressing Research Questions The research questions appear below as subsection heads. In each case, anger-related measures refers to (a) the repeated measures: the Pediatric Anger Expression Scale, the Pediatric Anger Scale, the Pediatric Anxiety Scale, and the Anger Management Questionnaire; and (b) the ratings assigned to the paintings for degree of anger and degree of defensiveness. Following each heading the results of the analyses pertaining to that question are reported. Where appropriate, selected results of ancillary analyses, notably correlation coefficients, are also reported. Research Question 1; Is there a relationship between subject responses on the anger-related measures and the nature of the program component (art or didactic) most recently experienced? The chi-square analysis comparing the distributions of Time 2 results for subjects in the did a c t i c - f i r s t program-order group with those for subjects in the a r t - f i r s t group yielded only one significant result: The d i d a c t i c - f i r s t group tended to score higher on the Anger Management Questionnaire than the a r t - f i r s t group {X^i2, N = 58] = 6.04, E < .05). The t test comparisons of Time 2 and Time 3 program-order group means showed no significant differences between subject responses on the repeated measures following the didactic component and those following the art component for either program-order group, except for the predicted rise in the a r t - f i r s t group's mean score on the Anger Management Questionnaire (Time 2 M = 13.76, Time 3 M = 14.55, t(28) = -1.96, E < «05, one-tailed). Research Question 2 : Is there a relationship between the order in which subjects experienced the sequenced situations (the didactic and art components) and subject responses on the anger-related measures? In the comparison of the repeated measures results for the two program-order groups, the only variable for which a significant chi-square was found was the Anger Management Questionnaire at Time 2 (reported above). For the comparison of the painting ratings of the program-order groups (and for a l l subsequent analyses involving the painting ratings), the two sets of ratings were aggregated (i.e., for each painting, the values assigned by both raters were used), effectively doubling the total sample size for the analysis. This procedure was deemed appropriate given the substantial to high correlation obtained between the two sets of ratings, reported below under "Ancillary Analyses." The analysis comparing the paintings ratings of the program-order groups showed that in the second painting, a greater degree of anger was expressed by the a r t - f i r s t group than by the didactic-first group (X^l^^ H = 116] = 8.55, p < .05). In the third painting, a greater degree of defensiveness was evident in the didactic-first group than in the a r t - f i r s t group (X^[2, N = 110] = 9.33, p < .01). Research Question 3; Is there a relationship between subjects' anger expression within the art component and the sequence of art activities? As the results displayed in Table 4 indicate, the ratings for degree of anger in the second painting were significantly higher than those in either the f i r s t or the third painting. Furthermore, the f i r s t and the third paintings did not di f f e r on degree of anger. These findings were consistent across the entire sample and within each of the program-order groups (see Tables 5 and 6). Differences Between Paintings 1. 2, and 3 on Degree of Anger: Total Sample fN = 58) Mean Painting No. of anger 104 98 Probability level no. cases^ rating t value df (one-tailed) 2.06^ 2.40 2 . 08' 2 . 06 0.23 97 -3.87 103 E < -01 ns 110 2.42 2.04 4.62 109 E < -01 Note, ns = not significant. ^Includes two sets of ratings per subject. Numbers vary due to subject absences from lessons. ^Means for each painting di f f e r between comparisons due to procedures used by s t a t i s t i c a l analysis program for handling missing data (e.g., for subjects absent for Painting 3, the ratings assigned for Painting 1 were included in the calculation of the Painting 1 mean for the comparison of Paintings 1 and 2, but not in that for the comparison of Paintings 1 and 3). Differences Between Paintings 1, 2. and 3 on Degree of Anger; Didactic-first Group (JQ = 29) Mean Probability Painting No. of anger level no. cases^ rating t value df (one-tailed) 1 2.02^ ^ 52 -1.81 51 E < .05 2 2.25 1 2.07^ » 46 0.20 45 ns 3 2.04 2 2.31 52 2.94 51 p < .01 3 2.00 Note, ns = not significant. ^Includes two sets of ratings per subject. Numbers vary due to subject absences from lessons. ^Means for each painting di f f e r between comparisons due to procedures used by s t a t i s t i c a l analysis program for handling missing data (e.g., for subjects absent for Painting 3, the ratings assigned for Painting 1 were included in the calculation of the Painting 1 mean for the comparison of Paintings 1 and 2, but not in that for the comparison of Paintings 1 and 3). Differences Between Paintings 1, 2, and 3 on Degree of Anger: A r t - f i r s t Group (n = 29) Mean Probability Painting No. of anger level no. cases^ rating t value df (one-tailed) 1 2.10 52 -3.72 51 E < -01 2 2.56^ » 1 2 .10 52 0.14 51 ns 3 2.08 2 2.52^ ^ 58 3.56 57 E < -01 3 2.07 Note, ns = not significant. ^Includes two sets of ratings per subject. Numbers vary due to subject absences from lessons. '^ Means for each painting may differ between comparisons due to procedures used by s t a t i s t i c a l analysis program for handling missing data (e.g., for subjects absent for Painting 3, the ratings assigned for Painting 2 were included in the calculation of the Painting 2 mean for the comparison of Paintings 1 and 2, but not in that for the comparison of Paintings 2 and 3). No significant results were found for degree of defensiveness. Research Question 4 : Is there a relationship between subject age and subject responses on the anger-related measures? A modest correlation was noted between age and state anger at Time 1 (r = .27, p < .05). No significant chi-square result was found in the analysis of the distribution of scores on the repeated measures across the program groups (Groups 1, 2, 3, and 4). The analysis of the distribution of paintings ratings across the groups yielded a significant result for def ensiveness in the second painting (X^t^' ^ ~ 24.24, p < .01), and in the third painting N = 110] = 15.50, p < .05). No direct correlation was found, however, between age and the ratings of paintings. Research Question 5: Is there a relationship between subject a b i l i t y level and subject responses on the anger-related measures? Moderate correlations were found between scores on the Test of Cognitive S k i l l s (TCS) and state anger at Time 1 (r = -.36, p < .05), Time 2 Anger Management Questionnaire scores (r = .32, p < .05), and degree of defensiveness in Painting 1 (r = -.26, p < .05). However, there were no significant results when chi-square analysis was used to compare the available TCS scores with any of the repeated measures scores, nor when the former were compared with the painting ratings. Similarly, no significant difference was found between the distributions of available TCS scores of the two different program-order groups. Research Question 6; Is there a relationship between subject gender and subject responses on the anger-related measures? Gender was found to correlate at a moderate level with state anger at Time 2 (r = .25, p < .05), the boys giving more evidence of state anger than the g i r l s . Time 2 Anger Management Questionnaire results also correlated significantly with gender (r = -.31, E < -05), with the g i r l s scoring higher than the boys. In the f i r s t painting, the boys were rated as somewhat more defensive than the g i r l s (r = .26, p < .01). Significant differences between the group means for males and females were found in the scores for state anger at Time 2 and for the Anger Management Questionnaire at Time 2. On state anger, the boys scored higher (M = 14.79) than the g i r l s (M = 11.83), t(56) = 2.03, p < .05 (two-tailed), while on the Anger Management Questionnaire, the g i r l s scored higher (M = 15.87) than the boys (M = 13.86), t(56) = 2.44, p < .05 (two-tailed). The chi-square analysis comparing gender with scores on the repeated measures supported only the result for state anger. Time 2, above, showing the boys' scores to be distributed more towards the high end than those of the One gender difference was found in the chi-square analysis of the painting ratings. On the f i r s t painting, the boys were rated as more defensive than the g i r l s : 2(2, N = 104) = 7.32, p < .05. Research Question 7: Is there a relationship between subjects' locus of control and their responses on the anger-related measures? Moderate correlations were found between scores on the locus of control measure and state anxiety at Time 3 (r = .26, E < .05), t r a i t anxiety at Time 2 (r = .28, E < .05), and t r a i t anxiety at Time 3 (r = .31, E < -05), indicating a positive intercorrelation between externality and each of these variables. On anger-suppression at Time 1, the four subjects with the most external locus of control scores tended to score higher than those with the most internal scores (^^[2, N = 8] = 8.00, E < -05)• In addition, on the third painting, the internals were rated as having expressed more anger than the externals: ^2(2, N = 12) = 6.67, p < -05. Research Question 8: What other regularities in pattern of interaction between situation and personal qualities are discernable in this context? Sample change scores on repeated measures. The t tests for significant differences between the sample mean score on g i r l s : E < .05. a l l measures at Time 1, Time 2 and Time 3 yielded significant results on two variables, t r a i t anxiety and the Anger Management Questionnaire. Table 7 shows the sample mean scores, t value, degrees of freedom and probability levels for each pair of variables. Trait anxiety decreased between Time 1 and Time 2, and remained significantly less at Time 3 than i t had been at Time 1. Trait anxiety did not dif f e r significantly between Time 2 and Time 3. As the table also shows, the sample mean score on the Anger Management Questionnaire was higher at both Time 2 and Time 3 than at Time 1. Changes in Sample Mean Scores on Repeated Measures Variable Mean t value df TRANXl 18.00^ TRANX2 16.99 TRANXl 17.91' TRANX3 16.96 AMQl 14.19 AMQ2 14.90 2.58# 56 2.08# 56 -1.75* 57 AMQl 14.19 AMQ3 15.29 -2.83** 57 Note. See Table 3 for explanation of variable abbreviations. ^Means di f f e r due to procedures used by s t a t i s t i c a l analysis program for handling missing data (i.e., for subjects absent at Time 3, Time 1 scores were included in the calculation of the TRANXl mean for the comparison of TRANXl and TRANX2, but not in that for the comparison of TRANXl and TRANX3). #P < .05, two-tailed. *p < .05, one-tailed. **p < .01, one-tailed. Anomaly in the state anger frequency distributions. The frequency distributions and histograms for state anger for the entire sample at Time 1, Time 2, and Time 3 displayed a marked tendency for scores to be distributed at the extreme low end of the scale (see Figure 1). In addition, a small number of scores appeared consistently at the extreme high end of the scale. The individual frequency distributions for each of the four program groups also manifested this "floor effect," with two exceptions: The frequencies for Group 2 at Time 3 and for Group 3 at Time 1 were more normally distributed, albeit within a smaller range in the bottom two-thirds of the scale. An exploratory analysis was undertaken to assess in what other ways, i f any, the small group of high scorers on state anger stood apart from the rest of the sample. For the purposes of this analysis, high scorers were defined as a l l subjects who had scored 23 or more (in a possible range of 9 to 27) on at least one state anger inventory (at either Time 1, Time 2, or Time 3). The defining score was set at 23 because the histograms showed the high scores to cluster above that point (see Figure 1). Ten subjects met the criterion. Of these, 3 subjects scored 23 or more at a l l three times. A majority, 7 subjects, were in the a r t - f i r s t program-order group, while the other 3 took the didactic component f i r s t . Figure 1. Histogram showing frequency distribution of sample scores on state anger at Time 1. MT VALUE 19 9 .00 6 10.00 3 1 1 .00 4 12.00 3 13.00 4 14 .00 2 15.00 3 16.00 3 17.00 2 18.00 0 19 .00 1 20.00 1 21 .00 0 22.00 1 23.00 0 24.00 3 25.00 0 26.00 2 27.00 ONE SYMBOL EQUALS APPROXIMATELY .40 OCCURRENCES I I I I I I 0 4 8 12 16 20 HISTOGRAM FREQUENCY The high scorers formed one subgroup on state anger, and the rest of the sample formed the other. The two subgroups were then crosstabulated with the scores each of the repeated measures, and with the ratings of the paintings. Table 8 shows the results of the repeated measures analyses. High scorers on state anger were shown to exhibit high state anxiety at Time 1, Time 2, and Time 3. In addition, high scorers tended to score higher than the rest of the sample on t r a i t anger at Time 3, and on anger-out at Time 2 and Time 3. They also scored lower on the Anger Management Questionnaire at Time 2 and at Time 3. Tables 9 and 10 show the results of the analysis of the painting ratings. On the f i r s t painting, more high scorers than others were rated as expressing no anger, at the same time that a larger percentage of high scorers than others was judged to have expressed extreme anger. Similarly on the f i r s t painting, a larger percentage of high scorers was given a rating of no defensiveness, while more high scorers than others were rated as extremely defensive. On the third painting, high scorers were rated, overall, as expressing more anger. It should be noted that the number of subjects for whom results are available for Paintings 1 and 3 i s 9, not 10 (one student being absent from each of these lessons). Chi-square Results for State Anger Subgroups by Scores on Repeated Measures Probability Variable N Chi-square^ level STANXl 57 22.17 E < .01 STANX2 58 9.81 E < .01 STANX3 58 21.68 E < .01 TRANG3 57 7.21 E < .05 ANG0UT2 58 11.04 E < .01 ANG0UT3 58 9.50 E < .01 AMQ2 58 13.51 E < .01 AMQ3 58 13.36 E < .01 Note. See Table 3 for explanation of variable abbreviations, ^df = 2 Chi-square Results for State Anger Subgroups by Degree of Anger Degree of anger (%)^ Subgroup None Moderate Extreme High scorers on STANG^ 33 A l l others«^ 17 Painting 1 28 59 39 23 High scorers on STANG® 6 A l l othersf 20 Painting 3 33 67 61 13 '2 b 6.01 < .05 21.17 < .01 ^% of n for each subgroup, rounded to nearest whole number, bdf = 2. ^n = 18 (painting ratings). ^n = 86 (painting ratings). ®n = 18 (painting ratings). ^n = 92 (painting ratings). Chi-square Results for State Anger Subgroups by Decfree of Defensiveness Degree of Defensiveness (%)^ Subgroup None Moderate Extreme /H^ ^ E Painting 1 High scorers on STANGC 39 11 50 A l l others<^ 29 48 23 9.10 < .05 ^% of n for each subgroup, rounded to nearest whole number. ^df = 2. *^ n = 18 (painting ratings) . ^n = 86 (painting ratings) . Correlations between state and t r a i t v ariables. Patterns were noted i n the i n t e r c o r r e l a t i o n s of scores on the various state and t r a i t measures. Table 11 shows the degree of c o r r e l a t i o n found between state anxiety, state anger, t r a i t anxiety and t r a i t anger. Substantial correlations were revealed between state anxiety and state anger at a l l times, except for state anxiety at Time 3 with state anger at Time 1 and Time 2. The strongest correlations occurred at the matching times for these two variables ( i . e . , state anger at Time 1 with state anxiety at Time 1, et c . ) . For the most part, state anxiety and t r a i t anxiety were i n t e r c o r r e l a t e d at a moderate l e v e l . The correlations between state anxiety and t r a i t anger ranged from nonsignificant to r = .49 (p < .01) at Time 3 f o r both variables. State anger and t r a i t anxiety were not correlated, i n most instances. Except at Time 1, state anger and t r a i t anger showed a marked c o r r e l a t i o n . T r a i t anxiety and t r a i t anger were also s u b s t a n t i a l l y correlated, except at Time 1 for t r a i t anger where the co r r e l a t i o n s were somewhat lower. Intercorrelations between State and Trait Anxiety and Anger STANG TRANX TRANG Scale 1 2 3 1 2 3 1 2 3 STANX 1 .66 .44 .50 .42 .36 .45 ns .36 .49 2 .53 .61 .51 .35 .34 .36 .28* .40 .41 3 .39 .38 .60 .28* ns .41 .31* .44 .49 STANG 1 - a a .33* ns ns .28* .44 .51 2 - a ns ns ns .44 . 55 .48 3 - .32* ns .37 .45 . 62 .69 TRANX 1 - a a .41 .48 .47 2 - a .38 .48 .50 3 _ .33* .46 .60 Note. See Table 3 for explanation of variable abbreviations. E < .01 unless otherwise indicated, ns = not significant. ^Intercorrelations between scores on the same measure taken at different times are not reported here, since the r e l i a b i l i t y coefficient for each repeated measure i s reported elsewhere. *E < .05. Table 12 shows the intercorrelations between the anger expression s t y l i s t i c s represented by the three factor subscales of the Pediatric Anger Expression Scale (PAES) (Jacobs et a l . , 1989): anger-reflection/control, anger-out and anger-suppression. There were substantial to high negative correlations between anger-reflection/control and anger-out. There were also substantial, though lower negative correlations between anger-out and anger-suppression. Anger-reflection/control and anger-suppression showed the weakest (positive) correlations overall, with the notable exception that at Time 3, r = .62 (p < .01). Table 13 shows the correlations between the three anger expression s t y l i s t i c s and state and t r a i t anxiety. State anxiety correlated with these scales at a moderate level in about half the cases, and not at a l l in the other half, with two exceptions: State anxiety and anger-out at Time 2 showed a positive correlation, r = .44 (p < .01), and state anxiety and anger-reflection/control at Time 3 were negatively correlated: r = -.48 (p < .01). Except for the modest positive correlations with anger-out at Time 3, t r a i t anxiety was not found to be significantly correlated with any of the anger expression s t y l i s t i c s . Intercorrelations between the Anger Expression S t y l i s t i c s Scale 1 ANGOUT 2 3 1 ANGSUPP 2 3 ANGRXC 1 -.60 -.61 -.40 .32* ns . 32* 2 -.51 -.72 -.47 .41 .42 .38 3 -.67 -.67 -.60 .33* .32* .62 ANGOUT 1 - a a -.53 -.30 -.45 2 - a -.54 -.54 -.52 3 — -.44 -.43 -.52 Note. See Table 3 for explanation of variable abbreviations. P < .01 unless otherwise indicated, ns = not significant. ^Intercorrelations between scores on the same measure taken at different times are not reported here, since the r e l i a b i l i t y coefficient for each repeated measure i s reported elsewhere. *p < .05. Intercorrelations between State and Trait Anxiety and the Anger Expression Sty l i s t i c s Scale 1 STANX 2 3 1 TRANX 2 3 ANGRXC 1 ns ns -.30* ns ns ns 2 ns -.31* -.26 ns ns ns 3 -.26* ns -.48 ns ns ns ANGOUT 1 .26* ns .39 ns ns ns 2 ns .44 .41 ns ns ns 3 .35 .30* .42 .28* .32* .33* ANGSUPP 1 ns ns ns ns ns ns 2 ns -.28* -.28* ns ns ns 3 ns ns -.38 ns ns ns Note. See Table 3 for explanation of variable abbreviations. P < .01 unless otherwise indicated, ns = not significant. *p < .05. Table 14 shows the correlations between the anger expression subscales and state and t r a i t anger. The state anger c o r r e l a t i o n s ranged from nonsignificant to moderate, the strongest relationships holding between state anger and the anger-out factor. Moderate to substantial correlations were found between t r a i t anger and a l l the anger expression scales. T r a i t anger at a l l times correlated negatively with anger-suppression at a l l times, at a moderate l e v e l . The co r r e l a t i o n s between t r a i t anger and anger-r e f l e c t i o n / c o n t r o l , also negative, were o v e r a l l stronger than those with anger-suppression. The p o s i t i v e c o r r e l a t i o n s of t r a i t anger with anger-out were the most marked, e s p e c i a l l y at Time 1 and Time 3, where high c o r r e l a t i o n s were noted. Intercorrelations between State and Trait Ancrer and the Anger Expression Sty l i s t i c s Scale 1 STANG 2 3 1 TRANG 2 3 ANGRXC 1 ns -.26* -.30* -.51 -.43 -.34* 2 ns -.38 -.35 -.41 -.43 -.38 3 ns -.41 -.50 -.56 -.59 -.57 ANGOUT 1 .28* .46 .52 .76 .67 .56 2 ns .41 .48 .67 . 69 .62 3 .41 .33* .56 .57 .58 .80 ANGSUPP 1 ns ns ns -.34* -.38 -.36 2 ns -.34 -.34 -.30* -.41 -.38 3 ns -.35 -.44 -.43 -.58 -.53 Note. See Table 3 for explanation of variable abbreviations. P < .01 unless otherwise indicated. ns = not significant. *E < .05. Ancillary Analyses Frequency Distributions and Histograms These were obtained for each variable, across the total sample and by program group, as a means of identifying patterns in the data and directions for supplementary s t a t i s t i c a l analyses. The most noteworthy pattern detected was the floor effect in the state anger distributions (STANGl, STANG2, and STANG3), in which most of the scores clustered at the low end of the scale. See Research Question 8, above, for a description of the results of the analyses which were performed consequent to this observation. Correlation Coefficients Significant correlation coefficients are reported above in the section or sections to which each variable pair i s most relevant. Test-Retest Consistency Table 15 reports the r e l i a b i l i t y coefficients obtained for each of the repeated measures. High r e l i a b i l i t y levels were found for a l l measures. R e l i a b i l i t y Coefficients for Repeated Measures Measure Alpha STANX1/STANX2/STANX3 .82^ STANGl/STANG2/STANG3 .89^ TRANXl/TRANX2/TRANX3 .90^ TRANG1/TRANG2/TRANG 3 .90^ ANGRXC1/ANGRXC2/ANGRXC3 .85^ ANGOUT1/ANGOUT2/ANGOUT3 .86^ ANGSUPPl/ANGSUPP2/ANGSUPP3 .74^ AMQ1/AMQ2/AMQ3 .80^ Note. See Table 3 for explanation of variable abbreviations. Calculation based on N = 55. Reduced N resulted from procedures used by s t a t i s t i c a l analysis program for handling missing data. ^Actual N = 57 (i.e., 1 case missing). '^Actual N = 56 (i.e., 2 cases missing). "^Actual N = 58 (i.e., 0 cases missing). Interrater R e l i a b i l i t y To obtain a measure of the consistency between the ratings given to the paintings by the two independent raters, the two sets of ratings on a l l of the paintings (n = 165) were correlated with each other. The correlation coefficient obtained for the two raters' assessments of degree of anger in the paintings was .72 (p < .01). The ratings for degree of defensiveness were less consistent (r = .55, p < .01). The test for overall consistency (measured with the ratings for anger and defensiveness taken together, i.e., over 330 cases) yielded a correlation of r = .63 (p < .01) . Analysis of Representativeness of Available Scores on A b i l i t y Measure The t tests conducted to assess the differences between the group of subjects having TCS scores and the group without TCS scores yielded significant results on three variables: age, state anger at Time 2, and anger-suppression at Time 2. The group with TCS scores was older (M = 148.2) than the group without (M = 142.5), t(56) = 2.03, p < .05. At Time 2, the group with TCS scores had higher state anxiety (M = 17.1) than the group without (M = 14.8), t(56) = 2.14, E < .05. Also at Time 2, the group with TCS scores (M = 8.6) scored lower than the group without (M = 10.1) on anger-suppression, t(56) = 2.37, p < .05 ( a l l probabilities based on the two-tailed test). CHAPTER V DISCUSSION This chapter reviews and interprets the significant findings reported in Chapter IV. The f i r s t section briefly summarizes the significant results pertaining to each of the research questions and two of the ancillary analyses. The following several sections, which constitute the main body of the chapter, present interpretations of the study's major findings, in order of importance, by topic. Next, the limitations of the study are discussed, and the chapter ends with a summary and conclusion. A few recommendations for future research are included in the main body of the chapter, while most are contained in the section on limitations and in the chapter summary. Summary of Results Results Pertaining to the Research Questions In this section, each of the numbered subsections refers to the research question of the same number. 1. Effect of the nature of the program component. Whether the subjects had most recently experienced the art or the didactic component appears not to have made a difference to their scores on any of the repeated measures, except for the expected difference on the Anger Management Questionnaire, the content of which was closely tied to that of the didactic component. 2. Order effect of program components. Analysis of the painting ratings reveals a possible order effect, showing an elevated degree of anger in the a r t - f i r s t group on Painting 2, and more defensiveness in the didactic-first group on Painting 3. These results are discussed in detail in the section on "Component Order and Content," below. 3. Degree of anger and defensiveness in the paintings. The degree of anger expressed in Painting 2 was greater than that in either of the other two paintings, both within each program-order group and across the sample as a whole. At the same time, there was no significant result for defensiveness. The section entitled "Changes in Anger Expression in the Paintings," below, interprets this finding. 4. Age effect. Other than a slight positive correlation between age and state anger at Time 1, the results for age are either nonsignificant or unclear. The findings are discussed below. 5. Abi l i t y . The only significant results from the a b i l i t y analyses are low correlations between the measure of a b i l i t y and a few of the anger-related measures. The section on "Ability Level," below, discusses these results in light of the possible unrepresentativeness of the available a b i l i t y scores. 6. Gender. The gender results reveal no trend, just indications that (a) the boys f e l t angrier than that g i r l s at Time 2, (b) the g i r l s could report more of the content of the didactic component than the boys at Time 2, and (c) the boys showed more defensiveness than the g i r l s in Painting 1. These results are discussed in detail below. 7. Locus of Control. Positive, low correlations between externality and t r a i t anxiety at Time 2, t r a i t anxiety at Time 3, and state anxiety at Time 3 suggest some relationship between locus of control and anxiety. Also, externals suppressed their anger more at Time 1, while internals expressed more anger on Painting 3. These findings are discussed below. 8. Other patterns noted. Three sets of results were analyzed: sample change scores on the repeated measures, the anomalous state anger frequency distributions, and correlations between the state and t r a i t variables. The sample change score analyses reveal that between Time 1 and Time 2, t r a i t anxiety appeared to decrease and subjects' scores on the Anger Management Questionnaire rose. These results are interpreted below. The analyses arising from the state anger frequency distributions, and the patterns noted in the correlations of state and t r a i t variables are also discussed in detail in the relevant sections below. Results of Ancillary Analyses Test-retest r e l i a b i l i t y . The r e l i a b i l i t y coefficients for the repeated measures scales are a l l quite high, with the exception of that for anger-suppression, which, though lower, i s s t i l l substantial. These results allow for a f a i r l y high degree of confidence in the s t a b i l i t y of the measures used. Interrater r e l i a b i l i t y . The consistency between the ratings given to the paintings by the two independent raters is high for degree of anger, and lower, though s t i l l substantial, for degree of defensiveness. The factors which may have contributed to the lower level of consistency between the two sets of ratings for degree of defensiveness are discussed below, in the section on "Changes in Anger Expression in the Paintings." As noted above, the correlations between the two raters' ratings were deemed to be high enough in both cases (anger and defensiveness) for i t to be appropriate to aggregate the two sets of ratings for the purposes of analysis. The Major Findings Changes in Anger Expression in the Paintings The predictions that the degree of anger expressed in the paintings would increase over the course of the art component, and that the degree of defensiveness would decrease, were based on the hypothesis that each time the subjects were given permission to express anger in this context, they would become more willing and able to do so. The results revealing that the greatest degree of anger was expressed in the second painting, not the third, and that there was no change in degree of defensiveness, clearly do not support this hypothesis. Rather, these results suggest an entirely different set of factors which may play a role in the degree to which children connect with their angry feelings and are motivated to express them. Each time the subjects were asked to paint, they were given a different stimulus activity and a different topic. For Painting 1, the stimulus component included discussion of a picture of angry people, and a very brief guided visualization. The painting topic was "The angriest time in my l i f e . " The stimulus activity for Painting 2 was a longer guided visualization, and the topic was "What I feel like in my body when I'm really angry." Painting 3 was preceded by a series of questions to be answered with a show of hands and the relating of personal experiences by a few students. The topic for this painting was " A time I f e l t good about my anger." It i s possible that the greatest degree of anger was expressed in Painting 2 because either the stimulus activity or the topic i t s e l f , or both together, incorporated content that was more evocative of anger than those for either of the other paintings. The guided visualization preceding the second painting activity (see Appendix B) invited the students to become aware of several specific physiological elements of anger. The stimulus for Painting 1 was a considerably shorter and less focussed visualization, while there was no guided visualization prior to Painting 3. Guided visualization (sometimes called guided fantasy) can be a powerful tool for releasing blocked emotions (Anderson, 1980; Kelly, 1972). Since a more detailed guided visualization preceded Painting 2 than Painting 1, and none was used to introduce Painting 3, i t i s possible to speculate that this factor alone accounts for the relatively high degree of anger expressed in the second painting. However, i t can also be argued that the topic assigned for each painting played at least as important a role in producing this result as the nature of the stimulus activity. The topic for Painting 2 asked the students to access more or less immediate bodily experiences while those for Paintings 1 and 3 required that they reconstruct incidents from the past. Perhaps i t i s easier to remember and symbolically express an emotion when one recalls the physical sensations associated with i t than when one simply recollects the events and dialogue surrounding i t . Theories such as bioenergetics (Lowen, 1975) hold that when feelings are suppressed, chronic muscular tension develops and a "muscular memory associated with the feeling" (Stark, 1982) is established. If this i s the case, accessing the muscular memory - in this instance through guided visualization - may be a powerful way of calling to awareness the inner feeling state with which i t is associated. From this perspective, the topic for Painting 2 was manifestly more evocative than those for the other paintings. It i s lik e l y that the two factors just discussed - the quality of the guided visualization, and the painting topic - both contributed to encouraging the expression of more anger in Painting 2. The relative proportion of their contributions would be an interesting avenue for further investigation. Two additional factors may have played a role in producing the result under discussion. First, in the f i r s t painting lesson, many students may have been intimidated by the "art work" aspect of the activity - many hesitated up to 10 minutes before starting to paint. It i s quite possible that their fear of being perceived as incompetent ar t i s t s i n i t i a l l y served to limit the intensity of their emotional expression in the painting. Second, i t may be that the topic for Painting 3 lacked sufficient c l a r i t y to be interpreted in a similar way by a l l the subjects. The content of the third art lesson was intended to parallel the third didactic lesson in i t s focus on constructive angry behaviour. Thus the topic for Painting 3, "A time I f e l t good about my anger," was explained to the students as "a situation in which you f e l t good either because you were able to solve the problem, or because you were able to say clearly that you were angry and why" (see Appendix B). It is possible that there was confusion among the students about the meaning of the topic - after a l l , feeling good and feeling angry are normally conceived of as contradictory states. In every program group, this topic was interpreted in at least three different ways. Of the total of 55 paintings produced in the third art lesson, 32 expressed a theme of revenge and/or hatred, that i s , the students seemed to be saying, "I f e l t good when I got sweet revenge," or "I'm so angry at these people that i t feels good to hate them." Another 8 paintings contained a theme of righteous anger: "I can feel good about my anger when i t comes from my being falsely accused and I am vindicated, that i s , when i t ' s clear that I have right to be angry." A variety of other themes were expressed in the remaining 15 paintings, including two which depicted the ar t i s t feeling good about angrily yelling at someone, and a few which contained the prescribed theme. Clearly, most of the students were not willing or able to paint "A time I f e l t good about my anger" in the way intended in the lesson plans. It is quite possible that very few of them had ever actually had such an experience. It is also conceivable that the topic, framed as i t was, and presented when i t was - just after the second art lesson when so many subjects had expressed a great deal of anger, and they were perhaps primed to express more - evoked a particular set of strong feelings associated with anger: the desire for revenge, pleasure in planning i t , the exhilaration of achieving i t . Supporting this hypothesis is the fact that in Group 3, in particular, in which 13 of 15 paintings reflected the revenge/hatred theme, a high level of excitement and agitation was evident throughout this lesson. The students may not have been ready to move on from exploring angry feelings to a genuine consideration of problem-solving in angry situations. In addition, many lik e l y experienced a conflict between what they were being asked to do and what they f e l t impelled to do (e.g., express revenge). The ratings of the paintings no doubt reflect the confusion there was over the painting topic and the conflict i t may have occasioned. The ratings were further influenced by the ways in which the raters interpreted the evidence of this confusion in the paintings. Both raters saw "defensiveness" in paintings in which, the researcher believes, the subject was merely attempting to follow the instructions with regard to the topic to be depicted (i.e., feeling good when angry). In addition, while rating the paintings for defensiveness, one of the raters developed an erroneous hypothesis about the themes of the paintings, and rated the remaining pictures accordingly. This fact may help to explain the lower degree of interrater consistency on degree of defensiveness than on degree of anger, and may also have a bearing on the fact that the ratings reveal no difference in defensiveness between the paintings. Given these conditions, a l l the defensiveness results must be interpreted with caution. Nevertheless, the results discussed in this section strongly suggest that subject matter - the content of instructions - makes a difference when i t comes to encouraging the expression of anger. They may also be interpreted as suggesting that before students can be open to the p o s s i b i l i t i e s of constructive angry behaviour, they must somehow complete the process (or a stage in the process) of exploring and expressing a f u l l range of feelings and fantasies associated with anger. Component Order and Content Evidence from the paintings. In two instances, there was a clear association between the painting ratings and order of exposure to the program components. The a r t - f i r s t group expressed a greater degree of anger than the didactic-f i r s t group in Painting 2, and the didactic-first group was rated as more defensive in Painting 3. In interpreting these results, i t i s useful to note that the a r t - f i r s t group produced Painting 2 in their second lesson, while the d i d a c t i c - f i r s t group painted for the second time in their f i f t h lesson. For the latter group. Painting 3 was the last lesson. A possible explanation for these findings may l i e in the different roles played by the experimenter in the two different program components. The didactic component required the experimenter to perform the functions of a teacher: to give instructions, provide information, ask questions (both open-ended and closed), test understanding, direct student learning toward specific goals, and manage student behaviour within a classroom setting. By contrast, in the art component the experimenter became more like a therapist: providing a safe setting for student self-discovery and offering non-judgemental reflections of the material produced. The students in the d i d a c t i c - f i r s t group knew the experimenter f i r s t as a teacher; by the time they began the art component the teacher-student relationship -which many students perceive as inherently c r i t i c a l and controlling - was well established. The a r t - f i r s t group entered directly into the more allowing, less structured therapist-client relationship with the experimenter and were thus perhaps less inhibited when asked to express anger in a painting. Anecdotal evidence from the experimenter's program implementation log supports this hypothesis. The students in both of the program groups which underwent the art component f i r s t (Groups 3 and 4) showed considerable resistance to participating in the "Anger Buttons" exercise in the f i r s t didactic lesson. Many students seemed afraid either to write down examples of things that would make them angry or to read their examples to the group. It i s possible that these students f e l t more vulnerable doing this exercise than the didactic-first group because through the painting, in the safety of the therapist-client relationship, the former had revealed - at least to themselves - some normally well-hidden psychic material, and, with the sudden shift back to teacher-student and classroom mode, they became frightened of exposing themselves to the world by putting this material into words. At the same time, many in the dida c t i c - f i r s t group had a lot of d i f f i c u l t y starting to paint in their f i r s t art lesson, expressing similar kinds of reluctance as had the a r t - f i r s t group in their f i r s t didactic lesson. Contrary to the researcher's expectations, knowing the experimenter did not increase these students' i n i t i a l comfort level with the painting activity. The students sensed that the relationship had changed and that a different kind of participation was being asked of them. The transition created some discomfort. Perhaps the coarseness of the measure used to acquire these results accounts for the fact that no difference in anger expression was found between the two program-order groups on Paintings 1 or 3. Alternatively, the evidence could suggest an interaction of effects: The a r t - f i r s t group's Painting 1 anger ratings were relatively low because the topic and the activity did not inspire a high degree of anger expression (and possibly also because of first-lesson anxiety); for Painting 1 the didactic-first group was neither inspired nor trusting of the new instructor-student relationship. For Painting 2, both groups were inspired, by the topic and the guided visualization, to express more anger than before, with the didactic-first group expressing less, for the reasons given above. For Painting 3, a less simulating topic and activity combined with confusion about the topic (on the part of both the subjects and the raters) to produce lower anger ratings in both groups. A possible explanation for the higher degree of defensiveness found in the third paintings of the didactic-f i r s t group i s that in the last lesson of the program this group, knowing that the program was about to end, may have been particularly conscious of the imminent termination of their access to the contained safety of the art media (Mills, 1985) and the special therapist-client relationship, which perhaps, they had just begun to u t i l i z e . Consequently they may have raised their defences to a degree that the a r t - f i r s t group (doing Painting 3 in the third lesson) had no need to do. The same degree of caution, however, must be applied to the interpretation of this defensiveness result as to a l l the others, as noted above. Overall, the results discussed in this section suggest that simply talking about anger f i r s t , within the context of didactic lessons on anger management, does not seem to provide any extra level of comfort or willingness to express i t through painting. The c r i t i c a l factor may be the nature of the relationship - didactic or therapeutic - established between the instructor and the students. Evidence from the repeated measures. With one exception, no relationship was shown between subjects' scores on any of the repeated measures and either the content of the two different program components, or the order in which the subjects experienced them. The one exception was the Anger Management Questionnaire, on which the didactic-first group scored higher at Time 2, and on which the a r t - f i r s t group improved i t s mean score between Time 2 and Time 3. This unsurprising result merely confirms that the content of the questionnaire pertained more directly to the didactic component than to the art component (as i t was designed to do), and that the students were capable of reproducing the expected answers. High Scoring Subjects on State Anger The analyses investigating how the small group of high scorers on state anger stood apart from the rest of the sample constituted an exploration by s t a t i s t i c a l means of a c l i n i c a l l y interesting phenomenon. Of course, the group's small n precludes any conclusions based on s t a t i s t i c a l significance. However, the results of these analyses do offer some clues and a point of departure for speculations of c l i n i c a l and theoretical relevance. Increasing anger expression. It i s perhaps not surprising that the high scorers showed more evidence of using the anger-out s t y l i s t i c than the rest of the sample. However, this difference only became evident at Time 2, quite soon after most of the group (7 out of 9) had painted for the third time. It was on this third painting that the high scorers as a group clearly expressed a greater degree of anger than the rest of the sample. (Perhaps this group chose to use the third painting lesson as an additional opportunity to express anger, rather than to attempt to follow instructions and produce a "feeling good" picture.) The elevated scores on anger-out (a t r a i t factor) were sustained at Time 3, at which time the high scorers also reported higher levels of t r a i t anger than their peers. Taking these results together, i t becomes possible to postulate a trend: From the beginning, as a group, the high scorers on state anger were willing to report high levels of state anger and anxiety; as the program progressed, they became more willing to admit to characteristic anger in addition to anger in the present tense. Since such a large proportion of the group took the program's art component f i r s t , i t may be reasonable to conjecture that the process of painting helped this group to become more aware and possibly more accepting of their tendencies to anger. Evidence of a trend of increasing anger expression among the members of the group can also be found in the paintings themselves. Of the 9 high scorers who completed Painting 3, 5 produced pictures that were rated as angrier than either their f i r s t or their second. For Painting 1, one of these (Subject A) produced "I'm sad," a picture of himself crying upon learning that he had a chronic illn e s s . (See Appendix N, "Paintings of High Scorers on State Anger.") The same subject's next painting, called "Mold," depicted mold, a broken heart, and someone throwing up in a t o i l e t : sickness and decay, sadness, and disgust. His f i n a l picture was "Ow! Ouch! Insult!" It showed him physically and verbally assaulting someone on four different occasions. A clear progression, from sadness, through disgust to rage i s evident over the course of these paintings. Another of the high scorers (Subject B; see Appendix N) began with a painting called "The Library," intended to show an incident in which the student was angry at a librarian. However, the student was unable or unwilling to finish the painting, so a l l i t showed was single female figure, standing calmly on a green surface. This student had tremendous d i f f i c u l t y deciding what to paint for her second picture. Finally she produced "The Boring Weekend," a l l blue sky, green grass and blank space. One of the raters saw anger at the moderate level under this restrained surface. In her third painting Subject B made a breakthrough with "The Bedroom Window," a depiction of herself and her parents at a time when she was "grounded" for something she didn't do, and "escaped" out her bedroom window. She refused to apologize (for what she hadn't done) and f e l t good about that - one expression of the theme of righteous anger. In the painting, she coloured herself a deep purple, and her parents light pink, surrounding the three figures with messy strokes of red, black and green. In this case, the student progressed from being unable or unwilling to depict an incident involving anger, through a stage of being unable to completely hide her angry feelings, to a f u l l and open representation of both an incident and her feelings about i t . A f i n a l example serves to illustrate a different kind of development of emotional expression over the course of the art component. One of the high scorers (Subject C; see Appendix N) began with a painting called, "Boy with B a l l , " showing a cutaway view of the inside of his house, with thick black walls, roof and chimney. He said, "This i s like my house in a dream." Inside are two small, indistinct figures. A purple b a l l hangs over the head of one of them. The student said his brother threw the ball at him. The painting seems to offer a tightly controlled glimpse of the scene of the student's anger. During the second lesson of the art component Subject C voluntarily produced two paintings. The f i r s t , called "The Rage," consisted of paint slathered and mixed a l l over the page, with red and yellow - "like flames, showing my anger" - and darker oranges and browns for "sadness." At the end, he used his fingers to draw a frowning face in the wet paint on the paper. "That's my face, I'm sad." For his second painting he splattered paint of a l l colours over the page. He said the dark colours were the problem, and the other colours were "helping the problem get better." He c l a r i f i e d : "The dark colours are like a disease, and the other ones are like the medicine." This painting was called "Helping Confusion." The f i r s t painting of this pair expresses the student's messy, uncontrollable, d i f f i c u l t feelings that range from rage to sadness. The second painting seems to indicate that, having put some of those feelings out on paper, he can now begin to see some bright spots of hope emerging from the confusion. This student's fi n a l painting, entitled "Death by Medicine B a l l " revisits some of the same themes. A large orange sun with long yellow rays dominates the picture. Two small figures depict the student and his brother, who has just called him a name. The student is throwing a flaming medicine b a l l at his brother. While painting, the student was very frustrated because he wanted to use a smaller brush to paint finer detail but none was available. He said that a l l the other times he painted he had used a big brush and that that's why his work had been so messy. Both his concern with neatness and control, and the carryover of the metaphor of medicine, seem to indicate a continuing progression for this student towards a desire for control and healing. In the f i r s t painting, the bal l hung ominously over his head; by the last he had taken control of i t . This progression i s not as linear as that noted in the paintings of some other students, but for this student also there was clearly movement through various stages of anger expression. Low scores on Anger Management Questionnaire. The high scorers' Time 2 Anger Management Questionnaire scores were lower than the rest of the sample's no doubt because 7 of the 10 subjects in this group had not yet taken the didactic component at Time 2. However, their scores were s t i l l significantly lower at Time 3. This result may suggest that a high level of state anger can interfere with a student's a b i l i t y or motivation to engage with the cognitive-behavioural content of an anger management program. If so. i t would support either Nannis' (1988) suggestion that a highly emotional situation can result in "the constriction of cognitive s k i l l s available for use" (p. 45), or the contention of Saylor et a l . (1985) that i t i s possible for children to feel too angry to care about anger management. State and t r a i t anxiety. The fact that the high scorers consistently reported high state anxiety suggests that there may be a close relationship between state anger and state anxiety. This hypothesis is supported by the substantial correlations found between the sample scores of these two variables (see Table 11 and "Correlations between State and Trait Variables," below). In contrast with state anxiety, t r a i t anxiety did not appear to be elevated among the high scorers on state anger. This result, along with the fact that the correlations noted between state anger and t r a i t anxiety were low or non-significant (see Table 11), provides reassuring support for the assumption that the state and t r a i t anxiety questionnaires were not measuring the same thing. Anger and defensiveness in Painting 1. The results of the analyses of Painting 1 provide a study in contrasts. More high scorers than others were rated as expressing both extreme anger and none. At the same time, more high scorers than others were rated as exhibiting both extreme defensiveness and none. A possible explanation for this phenomenon is that for this group there was no middle ground: Either they expressed anger, or they didn't; either they attempted to hide i t or they didn't. Fully half of the high scorers were rated as extremely defensive, compared to less than a quarter of the group of others. It i s also interesting to note that the percentage of high scorers who expressed extreme anger is the same as the percentage of high scorers who showed no defensiveness. This finding suggests that there may be something more than a logical relationship between degree of anger and degree of defensiveness. Of course, interpretation of the results for defensiveness can only be very tentative because of the problematic nature of the data upon which they depend (discussed above). Correlations between State and Trait Variables The marked correlations found between state anxiety and state anger (see Table 11) suggest a relationship between these two factors, one which makes intuitive sense, given the social unacceptability of anger. Further investigation of this association could well result in enhanced understanding of the phenomenon of anger. It i s interesting to note that, while there is a clear inverse relationship between anger-out and both anger-reflection/control and anger-suppression, the association between the latter two anger expression s t y l i s t i c s i s not as consistently strong (see Table 12). These findings suggest that children who characteristically express anger directly don't suppress i t , nor, more interestingly, do they attempt, while angry, to control their anger and resolve the problem. At the same time, anger-reflection/control does not necessarily involve anger-suppression or vice versa. Helping students learn anger management - when i t i s defined as anger control and conflict resolution - probably entails encouraging children to find a middle way between overtly expressing angry feelings and simply suppressing them. While state and t r a i t anxiety are associated only weakly, i f at a l l , with any of the anger expression s t y l i s t i c s (see Table 13), and state anger shows just slightly larger correlations with them (see Table 14), t r a i t anger i s markedly related to these factors, particularly to anger-out (see Table 14). This finding tends to confirm that anger-out i s a t r a i t variable, very similar to t r a i t anger. Children who see themselves as characteristically angry also see themselves as expressing that anger in an open, often hostile way. At the same time, they tend not to attempt to control their anger or resolve conflicts. Less strongly, the results suggest that these children don't suppress the anger they feel. Another pattern noted in the correlation matrices for the state and t r a i t variables shows the strength and significance of the association between 10 of the variable pairs increasing over time. That i s , for each of these pairs of variables, the size of the correlation coefficient increased between Time 1 and Time 2, and again between Time 2 and Time 3 (see Table 16; see also Tables 11 - 14). Variable Pairs with Increasing Correlations Between Time 1 and Time 3 Correlation coefficient Variable pair Time 1 Time 2 Time 3 STANX/TRANG ns .40 .49 STANG/TRANG .28* .55 .69 TRANX/TRANG .41 .48 .60 ANGRXC/ANGSUPP .32* .42 .62 ANGRXC/STANX ns -.31* -.48 ANXRXC/STANG ns -.38 -.50 ANGOUT/STANG .28* .41 .56 ANGSUPP/STANX ns -.28* -.38 ANGSUPP/STANG ns -.34 -.44 ANGSUPP/TRANG -.34* -.41 -.53 Note. See Table 3 for explanation of variable abbreviations. P < .01 unless otherwise indicated, ns = not significant. *p < .05. Conceivable explanations for this pattern include the poss i b i l i t y that i t reflects some effect of the interaction that occurred between the experimenter and the subjects as teacher and students, as therapist and clients - over the course of the anger management program. Another plausible interpretation of these results i s that they manifest the influence of a demand characteristic: As the subjects repeatedly encountered the same set of questionnaires, they became more and more aware of the connection between the anger management program and the "tests," and became more and more faci l e , practised or deliberate in their responses. A third possibility is that both of these factors came together in some combination to produce these results. With such a small N i t is not possible to determine the actual reason for this pattern. Simply noting i t here can serve the purposes of future research by calling attention to these p o s s i b i l i t i e s . Gender The gender results seem to indicate that, at Time 2 (and only at Time 2) the boys were angrier than the g i r l s . Correlational, chi-square and t test analyses a l l support this conclusion. Possible explanations for this result include the following. A playground or classroom event involving mostly boys rather than g i r l s in the hours or days preceding administration of the intermediate-test questionnaires may have precipitated a general rise in state anger levels among the boys. Alternatively, having been encouraged in the f i r s t half of the anger management program to discuss and depict anger, the boys, (for some reason) more than the g i r l s , may have taken this encouragement as permission to acknowledge their anger more freely on the questionnaires. Or the g i r l s may have taken the questionnaires more seriously than the boys, tending to exaggerate their responses less or report them with more honesty. However, speculative interpretations such as those just offered must be tempered by the facts that f i r s t , the intercorrelation i s not strong, and second, the state anger results may not be particularly revealing because of the floor effect in the state anger distributions, where most of the scores clustered at the low end of the scale. This phenomenon, an effective reduction of the range, indicates the measure's insensitivity and eliminates the possibility of being confident about any of the state anger results. Also at Time 2, the g i r l s scored higher than the boys on the Anger Management Questionnaire. This overall gain by the g i r l s between Time 1 and Time 2 could be evidence of pre-adolescent g i r l s ' superior verbal a b i l i t y (Maccoby & Jacklin, 1974a), or i t may be a manifestation of the tendency for school-age g i r l s to be more highly motivated than boys and so to reach higher levels of academic achievement at this age (Frymier, 1980). However, in either case, the difference should have been sustained at Time 3, and i t was not. Another possibility is that the difference simply reflects g i r l s ' greater tendency to obey the directions of adults (Maccoby & Jacklin, 1974b), and so to supply the expected answers. Finally, on Painting 1, the boys were rated as more defensive than the g i r l s . This result supports the conventional wisdom that boys are more reluctant than g i r l s to express themselves through art media. However, this pattern was not sustained over the other paintings. It i s possible that the nature of the instruction and/or ac t i v i t i e s during the art component allowed the boys to grow more comfortable with expressing angry feelings, at least on paper. In sum, the present study offers some evidence, i f inconclusive, of possible gender effects pertaining to state anger, defensiveness, and scores on the Anger Management Questionnaire. These findings point directions in which further investigation may be f r u i t f u l , especially with regard to the possible difference between the sexes in defensiveness around art work, and the interplay between gender and state anger. (It is interesting to note, that of the 10 high scorers on state anger, only 1 was female.) However, the fact that so few differences were noted suggests that, contrary to expectation, on many of the factors relevant to anger and anger management, boys and g i r l s may not be dissimilar. Further research using a wider variety of more precise measures is required before the relationship between gender and such anger-related factors can be c l a r i f i e d . Change Scores on Repeated Measures Trait anxiety was seen to decrease between the f i r s t administration of the questionnaires, prior to the beginning of the anger management program, and the second, halfway through i t . The lower level of anxiety noted at Time 2 was sustained at the posttest. Time 3. This finding could indicate that the subjects experienced i n i t i a l anxiety regarding the questionnaires, which then abated, as the context from which they arose became apparent. This interpretation is supported by anecdotal evidence from the experimenter's program log. During the lessons, a number of students referred to the questionnaires as "tests." The f i r s t set came out of the blue, as i t were, and the teachers administering them had been instructed not to inform the students of their connection with the anger management program (of which, at that point, the students were unaware). The intermediate and posttests were probably less threatening to the students than the pretest because by Time 2 they had realized that the questionnaires were connected with the anger management program and were not tests in the usual sense. However, i f this interpretation i s correct, i t would seem more l i k e l y for the effect to be revealed in the state anxiety data rather than in t r a i t anxiety. There was no change in the sample scores on state anxiety. It i s possible that the t r a i t anxiety result is some kind of measurement error anomaly. Nevertheless, an interpretation of this result (such as that offered above) based on changes in the subjects' anxiety levels remains plausible. The t r a i t anxiety change scores, along with the substantial correlations found between state anger and state anxiety together suggest that students' level of anxiety over the course of an anger management program may be worth further investigation. Sample scores on the Anger Management Questionnaire were found to increase between the pretest and the posttest. Because the properties of this instrument are unknown, a l l this result can suggest i s that the subjects were capable of reproducing the responses they had deduced to be the expected ones. After the program, the students were better able to answer questions in conformance with program content than they had been before the program. This result i s not an indication of the students having learned the anger coping strategies referred to in the questionnaire. Age as a Factor The correlation found between age and state anger at Time 1, the only correlation noted between age and any of the anger-related measures, was rather small and not particularly compelling conceptually. It can therefore be regarded as anomalous and probably specious. The significant chi-square result for defensiveness in the second and third paintings indicates that there may be some sort of relationship between a subject's program group and the degree of defensiveness noted in his or her paintings. The question is whether this relationship has to do with age or some other function of program group. Two kinds of evidence stand against the age hypothesis. First, age did not correlate in any case with the ratings of the paintings. Second, as indicated in Table 17, for Painting 3 the two youngest groups (Groups 2 and 3) had quite different percentages of scores in both the none (no defensiveness) c e l l and the extreme (high defensiveness) c e l l . In addition. Groups 3 and 4 (the oldest and second-youngest groups) - together comprising the a r t - f i r s t program-order group - had much higher percentages of scores in the none c e l l than Groups 1 and 2 (the youngest and second-oldest groups) - which together comprised the did a c t i c - f i r s t program-order group. These data suggest that the factor determining the relationship between defensiveness and program group i s not age, but rather order of exposure to program components, and they reflect the results discussed above showing that the did a c t i c - f i r s t group was more defensive in Painting 3. However, Table 17 also shows that the results for Painting 2 do not suggest the same conclusion. Here Group 1 seems to stand apart from the others with a very high percentage of paintings showing either moderate or extreme defensiveness (94%), while in Group 3, 50% of the paintings were rated as showing no defensiveness at a l l . At the same time, the results for Groups 2 and 4 cluster too close together to support either the age hypothesis or the program-order hypothesis. It i s quite possible that the confusion noted above regarding the rating of the paintings for defensiveness contributed to the ambiguity of these results. It i s also possible that the mean ages of the program groups were not sufficiently different to guarantee that an age effect would emerge from this analysis. A multiple regression analysis would be an appropriate way to give more order and c l a r i t y to the details of the present data and might reveal an age effect hidden in the current analysis. Crosstabulations of Program Group by Degree of Defensiveness Degree of defensiveness (%)^ Group Program Mean age No. order'^ (mos. ) None Moderate Extreme Painting 3 1 D - A 154 12 31 58 2 D - A 135 19 50 31 3 A - D 141 37 50 13 4 A - D 155 36 36 29 Painting 2 1 D - A 154 7 57 37 2 D - A 135 43 21 36 3 A - D 141 50 37 13 4 A - D 155 39 14 46 of n for each group, rounded to nearest whole number. ^Order of exposure to program components: D - A = didactic f i r s t , A - D = art f i r s t . A b i l i t y Level A b i l i t y was included as a variable in the current study because i t was deemed necessary to distinguish this factor from that of chronological age. The constraints of the study, however, prevented the direct administration of an a b i l i t y test and forced reliance on an incomplete set of somewhat dated scores. The incompleteness of the data batch in turn necessitated an extra stage of analysis to determine the representativeness of the available scores. Representativeness of available scores. The fact that the subjects with scores on the a b i l i t y test (TCS) were older than those without can be explained with reference to the datedness of the scores. The test was administered to the subjects, by class, at the beginning of the school year prior to the school year during which they participated in the present study. At that time, class compositions differed, and i t appears that only one younger class - which included only some of the younger participants in the study - was given the test. More of the older students were members of classes which were tested. The other differences between the group of subjects with TCS scores and those without are not so readily explained. Even taking into account the standard error of measurement, the differences between them on state anger and anger-suppression at Time 2 remain significant. If i t could be shown that these two TCS groups align with the program-order groups, then i t might be argued that the differences are a function of order of exposure to program components. But in fact the program-order groups had almost equal numbers of missing cases, with the didac t i c - f i r s t group accounting for 53% of them and the a r t - f i r s t group encompassing 47%. It could be speculated that the differences between the two TCS groups were related to the age difference. The older group (with scores) exhibited greater Time 2 state anxiety and scored lower on anger-suppression at Time 2 than the younger group (without scores). Many subjects in the group with scores were at or approaching adolescence: 62% were aged 12 years or older, while 42% of the group without scores were this age. It i s possible that the older group was beginning to exhibit some of the totalism of adolescents in a tendency to insist on showing their true feelings. However, there i s no correlational evidence in the current study to support such a speculation. It i s worth noting that on two of the three occasions on which state anxiety and anger-suppression were measured there was no difference between the two TCS groups. States are by definition unstable, so perhaps i t should not be surprising to find evidence of a f a i r l y wide variation in scores over time on a measure such as state anxiety. As a t r a i t , anger-suppression should show less v a r i a b i l i t y over time than state measures. But i f anger-suppression can be conceived as being inversely related to anger arousal, and to the extent that anxiety may play a role in the latter (Wickless and Kirsch, 1988), i t may be possible to argue that there is a similar relationship between state anxiety and anger-suppression. In the current study, a slight inverse correlation was noted between state anxiety at Time 2 and anger-suppression at Time 2 (r = -.28, P < .05), a relationship which increased in strength at Time 3 (r = -.38, p < .01). However, further investigation, beyond the scope of the current study, i s required before any conclusion can be reached in this regard. While i t cannot be maintained that the group of subjects with TCS scores is entirely representative of the group without, the anomalous quality of the few differences observed suggests that i t may be appropriate to use the available scores, with caution, in the analysis of a b i l i t y level as i t relates to the anger-related measures. Ab i l i t y results. The correlations noted between TCS scores and state anger at Time 1, Anger Management Questionnaire scores at Time 2, and degree of defensiveness in Painting 1, are a l l f a i r l y low and are unsupported by any significant distributional differences. It can therefore be concluded, with the degree of caution warranted by the inconclusive results discussed in the previous section, that a b i l i t y level was not an important factor in determining the subjects' responses on any of the anger-related measures. Locus of Control The correlational evidence suggests that children with a more external locus of control have a tendency to greater State and t r a i t anxiety. These results align with the results of several studies summarized by Phares (1976), indicating positive correlations between anxiety and externality. However, Biaggio's (1985) finding of a high negative correlation between externality and t r a i t anxiety suggests that the relationship between locus of control and anxiety may not be so straightforward. In the current study the results of the exploratory analysis undertaken using the four most extreme scores at either end of the scale did not provide any support for the hypothesized relationship between externality and anxiety. Rather, the only variable for which a relationship was revealed was anger-suppression, with the externals scoring higher than the internals at Time 1. It i s possible to speculate that the externally oriented children claim to suppress anger because they believe that i f they don't suppress i t i t w i l l take them over, becoming an external force over which they have no control. Or i t may be that they attempt to hide their anger for fear of not being able to control the consequences i f they express i t . Unsupported by other evidence, however, this result does not admit of interpretation. The result for Painting 3, showing that the internals expressed more anger than the externals (perhaps because they f e l t more able to control i t ) , would at least not be at cross-purposes with the aforementioned anger-suppression result. However, i t is not possible to conclude anything from the painting result, since the size of the sample portion used in this analysis was even smaller than the 4 per group intended: Student absences from the lesson in which Painting 3 was produced reduced the number of subjects in each group to 3. The locus of control analyses were performed in an attempt to address some of the conative aspects of anger and anger management. The analysis using only the most extremely internal and external scores was undertaken as a potentially f r u i t f u l avenue of exploration, but under the circumstances of the current study, i t has proven to be ineffectual even in pointing directions for further investigation. Yet both locus of control in particular, and conation in general, should not be discounted in future studies of children's anger and anger management. Research involving the use of larger sample sizes may well reveal trends and patterns in this area that were undetectable in the present study. Limitations of the Study The limitations of the current study f a l l into two categories: f i r s t , the methodological problems inherent in the study design, data collection, and analysis; and second, the shortcomings of the anger management program used in the study. These two different kinds of limitations are treated in separate sections below: Research Limitations and Program Limitations. Research Limitations Since the current study was designed as an exploratory investigation, i t s limitations are less important with respect to the generalizability of the results, than they are in reference to their interpretation. Several aspects of the study's design are tenuous enough to cast doubt on the tentative conclusions reached above. Unproven measures. The research limitations include the unproven nature of several of the measures used. The Pediatric Anger Expression Scale (Jacobs et a l . , 1989) i s a new instrument s t i l l in the process of being developed, for which few r e l i a b i l i t y and validity studies have been conducted. Similarly, no supporting data i s available for the Pediatric Anger Scale (PANG) and the Pediatric Anxiety Scale (PANX) (Jacobs, cited in Jacobs et a l . , 1989). The Anger Management Questionnaire was developed specifically for use in this study, and i t s properties are unknown. As already noted, the floor effect in the state anger distributions calls into particular question the valid i t y of this measure and any of the analyses performed using the state anger results. However, very few children's measures of any aspect of anger exist. The current study can make a valuable contribution to knowledge in the f i e l d by adding to the database on the PAES, the PANG and the PANX. Self-reports. A limitation which the present study shares with much psychological research i s the reliance on self-report measures. The demand characteristics of such measures, especially when what is being reported i s something as socially unacceptable as anger, make suspect the results they produce. However, as A v e r i l l (1983) pointed out, "there is a certain necessity - not just a convenience - about the use of self-reports for investigating a phenomenon like the everyday experience of anger" (p. 1155). Av e r i l l contends that much can be learned from what people can t e l l us about their everyday lives, in "ordinary language" (p.1155). Although the possibility must be borne in mind that the results of the self-reports were confounded by reactivity or by a social desirability response set, or both, i t would not be appropriate to discount entirely the information they offer. The quality of the particular self-report measure may be more significant than the nature of self-reports in general. The qualitative limitations of the measures used in the current study have already been noted above. Anomalies in the raw data results from repeated measures questionnaires. The anomalies in the raw data from the questionnaires (described in Appendix K) reflect a number of other limitations of self-report measures. One i s that they leave open the possibility of subjects f a i l i n g to complete them in good faith. The several instances of the same response being given to every item on a single questionnaire are probably examples of such "malicious compliance" (L. D. Travis, personal communication, 1993), as may be the cases of apparent conceptual discrepancy between the responses to different items on the same questionnaire. However, because i t i s not possible to be completely certain that these items were completed in bad faith, as i t were, they were scored at face value and the results were included in the present analysis. Another limitation of self-reports reflected in the anomalous questionnaire data is that such measures are incapable of capturing the f u l l range of possible subject response to any given item. Some of the anomalous data resulted from subjects attempting to respond with options not available on the given scale. This was particularly true on the locus of control scale and the Anger Management Questionnaire. On the Anger Management Questionnaire, confusion was also l i k e l y caused by the wording of some of the questions, particularly numbers 14 and 15 (see Appendix G). In addition, on this questionnaire i t is possible not to know the answers. However, the greatest confusion, as evidenced by the anomalous responses to questionnaire items, occurred not at Time 1, but at Times 2 and 3. Perhaps having a second and third opportunity to think about the answers allowed the subjects to doubt their i n i t i a l certainty. More lik e l y , by Time 2 for half of the subjects, and by Time 3 for a l l of them, their exposure to the didactic component of the program had led them to believe that there were "right" answers to the questions, not just individual opinions, and they therefore approached the questionnaire with more se l f -doubt and less certainty. Whatever the reasons for their occurrence, taken together, the anomalies in the responses on the repeated measures constitute a notable limitation of the current research, and provide another reason for provisional interpretation of the results. Anomalies in paintings analysis. A number of problems became evident during the analysis of the paintings. As noted above, one of the raters developed an erroneous hypothesis about the themes of the paintings, and rated the remaining paintings accordingly. The relatively low degree of interrater consistency on degree of defensiveness l i k e l y resulted, at least in part, from this circumstance. It may also have reflected the fact that the construct of defensiveness. as defined in the present study, lacked the intuitive c l a r i t y of the idea of anger. The correlation between the two sets of ratings on defensiveness was s t i l l considered high enough for them to be aggregated for the purposes of analysis. However, i t could be argued that this aggregation of data was not, in fact, appropriate, and that the results for degree of defensiveness cannot be regarded with any degree of confidence. There i s an additional factor contributing to the uncertainty of the defensiveness results. The raters were blind to the nature of the study, having been told only that i t dealt with anger in children. They were not aware that each subject had painted three different times, or that specific topics had been assigned for each painting. These conditions were established, of course, so as to avoid contaminating the results with the raters' preconceptions. However, the raters' unawareness of the context in which the paintings had been produced caused a different kind of distortion in the results: In many of the third paintings the raters saw "defensiveness" where the subject was merely attempting to follow the experimenter's instructions with regard to the topic to be depicted -feeling good when angry. While this d i f f i c u l t y was, in part, a function of the perceived ambiguity of the topic (as discussed above), i t would have been obviated i f the raters had known what the subjects had been instructed to paint. Alternatively, the use of a more structured method for analyzing the paintings might have built in some safeguards against this kind of confound. For example, raters could be asked to base their assessments of the paintings on a specific set of c r i t e r i a which operationally define anger and defensiveness (or any other dimension of interest) for rating purposes. Another possibility would be to pursue a different but perhaps equally f r u i t f u l avenue of investigation by deliberately informing the raters of the painting topics; they could then be asked to assess whether or not each painting in fact depicted the assigned topic, and in what different ways the topic was interpreted or expressed. Future research involving an analysis of artwork similar to that undertaken in the current study w i l l require the careful development of appropriate methodologies for assessing the emotional content of the art. A b i l i t y measure. A major shortcoming of the study was the failure to obtain a f u l l data batch for the measure of ab i l i t y . Since the representativeness of the available scores was not definitively established, the results involving these scores must be interpreted with some caution. Replicability of lessons. One inevitable limitation of the study was that the six lessons of the anger management program, as taught, were not exactly the same for each of the four program groups. They varied from group to group in mostly minor and subtle ways. During the art lessons for both didactic-first program groups, the experimenter used some of the vocabulary which had been introduced to these groups in their didactic lessons. Of course during their art lessons, the a r t - f i r s t groups had not yet been exposed to this vocabulary and so did not hear i t at a l l during the art component. With each new teaching of the same lesson plan, the experimenter made small improvements - in emphasis, timing, amount of probing, and so on. This was not done deliberately, and hardly consciously, but simply as a part of the process of teaching. In addition, inevitably, the experimenter brought less enthusiasm to the third and fourth teachings of a lesson than to the f i r s t and second. (See the footnotes to Appendix B, Lesson Plans for the Anger Management Program, for the details of concrete differences in the material presented to each of the program groups.) Teaching is interactive, and much of what happens in didactic interactions depends upon the characteristics of the teacher and of the individual students, and upon the current physiological and emotional states of both teacher and taught. For this reason alone, i t i s virtu a l l y impossible to replicate the teaching of a lesson, either within a study or from one experiment to another. While r e p l i c a b i l i t y i s not a major issue in an exploratory study such the present one, i t must be noted that further investigation of the issues explored herein cannot simply rely upon attempts to replicate the anger management lessons. Future research must be designed to take into account the inherent i r r e p l i c a b i l i t y of classroom events, and to make the impact of instructor variables in anger management programs an object of investigation. Sample size. Although appropriate for an exploratory study, the small sample size limited the possibility of significant effects being revealed. It also necessitates that the results be viewed as provisional. Discrepancies in administration of repeated measures. The fact that a number of subjects were not administered the repeated measures at the designated pre-, intermediate, and posttest times contributed a degree of error to the repeated measures results. However, for each test time, the number of subjects involved was small enough to suggest that the degree of error thus introduced was not large (see Appendix J) . Nonrandom sampling. The constraints of the study did not allow the population of intermediate-age elementary school children to be randomly sampled, or the subjects to be randomly assigned to treatment groups. In fact, two of the groups were drawn from a class that had been identified by school personnel as the school's "most d i f f i c u l t . " However, the necessity of working with intact groups formed (largely) on the basis of age, allowed for the poss i b i l i t y of comparing the responses of children of different ages on the anger-related measures. As well, the necessity of working with children prone to conflict allowed for the investigation of personal and situational factors important in anger management to be carried out with just the sort of nonclinical group most likely to be identified as in need of anger management training. Finally, the exploratory nature of the study made s t r i c t generalizability unnecessary. Non-separation of researcher and experimenter. Limited resources prevented the separation of the roles of researcher and experimenter. This was not, however, as serious a limitation as i t would have been in an experimental study, since the current study advanced few hypotheses and no claim has been made as to the independence of the experimenter. n o Same instructor for each cfroup. The constraint of having the same instructor for each group, while making i t impossible to determine the impact of instructor variables such as personality or teaching style, at the same time meant that the problem of having to control for differences between instructors could be avoided. Names on paintings. A minor but time-consuming methodological shortcoming of the study was that the students were asked to put their names on the front of their paintings. Maintaining the subjects' confidentiality required the researcher to take pains to conceal the name on each of the 165 paintings prior to photographing them for rating purposes. Program Limitations The sequence of lessons that served as the context for the current investigation is inadequate in a number of respects as an anger management program. First, i t i s not a comprehensive unit. It does not even touch upon many topics in cognitive-behavioural anger management, and the topics that are mentioned are not covered in sufficient depth. Second, as they stand, the lessons include too much to be adequately addressed in the 40 minutes allotted. In the didactic component, there are too many topics, and not enough opportunity for review and practice. In the art component, 20 minutes is insufficient for many students to complete paintings, and there i s not enough time allowed for display and discussion of the pictures produced. Ideally, longer periods would be dedicated to the painting lessons, and the didactic component would consist of a greater number of - perhaps shorter - lessons. The most important of the specific content-related inadequacies of the program is the omission of an instruction to the teacher to be prepared to use any situation involving anger that arises during the lessons as a "teachable moment." Students get angry at students, students get angry at the teacher, the teacher gets angry at students - a l l of these everyday occurrences can serve as living examples of how to apply the s k i l l s and concepts presented in the program. Summary and Conclusion The present study was designed as an exploratory, descriptive investigation of some of the factors at play during the implementation of one anger management program. It took account of variability in program characteristics (i.e., component content and sequence of components) which might be associated with indicators of impact in the population sampled. It also took account of the possible di f f e r e n t i a l effects of subject variab i l i t y within the context of the anger management program, as reflected in subjects' responses on a variety of measures related to anger and anger management. A greater number of significant results of interest were obtained relating to program var i a b i l i t y than pertaining to variation in subject characteristics. Both the content and the sequence of the program components were found to be important. The results suggest that an art therapy component can be a useful complement to the teaching of cognitive-behavioural anger management techniques. They also suggest that the relative timing of these two different program elements may be of some consequence. Within an art component, the purpose of which i s to encourage the exploration and expression of personal experiences of anger, the present findings indicate that different amounts of exploration and expression may be f a c i l i t a t e d , depending on the topic assigned for the art work and the nature of the stimulus activity that precedes i t . The results of the study suggest that topics and a c t i v i t i e s that help children access bodily sensations and memories may be powerful facilitators of expression. Potentially f r u i t f u l areas for further research, then, are the impact on the expression of anger of different art topics, and the use of guided visualization as a stimulus to expression. One of the interpretations offered to explain the difference between the amount of anger expressed by the art-f i r s t group and the didactic-first group was that i t was a function of the nature of the relationship - didactic or therapeutic - f i r s t established between the instructor and the students. An implication of this conclusion i s that the two different elements of this kind of anger management program - the expressive and the cognitive - require the instructor to take two different roles. Ideally, even within the didactic component, the instructor would be able to allow and accept the verbal expression of anger and respond to i t as therapist rather than classroom manager. But the present study raises the question of whether the same person, especially someone the students already know as a teacher, can f i l l both roles. During the didactic lessons, the experimenter f e l t great tension between the need to cover the material in a certain amount of time - a normal pressure of classroom teaching - and the need to respond to the students in a manner more allowing of self-exploration and self-expression. Yet, r e a l i s t i c a l l y , anger management programs for nonclinical populations of children are of necessity presented by one person, usually the classroom teacher. An important direction for future research w i l l be to investigate further the nature of the relationship between instructor and students in the different parts of an anger management program, and explore how the components of a program can be manipulated to increase the likelihood of one person successfully f i l l i n g both roles. The importance of the relative timing of the two different components of the anger management program is revealed in the results suggesting that students need ample opportunity to explore and express their personal experiences of anger before they can be ready to consider constructive approaches to managing i t . The implications of this conclusion are, of course, that in an anger management program like that used in the current study, the art component should come before the didactic component, and that i t should incorporate more than three lessons. A further implication is that a consideration of problem-solving, either within the art component or as part of the didactic component, should not be introduced too soon. Of course further research is required to support this conclusion, as well as to ascertain what factors determine how much exploration and expression is enough, and how to maximize students' readiness to learn cognitive-behavioural anger management techniques. Though the current findings did not support the expectation that the degree of anger expressed in the paintings would increase between the beginning of the art component and the end, the small group of high scorers on state anger did give some evidence of increasing anger expression over the course of the program. It was conjectured that the process of painting helped them become more aware and more accepting of their angry feelings. If this i s the case, i t may be that an expressive approach to the exploration of anger, such as that embodied in the art component of the program used in this study, i s more effective or useful for children who, like the high scorers in this study, have high levels of state anger much of the time. If i t is true, as suggested above, that a high level of state anger can interfere with a student's a b i l i t y or motivation to engage with the cognitive-behavioural content of an anger management program, then i t may even be the case that children with elevated state anger require more opportunities than others to develop awareness and acceptance of anger through expressive act i v i t i e s , in order to attain a readiness to learn anger management s k i l l s . Further research on the differences between children who consistently score high on measures of state anger and those who do not could shed light on the question of who would benefit most from the art therapy component of an anger management program. The results of the current study provide support for the hypothesis that there is a relationship between state anger and state anxiety in children. Future research might profitably explore this association in greater depth. It would also be useful to investigate the relationship between anger and other emotions, such as sadness, self-hatred, disgust, fear, and so on, either directly through se l f -reports or other scale measures, or through analysis of art work similar to that produced in this study. Of the four areas of individual difference investigated in the study (level of maturity, abi l i t y , gender and locus of control), only the gender analyses provided results that allowed of much interpretation. Additional research, involving larger samples and aimed more precisely at exploring the effect of each kind of individual difference on anger management outcomes, is required to address these outstanding issues. Finally, an important factor not addressed in the current study is the role that cultural differences might play in how children respond in the context of anger management programs. Future research on the impact of individual differences on children's behaviour within this context should not f a i l to include this factor. This study has described some of the important factors at play in the experience of one group of students of a particular anger management program. Though the results of the study are not generalizable, i t i s safe to postulate that at least some of these factors would help to determine the course of events i f any group of elementary school students were to undergo a similar program. More research is required to explore further and eventually identify the c r i t i c a l factors which determine the success or failure of children's anger management programs. Meanwhile the present research may be of use as a guide to practitioners -teachers and counsellors - in their liv i n g classroom experiments with anger management programs. Allan, J. A. B. (1988). Inscapes of the child's world; Jungian counselling in schools and c l i n i c s . Dallas, TX: Spring. Allan, J. A. B., & Bertoia, J. (1992). Written paths to healing; Education and Jungian child counselling. Dallas, TX: Spring. Allan, J. A. B., & Nairne, J. (1984). Class discussions for teachers and counsellors in elementary school. Toronto, ONT: Guidance Centre, Faculty of Education, University of Toronto. Anderson, R.F. (1980). Using guided fantasy with children. Elementary School Guidance and Counselling. 14, 39-46. A v e r i l l , J. R. (1983). Studies on anger and aggression; Implications for theories of emotion. American Psychologist. 38, 1145-1160. Biaggio, A. M. B. (1985). Relationships between state-trait anxiety and locus of control - experimental studies with adults and children. International Journal of Behavioural Development. 8., 153-166. Camp, B. W., & Bash, M. A. (1981). Think aloud: Increasing social and cognitive s k i l l s - a problem-solving program for children. Champaign, IL: Research Press. Carroll, J. J., & Steward, M. S. (1984). The role of cognitive development in children's understanding of their own feelings. Child Development. 55, 1486-1492. Chaplin, J. P. (1968). Dictionary of Psychology (rev. ed.). New York: Dell. Committee for Children. (1989). Second step: A violence prevention program. Seattle, WA: Author. Cowan, P. A. (1982). The relationship between emotional and cognitive development. In D. Cicchetti & P. Hesse (Eds.), Emotional Development (pp. 49-81). San Francisco: Jossey-Bass. Crary, E. (1984). Kids can cooperate: A practical guide to teaching problem-solving. Seattle, WA: Parenting Press. CTB/McGraw-Hill. (1981). Test of cognitive s k i l l s . Monterey, CA: Author. Elardo, P. T., & Caldwell, B. M. (1979). The effects of an experimental social development program on children in the middle childhood period. Psychology in the Schools, 16/ 93-100. Fishman-Alschuler, C, & Alschuler, A. S. (1984). Developing healthy responses to anger: The counselor's role. Journal of Counseling and Development. 63, 26-29. Franko, D. L., Powers, T. A., Zuroff, D. C., & Moskowitz, D. S. (1985). Children and affect: Strategies for self-regulation and sex differences in sadness. American Journal of Orthopsychiatry. 55. 210-219. Frymier, J. R. (Presenter). (1980). Motivation and teaching. Program 1: The learner [Videorecording]. Bloomington, IN: Phi Delta Cappa and Indianapolis Public Schools. Goldstein, A. P., & Click, B. (1987). Aggression replacement training: A comprehensive intervention for aggressive youth. Champaign, IL: Research Press. Gordon, T. G. (1974). Teacher Effectiveness Training. New York: Wyden. Howard, G. S., & Conway, C. G. (1986). Can there be an empirical science of volitional action? American Psychologist. 41. 1241-1251. Jacobs, G. A., Phelps, M., & Rohrs, B. (1989). Assessment of anger expression in children: The Pediatric Anger Expression Scale. Personality and Individual Differences. 10. 59-65. James, W. (1981). The Principles of Psychology (Vol. 2)(F. H. Burkhardt, F. Bowers, & T. K. Skrupskelis, Eds.). Cambridge, MA: Harvard University Press. (Original work published 1890). Kalmakoff, S., & Shaw, J. (1987). Peer Conflict Resolution Through Creative Negotiation. Vancouver, BC: Public Education for Peace Society. Kazdin, A. E., Esveldt-Dawson, K., French, N. H., & Unis, A. S. (1987). Problem-solving s k i l l s training and relationship therapy in the treatment of antisocial child behavior. Journal of Consulting and C l i n i c a l Psvchology. 55, 76-85. Kelly, G. (1972). Guided fantasy as a counselling technique with youth. Journal of Counseling Psychologv. 19, 355-361. Kendall, P. C., & Braswell, L. (1982). Cognitive-behavioral self-control therapy for children: A components analysis. Journal of Consulting and Cl i n i c a l Psychology. 50, 672-689. Knight, R. G., Chisholm, B. J., Paulin, J. M., & Waal-Manning, H. J. (1988) . The Spielberger anger expression scale: Some psychometric data. British Journal of Cl i n i c a l Psychology. 27. 279-281. Kriedler, W. J. (1984). Creative conflict resolution in the classroom: More than 200 activities for keeping peace in the classroom. Glenview, XL: Scott, Foreman and Co. Lochman, J. E., Curry, J. F., Burch, P. R., & Lampron, L. B. (1984). Treatment and generalization effects of cognitive-behavioral and goal-setting interventions with aggressive boys. Journal of Consulting and Cli n i c a l Psychology. 52, 915-916. Lowen, A. (1975). Bioenergetics. New York: Coward, McCann & Geoghegan. Maccoby, E. E., & Jacklin, C. N. (1974a). The psychology of sex differences. Stanford, CA: Stanford University Press. Maccoby, E. E., & Jacklin, C. N. (1974b). Myth, real i t y and shades of gray: What we know and don't know about sex differences. Psychology Today. 8(7), 109-112. Maqsud, M. (1980). The relationship of sense of powerlessness to antisocial behavior and school achievement. The Journal of Psychology. 105. 147-150. Meichenbaum, D. (1981) . Application of Cognitive-behavior modification procedures to hyperactive children. In M. Gittelman (Ed.), Strategic interventions for hyperactive children (pp. 122-132). Armonk, NY: M. E. Sharpe. Meichenbaum, D., & Goodman, J. (1971). Training impulsive children to talk to themselves: A means of developing self-control. Journal of Abnormal Psychology. 77, 115-126. Mil l s , A. (1985). Art therapy on a residential treatment team. Journal of Child Care. 2(4), 61-71. Nannis, E. D. (1988). Cognitive-developmental differences in emotional understanding. In E. D. Nannis & P. A. Cowan (Eds.), Developmental psychopathelogy and i t s treatment (pp. 31-49). San Francisco: Jossey-Bass. Natale, M. (1978). Effect of induced elation and depression on internal-external locus of control. The Journal of Psychology. 100. 315-321. Novaco, R. W. (1985). Anger and i t s therapeutic regulation. In M. A. Chesney St R. H. Rosenman (Eds.), Anger and Hostility in Cardiovascular and Behavioral Disorders (pp. 203-226). Washington, DC: Hemisphere. Nowicki, S., Jr., & Duke, M. P. (1983). The Nowicki-Strickland life-span locus of control scales. In H. M. Lefcourt (Ed.), Research with the locus of control construct: Vol. 2. Developments and social problems (pp. 9-51). New York: Academic Press. Nowicki, S., Jr., & Strickland, B. R. (1973). A locus of control scale for children. Journal of Consulting and Clinical Psychology. 40, 148-154. Ollendick, D. G. (1979). Locus of control and anxiety as mediating variables of locus of conflict in disadvantaged youth. The Journal of Psychology. 101, 23-25. Omizo, M. M., Hershberger, J. M., & Omizo, S. A. (1988). Teaching children to cope with anger. Elementary School Guidance and Counseling^ 22, 241-245. Palormares, U. (1975). A curriculum on conflict management. San Diego, CA: Human Development Training Institute. Phares, E. J. (1976). Locus of control and personality. Morristown, NJ: General Learning Press. Rengert, L. A. (1986). Anger Management for Youth. Workshop materials presented at the Justice Institute of British Columbia, Vancouver, BC. Rotenberg, K. J. (1985). Causes, intensity, motives, and consequences of children's anger from se l f -reports. The Journal of Genetic Psychology, 146. 101-106. Rotter, J. B. (1966). Generalized expectancies for internal vs. external control of reinforcement. Psychological Monographs^ 80, 1-28. Rubin, J. A. (1984). The art of art therapy. New York: Brunner/Mazel. Saylor, C. F., Benson, B., & Einhaus, L. (1985). Evaluation of an anger management program for aggressive boys in inpatient treatment. Journal of Child and Adolescent Psychotherapy. 2, 5-15. Schmidt, F., & Friedman, A. (1983). Creative conflict solving for kids; Grades 4-9. Miami Beach, FL; Grace Contrino Abrams Peace Education Foundation. Schneider, M. (1974). Turtle technique in the classroom. Teaching Exceptional Children. 7, 22-24. Shennum, W. A. (1987). Expressive activity therapy in residential treatment; Effects on children's behavior in the treatment milieu. Child and Youth Care Quarterly. 16, 81-90. Spielberger, C. D., Johnson, E. H., Russell, S. F., Crane, R. J., Jacobs, G. A., & Worden, T. J. (1985) . The experience and expression of anger: Construction and validation of an anger expression scale. In M. A. Chesney, & R. H. Rosenman, (Eds.), Anger and ho s t i l i t y in cardiovascular and behavioral disorders (pp. 5-30). Washington, DC; Hemisphere. Stark, A. (1982). Dance-movement therapy. In L. E. Abt & I. R. Stuart (Eds,), The newer therapies: A sourcebook (pp. 308-325), New York: Van Nostrand Reinhold, Stratton P., & Hayes, N. (1988). A student's dictionary of psychology. London: Edward Arnold. Strickland, B. R. (1973). Delay of gratification and internal locus of control in children. Journal of Consulting and Clinical Psychology^ 40, 338. Temrick, P. (1986). Steps for Students. Workshop materials presented at the Justice Institute of British Columbia, Vancouver, BC. Travis, L. D,, & Cote, A, D. J. (1989), Piaget, Plato and Pedagogy. In C. Violate & A, Marini (Eds.), Child Development (pp. 97-112). Calgary: Detselig. United Nations Education (1986). Creating a caring communitv: Conflict resolution in the classroom (grades 4-6). Vancouver, BC: United Nations Association. Vogli-Phelps, V. (1985). Letting the monsters out! The Pointer. 29(3), 35-39. Wickless, C., & Kirsch, I. (1988). Cognitive correlates of anger, anxiety, and sadness. Cognitive Therapy and Research, 12, 367-377. Winefield, H. R. (1981). Anger expression in the Picture-Frustration Study under stressful conditions. Journal of Personality Assessment, 45, 370-374. Appendix A Letter to Parents of Subjects April 17, 1990 Dear Parents, Beginning on April 23, Ms. Sandy Kalmakoff, teacher and UBC counselling student, w i l l be teaching a guidance program on anger management to Mr. Egelstad's Grade 5/6 class and Mrs. Bland's Grade 7 class. The program w i l l involve such topics as: what is anger, how angry feelings are different from angry behaviour, ways to avoid inappropriate angry behaviour, and ways to communicate assertively rather than aggressively. These topics w i l l be covered through discussion, writing and role play exercises, and art a c t i v i t i e s . The lessons w i l l be about 40 minutes long, once a week for 6 weeks. The program is part of the Health and Life S k i l l s Curriculum for grades 5 to 7. To measure how the students respond to the unit, Ms. Kalmakoff w i l l be asking them to complete a number of questionnaires before, during and after the program. These w i l l include a scale measuring the children's style of anger expression, one that measures their knowledge of anger coping strategies, and one that measures their locus of control (whether they think that what they do can have an effect on how things go for them). Each student w i l l also complete a brief evaluation of each lesson and of the whole program. A l l information gathered from the students w i l l be kept confidential, and individual names w i l l be removed when the results are analyzed. A parental consent form i s attached. Parents have the right to refuse the participation of their child or to withdraw their child at any time, and alternate study arrangements would be made. If you would like further information on the program or the questionnaires before you sign and return the form, please feel free to contact Ms. Kalmakoff through the school phone number: 936-0478. We hope that participating in this program w i l l help your child to develop constructive ways of handling his or her anger. Thank you for your attention and your co-operation. I consent / I do not consent to my child, participating in the Anger Management guidance unit. Signature of Parent/Guardian Appendix B Lesson Plans for the Anger Management Program Didactic Component - Lesson 1 Objectives: 1. Students w i l l become aware of the nature of the program, and the group guidelines to be followed during the program.1 2. Students w i l l begin to feel comfortable talking about anger.^ 3. Students w i l l be able to distinguish between angry feelings and angry behaviour. 4. Students w i l l increase their awareness of the things that anger them. 5. Students w i l l become aware that most people sometimes get angry. Materials : 1. L i s t of adjectives printed on large Bristol board (see Table B-1).^ 2. Group guidelines printed on large Bristol board (see item number 3 under "Procedure" below) 3. F l i p chart paper and f e l t markers for brainstorm. 4. Copies of "Anger Buttons" worksheet, one per student (see Figure B-1). 5. Sticky-backed tags, one per student. 'These items were omitted with Groups 3 and 4, for whom the didactic component of the program came second, and who had therefore already been introduced to the program. L i s t of Adjectives for use with Introductory Exercise in Didactic Component - Lesson 1 happy gorgeous calm cool great talkative cheerful athletic neat jumpy strong nimble hopeful punctual generous graceful kind friendly tender j o l l y musical helpful stubborn active sensitive careful famous funny knowledgeable honest Figure B-1. Anger Buttons worksheet. Note. From Peer Conflict Resolution through Creative Negotiation (p. 54) by S. Kalmakoff and J. Shaw, 1987, Vancouver, BC: Public Education for Peace Society. Reprinted by permission. ANGER BUTTONS T H E S E THINGS REALLY PUSH MY BUTTON! Procedure: 1. Experimenter introduces self and the program: For the next six weeks, I'll be meeting with you once a week, and when we get together we'll be talking about -and doing a variety of activities that have to do with - the topic of anger. The work that we'll be doing together is a special part of your Health and Life Skills program.^ (1 minute) 2. Have each student in turn give his or her name, and pick one adjective that best describes them (either as they usually are or as they are right now) from the l i s t provided, or another of their choice.^ (4 minutes) 3. Introduce and explain the group guidelines: Conf identiality Raise your hand to speak Listen carefully when others are speaking No put-downs.^ (5 minutes) 4. Give the students an opportunity to ask a few brief questions about the program or about the group guidelines.' (1 minute) 'These items were omitted with Groups 3 and 4, for whom the didactic component of the program came second, and who had therefore already been introduced to the program. 5. Have the students answer the following warm-up questions with a show of hands: How many of you think everyone gets angry sometimes? How many of you think some people never get angry? How many of you get angry sometimes? (Allan & Nairne, 1984, p. 31) (1 minute) 6. Have students "brainstorm" words and phrases they think of in connection with anger. Stress the following: Don't use any names; speak of actions, not persons. If you're thinking of a person, try to say what they do rather than who they are (for now and for a l l the rest of the time we're talking about anger) Write ANGER in large letters in the middle of f l i p chart paper. Record words and phrases suggested by students in web chart format around the word ANGER, grouping together those that represent feelings, those that are impulses^ (Temrick, 1986), and those that are behaviours. Have students try to identify the categories (feelings, what you feel like doing, what you do). Have students suggest what the differences between the categories are. Address any other patterns or concepts that emerge. ^Not done in Group 1 until after the Anger Buttons exercise had started. The students found i t extremely d i f f i c u l t to adhere to this guideline. A l l groups needed many reminders. ^In Group 3 the concept of impulse was omitted in this lesson and mentioned in the following lesson. Group 4 had the most extensive treatment of this concept of a l l the groups. (10 - 14 minutes) 7. Distribute "Anger Buttons" sheet and explain: Each circle stands for an anger button - something that makes you mad. (Give a personal example: "Being interrupted - that really pushes my button!") Tell students to write in each c i r c l e one thing that makes them angry, one they would feel comfortable sharing with the group (using just enough words to convey the meaning). While they do the sheet, distribute one sticky-backed tag to each student. After 3 minutes, ask students to choose one item from their sheet and write i t on their tag. (Give a personal example.) T e l l students to put on their anger buttons (i.e., by sticking the tag to their clothes), and to get up and look at each others* buttons. (Allow some interaction.) T e l l students to stand in a group with others who have the same or similar buttons. Lead a discussion based on the following questions: How many found others with the same or almost the same button? What was it?'* How many had a button unlike anyone else's? What was it?^ How did it feel wearing your buttons? Why do you think it felt like that? What have you learned about anger? ""What was i t ? " was not asked in Group 2. (Exercise adapted from Rengert, 1986.) (18 - 25 minutes) Didactic Component - Lesson 2 Objectives: 1. Students w i l l understand the distinction between angry feelings and angry behaviour. 2. Students w i l l identify several body sensations experienced during episodes of anger. 3. Students w i l l become familiar with the anger arousal cycle ("Anger Mountain"). 4. Students w i l l become familiar with "STOP" (J. Snider, personal communication, October 6, 1989) as an anger coping strategy. 5. Students w i l l understand the concept of self-talk and w i l l identify one self-instructional statement (Meichenbaum, 1971) that w i l l help them cope better when they are angry. Materials: 1. Li s t of feelings and actions drawn from brainstorm in Lesson 1. 2. Materials for drawing Anger Mountain diagram (coloured chalk or f e l t markers, blackboard or f l i p chart paper) and/or completed diagram on large Bristol board (see Figure B-2). 3. Fli p chart paper for "STOP" and "self-talk" discussions. 4. "My Self-Talk" worksheet, one per student (see Figure B-3). Figure B-2. Diagram of Anger Mountain. Note. From Peer Conflict Resolution through Creative Negotiation (pp. 51, 52) by S. Kalmakoff and J. Shaw, 1987, Vancouver, BC: Public Education for Peace Society. Reprinted by permission. Low THINKING ABILITY Figure B-3. My Self-Talk worksheet. Note. From Peer Conflict Resolution through Creative Negotiation (p. 61) by S. Kalmakoff and J. Shaw, 1987, Vancouver, BC: Public Education for Peace Society. Reprinted by permission. MY SELF-TALK When I feel angry, I can say to myself: Procedure: 1. Last lesson you listed a lot of different words you think of when you think of anger. I've rewritten some of them in these two different l i s t s . (Show l i s t s and briefly discuss their names - "Feelings" and "Actions" - and the idea that feelings are "inside" and actions are "outside.") Sometimes getting angry gets us into trouble. Which of these two parts of anger is most likely to get us into trouble? Why? Allow students to discuss these questions brief l y . So there's a big difference between the emotional feelings that go with anger (and the things we do "inside"), and the actions we take, or the way we behave on the "outside". The feelings may be difficult to put up with, but they're really okay. It's what we do when we're angry that sometimes causes the problem or makes things worse. A bit later we'll be talking about how we can do things differently so getting angry doesn't get us into trouble. (2 minutes) 2. Right now I want to talk about the physical feelings that go with anger. Some of you may have noticed that quite often when you get angry certain changes start happening in your body. Ask the following warm-up questions (to be answered with a show of hands): How many of you find your voice gets louder when you get angry? How many of you breathe faster when you get angry? How many of you feel your heart beating faster when you get angry? How many of you feel your muscles tighten when you get angry? How many of you can feel yourself getting angry slowly and building up to real anger? How many of you find yourself at a point of being really angry almost before you know it? (5 minutes) 3. Draw and describe the anger arousal cycle (see Figure B-2) : When something pushes one of our anger buttons and makes us mad, we start climbing Anger Mountain. As soon as we start to get angry, the things we talked about before start to happen in our bodies. As we climb Anger Mountain, we can usually feel these things more and more. When we are as angry as we're going to get, we're at the top of the Mountain. Usually we are so angry here that we feel like we need to do something drastic. Lots of people really need to move because they have so much energy wound up inside their tight muscles. This is the time when we sometimes move in inappropriate ways. (Ask for an example or give one: hitting someone.) During the "calm down" stage, another peak is possible if something else upsets us. What do you think in happening in the body during the "calm down" stage? During the low, the body needs rest and feels tired, and the person may have worried or guilty feelings about what has happened. This is a good time to talk to someone who will listen sympathetically. Draw the "thinking cycle" on the diagram of Anger Mountain. For a few moments right after we start to get angry, we can often think a l i t t l e better. It's like when you are about to fall off your bicycle or walk into something. Sometimes it seems as if everything moves very slowly and you automatically make the right moves. When this happens, it's a good time to try and solve the problem. But often, there isn't time - someone pushes your button and you find yourself suddenly at the top of Anger Mountain. The higher we climb on the mountain, the more we go down in thinking ability. This is why, when we're really angry, we sometimes do things that we get into trouble for or are sorry about later. As we calm down, our ability to think clearly returns. Another good time to problem-solve is in the low after we have calmed down. (Adapted from Rengert, 1986.) Answer any questions students may have about Anger Mountain. (8 minutes) 3. Introduce "STOP" as an anger coping strategy: Up to now we've been talking about what anger is. For the rest of this lesson and the next one, we're going to be talking about what you can do about your anger. In other words, we're going to look at some ways you can take shortcuts down Anger Mountain. The first shortcut is called S-T-O-P. Write "STOP" on board or f l i p chart. This is of course a word that means something, and that word in itself can be very helpful, but it's also a kind of code word because each letter stands for something you can do when you find yourself getting angry. Ask students to suggest what each letter might stand for. Accept appropriate responses and/or introduce the following: S - stop T - count to ten O - open your eyes and look around P - ponder. (J. Snider, personal communication, October 6, 1989) Have students discuss which of their suggestions might be the most helpful things to do during an episode of anger. (10 minutes) 4. Have students think back to a particular time when they f e l t angry, or think of the kind of situation that makes them angry. Give them a few moments to think about what i t was really l i k e . Suggest that they close their eyes in order to imagine i t better. Ask them to try and remember what i s going through their minds or what they say to themselves at the moment when they are feeling angriest. (Offer a few examples: e.g., " I ' l l show him!") Write "self-talk" on the board. Define i t , or have students define i t ("things you say to yourself"). What is your self-talk like when you feel really angry? (Remind students not to refer by name to particular people in their responses.) Record students' responses, Discuss which self-talk examples would probably be unhelpful and which would be helpful, i.e., which might send a person higher up Anger Mountain and which might calm them down. Have students generate a l i s t of helpful self-talk statements that could be used as shortcuts down Anger Mountain. Examples : - I can handle this. - I do not have to hit the guy. - Stay calm, cool i t . - I'm not going to let her bug me. - Take a deep breath. - Remember STOP. - He i s wrong, but I'm not going to blow i t . - Forget i t . It's not worth being upset about. - I can stay in control. - I'm not going to climb Anger Mountain. (Adapted from Rengert, 1986.) Hand out "My Self-Talk" sheets (see Figure B-3) and have students choose and record on their sheet one self-talk statement that they think w i l l help them stay in control when they are angry. If there is time, have students volunteer to read their chosen self-talk statement to the group and explain why they chose i t . (15 minutes) 5. Recruit three students to coach for the next lesson's role-plays. Didactic Component - Lesson 3 Objectives: 1. Students w i l l learn how to take a time out as a coping strategy in situations involving anger. 2. Students w i l l be able to distinguish "you" messages from " I " messages (Gordon, 1974) (assertive statements). 3. Students w i l l be able to describe the components of an "I" message. 4. Students w i l l be able to formulate "I" messages. Materials and preparation: 1. Coach two students to participate in a role-play demonstrating how to take a time out. (See Figure B-4 for role-play script.) 2. Coach one student to participate in a role-play demonstrating "you" messages and "I" messages. (See item number 3 under "Procedure" below for role-play script.) 3. "I" messages components and formula printed on large Bristol board (see item number 3 under "Procedure" below). 4. Copies of two "I" Messages worksheets (see Figures B-5 and B-6), one each per student. Figure B-4. Script for role play from Didactic Component -Lesson 3. Note. From Peer Conflict Resolution through Creative Negotiation (p. 69) by S. Kalmakoff and J. Shaw, 1987, Vancouver, B.C.: Public Education for Peace Society. Adapted by permission. Student A: Why didn't you stay after school yesterday to work on our project like you said you would?! We've s t i l l got work to do and i t ' s supposed to be handed in on Friday! Student B: Well, I was going to, but then I really wasn't feeling well, so I ended up going home. A: That was really retarded! You didn't even t e l l me! B: Don't c a l l me retarded! I've done most of the work on this project! A: You've done most of the work! Now I've heard everything ! B: Look, I really want to solve this problem, but I'm feeling very angry and I need to take a time out. I'd like to talk about i t later, maybe after school. A: (reluctantly) Well, okay, I guess. B: Okay, see you then. Bye. Figure B-5. Fi r s t "I" Messages worksheet. Note. From Peer Conflict Resolution through Creative Negotiation (pp. 25-26) by S. Kalmakoff and J. Shaw, 1987, Vancouver, B.C.: Public Education for Peace Society. Reprinted by permission. '7" MESSAGES Make "I" messages by saying: "I feel when I . . . " (say exactly what happened). OR "I feel when my . . ." (say exactly what happened). FOR EACH STORY BELOW MAKE AN "I" MESSAGE THAT YOU WOULD SAY INSTEAD OF THE "YOU" MESSAGE GIVEN: 1. You want to use your glue. You lent it to Marsha and she hasn't returned it yet. You feel like saying: i uu always take my glue. Would you mind giving it back right now!!!? A Instead you make an "I" message and you say to her: 2. You are walking home from school. Chris comes up behind you and says, "Your hair looks really stupid, you know that?" You feel like saying: Well. YOUR hair looks like a monkey's head! Instead you make an "I" message and you say to him: I feel when 3. You bring a new game to school. Jodi takes it out for recess without asking you. You feel like saying: If you don't give me my game back right now, I'll tell the teacher on you! Instead you make an "I" message and you say: - - - ^ ^ ^ ^ F i g u r e B - 6 . S e c o n d " I " M e s s a g e s w o r k s h e e t . N o t e . F r o m P e e r C o n f l i c t R e s o l u t i o n t h r o u g h C r e a t i v e N e g o t i a t i o n ( p p . 27 -28) b y S . K a l m a k o f f a n d J . Shaw, 1987 , V a n c o u v e r , B . C . : P u b l i c E d u c a t i o n f o r P e a c e S o c i e t y . R e p r i n t e d b y p e r m i s s i o n . FOR E A C H STORY BELOW M A K E A N "I" M E S S A G E T H A T Y O U W O U L D S A Y IF IT H A P P E N E D : 1. You made a special effort to get the red ball to play with at recess. Peter tries to grab it out of your hands. You stop and think about "I" messages and then you say to him: I feel when 2. Your team has just lost the soccer game. Penny comes up to you and says, "Ha, ha. You lost, we won!" You stop and think about "I" messages and then you say to her: I feel when 4. Michael sticks his foot out every time you go past. You have just tripped for the third time today. You stop and think about "I" messages and you say to him: Procedure; 1. Review Anger Mountain and shortcuts down i t discussed so far. (3 minutes) 2. Refer to students' self-talk examples such as "I'd better walk away." Suggest that one can not only say this kind of thing to oneself but one can also do i t . Taking the action implied in this kind of self-talk is another shortcut down Anger Mountain, sometimes called taking a time out. Ask i f the students have heard this term before and i f so where. (It may be necessary to check out the students' understanding of the term and to distinguish being involuntarily "timed out" in sports or a classroom from the voluntary taking of a time out from a situation.) Have two previously coached students role play an angry confrontation which ends with one of them taking a time out. Lead a discussion of the role play based on the following questions: Why did s/he take a time out? What did the person taking the time out actually say and do? Why didn't s/he just walk away? Where on Anger Mountain might it be most helpful to take a time out?' 'Not asked in Group 3 or 4. Taking off from students' answers to the second question, go over the four steps for taking a time out: 1 - Say: "I want to solve this problem." 2 - Say: "I need to take a time out." 3 - Say: "I'd like to meet you at ." (name a time and place) 4 - Leave. Have the students role play taking a time out: Divide them into two groups, lined up facing each other, each student with a partner. Designate one line as A, the other as B. Describe the scenario the students w i l l be enacting with their partners: Both A and B are on the same basketball team, and the team has just lost a game. A is angry at B for not playing well and losing the game for the team. A goes up to B and lights into him/her. B gets angry at A for blaming him/her. After a couple of minutes or so, B decides that s/he is so angry that s/he needs to take a time out. B then does so, using the four steps. T e l l students there are two role play rules: no touching, no pretend hitting. Instruct line B to indicate when they have taken the time out by stepping back and s i t t i n g down. If there is time, when students have completed the role play, repeat, reversing the roles.* *Not done in Group 1. Briefly discuss students' experience during the role play. Ask how partner A f e l t when partner B took the time out. (12 minutes) 3. With a previously coached student, role play a "you" message and then an "I" message, instructing the students to watch for differences between them: (Set the scene: J lent a sweater to [the student] and when she gave it back to me it was torn and dirty.) "You" message — You're such a slob! You don't know how to look after anything. You're careless and inconsiderate. You really make me really mad. " I " message — I feel angry and disappointed when I lend my things and they come back messed up. Have students discuss the differences between these two messages, answering the following questions: Did the listener react differently to the two messages? If so, how? How do you think the listener felt the first time? How do you think the listener felt the second time? What was the main difference between the words used in the two messages? Which message do you think would be more helpful? Why?'^ ^Not asked in Group 1. Describe and discuss the components of an "I" message: Talk about yourself. Say how you feel. Describe the problem, not the person. Avoid blaming. Say exactly what happened and the "I" message formula: "I feel when I " OR "I feel when my. ... " Examples : "I feel hurt when I get teased." "I feel annoyed when my book isn't where I l e f t i t . " Have students work through the f i r s t " I" messages worksheet orally as a group. Have students work individually on the second "I" messages worksheet. (25 minutes) Art Component - Lesson 1 Objectives: 1. Students w i l l become aware of the nature of the program, and the group guidelines to be followed during the program.^ 2. Students w i l l begin to feel comfortable talking about anger.^ 3. Through art work and discussion students w i l l express their ideas and feelings about their own anger. 4. Students w i l l become aware of similarities between their own experiences of anger and those of others. Materials: 1. L i s t of adjectives printed on large Bristol board (see Table B-1) 2. Group guidelines printed on large Bristol board (see item number 3 under "Procedure" below) 3. Photograph of an angry person or persons, with "I feel ANGRY!" printed in a speech balloon above their heads. 4. Art supplies. Procedure: 1. Researcher introduces self and the program: For the next six weeks, I'll be meeting with you once a week, and when we get together we'll be talking about -^These items were omitted with Groups 1 and 2, for whom the art component came second, and who had therefore already been introduced to the program. and doing a variety of activities that have to do with - the topic of anger. The work that we'll be doing together is a special part of your Health and Life Skills program.^ (1 minute) 2. Have each student in turn give his or her name, and pick one adjective that best describes themselves (either as they usually are or as they are right now) from the l i s t provided, or another of their choice.^ (4 minutes) 3. Introduce and explain the group guidelines: Confidentiality Raise your hand to speak Listen carefully when others are speaking No put-downs.S Give the students an opportunity to ask a few brief questions about the program or about the group guidelines.^ (2 minutes) 4. Stress the following: During this program we'll be talking a lot about how you get angry and what you get angry about. In our discussions, don't use any names; speak of actions, not persons. If you're thinking of a person, try to say what they do rather than who they are ^These items were omitted with Groups 1 and 2, for whom the art component came second, and who had therefore already been introduced to the program. 5. Show stimulus picture of angry person or persons. Lead a discussion with the following questions: How are these people expressing their anger?^ How do you usually express your anger? What kinds of things make you angry? (5 minutes) 6. Explain the format of the next three lessons: a brief beginning activity, then painting or drawing,'" ending with sharing of art and discussion. Address students' potential fear of producing art: Sometimes art activities make some people uncomfortable, because they're afraid that others will think they can't paint. But I want to stress that no one is going to be c r i t i c a l l y judging anyone else here. This is not a competition to see who can paint the best picture. This is art that comes from inside, art that is self-expression, not performance. And since everyone is unique, every painting will be unique and can't be judged by any kind of outside standards. I do expect you to take care when you are doing your pictures, but I don't want anyone to get hung up on getting it perfect. 'In Group 4 "What's going on in this picture?" was asked before this question. '"Drawing as an alternative to painting was mentioned in every group except Group 1, but was downplayed - students were asked to bring their own equipment for i t i f they wished to draw. Add to the group guidelines:" - No coitimenting on the a r t i s t i c value of anyone's paintings. - The paint stays on the paper. (3 minutes) 7. Lead a brief guided visualization, having the students "go inside" themselves to remember and get a clear picture of the time in their l i f e they were most angry: Close your eyes....Take a few deeper breaths....Go back to a time in your l i f e when you got really angry, the angriest you've ever been. Put yourself back in that time....You can see where it all happened....You're there now, you can feel it all over again....Notice who is there....Notice what is going on....And when you have a really clear picture of it in your mind, you can open your eyes and begin to paint or draw a picture of the angriest time in your l i f e . Have students paint. T e l l them to supply a t i t l e for their picture and sign i t . As students work, circulate and discuss their paintings with them individually.'^ (20-25 minutes)'^ "The original four group guidelines were reviewed here with Groups 1 and 2. '^This was done to varying degrees in each lesson and with each group. '^Groups 1 and 2 had more time for painting than Group 3 or 4. 8. Bring students together and ask for volunteers to show and talk about their pictures. You can decide if you want others to react to your picture. Each individual is his or her own expert, the only authority on the meaning of his or her art. So when you comment on someone else's art, it's like saying what it reminds you of, and that may not have anything to do with what the person meant to express for themselves in their picture. (5-10 minutes)'" '''in Group 1 there was no time for any discussion of the paintings. Art Component - Lesson 2 Objectives; 1. Through guided visualization, art work and discussion students w i l l develop and express awareness of the body sensations they experience when angry. 2. Students w i l l become aware of similarities between their own experiences of anger and those of others. 3. Students w i l l deepen their awareness of the emotional processes they experience when angry. Materials; 1. Text for guided visualization (see Figure B-7). 2. Art supplies. Procedure ; 1. Lead guided visualization. (10 minutes) 2. Have students paint; "What I feel like in my body when I'm really angry." T e l l them to supply a t i t l e for their picture and sign i t . As students work, circulate and discuss their paintings with them individually.'' (20 minutes) 3. Bring students together and ask for volunteers to show and talk about their pictures. (10 minutes) "This was done to varying degrees in each lesson and with each group. Figure B-7. Text for guided visualization in Art Component -Lesson 2. Note. From Peer Conflict Resolution through Creative Negotiation (p. 50) by S. Kalmakoff and J. Shaw, 1987, Vancouver, BC; Public Education for Peace Society. Adapted by permission. Close your eyes. Notice the feelings around your eyes. They may be warm, tingly, relaxing sensations, or they may be normal sensations, or there may be no sensation at a l l . Whatever they are, just let those feelings around your eyes be there. Gradually let these sensations flow down, down through your face and neck and down through a l l the other parts of your body. And now I want you to become aware of your breathing. Feel the oxygen going a l l the way down into your lungs, and when i t has gone a l l the way down through your body, gradually let i t out. Feel what that feels like when you let a l l your breath a l l the way out. Let yourself relax a l i t t l e more with every out breath. And now think of a time when you f e l t really angry....Think of that time and try to picture what happened....Picture i t in slow motion. Think back to what was happening that made you feel angry. Imagine that time as i f i t i s happening right now. Pay attention to what is happening in your body as you feel yourself getting angry. Keep i t happening in slow motion so you can see everything that i s going on. Notice any tightness in your body....Notice any changes in your breathing.... See i f there are any changes in your heartbeat....Notice where in your body you feel your anger the most.... Now, picture the problem being solved, or the reason you feel angry going away. Imagine yourself getting what you want....and begin to let your anger drain away. Feel your body gradually returning to normal...calming down...feeling relaxed...as a l l your anger drains away. Slowly become aware of yourself in this room again, and when you are ready, open your eyes and get ready to paint. Art Component - Lesson 3 Objectives: 1. Students w i l l become aware of and express their positive and/or negative feelings about their own anger. 2. Students w i l l become aware of similarities between their own feelings about anger and those of others. 3. Students w i l l identify positive aspects of anger. 4. Students w i l l become aware of the concept that getting angry i s not incompatible with feeling okay about oneself. Materials: 1. Art supplies. Procedure ; 1. Have students respond to the following questions with a show of hands: How many of you have ever got angry and then later you felt bad about yourself for getting angry? How many of you have ever felt good about getting angry (or, felt glad you got angry) ? How many of you have ever felt good about the way you handled your anger? How many of you have ever solved the problem that was making you angry? How many of you have ever told the person you were angry at that you were angry and why? 2. Have a few students volunteer to t e l l the group about a personal experience of feeling good about being angry, i.e., a situation in which they f e l t good either because they were able to solve the problem, or because they were able to say clearly that they were angry and why. (5 minutes) 2. Have students paint: "A time I f e l t good about my anger." T e l l them to supply a t i t l e for their picture and sign i t . As students work, circulate and discuss their paintings with them individually.'* (25 minutes) 3. Bring students together and ask for volunteers to show and talk about their pictures. (10 minutes) '*This was done to varying degrees in each lesson and with each group. Learning Outcomes of the Anger Management Program Upon completing the program, students w i l l be able to: 1. Understand that i t ' s okay to feel angry. (1, 2, 11, 12) * 2. Distinguish between angry feelings and angry behaviour. (3, 13) 3. Identify several behaviours that would be helpful to use when they feel angry (i.e., assertive or coping behaviours), and several that would be unhelpful (i.e., aggressive behaviours). (4, 14) 4. Identify several body sensations experienced during episodes of anger. (5, 15) 5. Distinguish helpful from nonhelpful self-talk to use when angry. (6, 16) 6. Use STOP - Stop, count to Ten, Open your eyes and look around, and Ponder (J. Snider, personal communication, October 6, 1989) - and understand why i t can be helpful. (7, 17) 7. Take a "time out" without escalating the situation. (8, 18) 8. Recognize "I" messages (Gordon, 1974) and their components. (9, 10, 19, 20) *Numbers in parentheses after each learning outcome refer to the corresponding items from the Anger Management Questionnaire (see Appendix F). Concfruence of the Anger Management Program with the Intermediate Health and Life S k i l l s Curriculum The anger management program covers concepts in the following content areas: what anger is and how i t influences our lives, awareness of personal reactions in situations involving anger, the distinction between angry feelings and angry behaviour, strategies for avoiding inappropriate angry behaviour, and strategies for enhancing assertive communication. In general, the objectives are (a) to increase students• awareness and understanding of their personal emotional and behavioural patterns when they feel angry, (b) to help them develop the concepts and vocabulary that w i l l enable them to think and talk more clearly about anger, and (c) through this enhanced personal and conceptual awareness, to open the way for the learning of new, constructive responses when anger arises. The objectives of this guidance unit are thus supportive of Goal 9 in the Intermediate Health and Life S k i l l s curriculum: "To help students learn to identify emotions in themselves and others and learn acceptable ways of expressing emotions." The objectives of the program may also be seen to be directed at "the development of successful relationships beyond the family unit" (Goal 3) -and may even help "enhance family relationships" (Goal 2) -in that an individual improvement in personal anger management can be a f i r s t step towards improved interpersonal relationships of a l l kinds. Since a primary aim of the program is to foster students' understanding and acceptance of their own and others' angry feelings, i t w i l l naturally "provide a positive environment in which respect, understanding, acceptance and caring are encouraged" (Goal 4). Finally, children who often act on their angry feelings in inappropriate ways don't usually like themselves for doing so, and their self-esteem may consequently be very low. In encouraging children to accept their own angry feelings while developing alternative, more appropriate, angry behaviour, the unit should thus support Goal 6: "To enhance students' self-concept." R e l i a b i l i t y and Validity of the Pediatric Anger Expression Scale The Pediatric Anger Expression Scale (PAES) (Jacobs et a l . , 1989) i s a new instrument s t i l l in the process of being developed. It was designed on the same model as the adult version, the Anger Expression Scale (Spielberger, Johnson, Russell, Crane, Jacobs, & Worden, 1985), for which Spielberger et a l . (1985) provided evidence of concurrent and construct validity, and for which Knight, Chisholm, Paulin, and Waal-Manning (1988) provided further evidence of construct valid i t y and "satisfactory levels of r e l i a b i l i t y " (Knight et a l . , 1988, p. 279). Jacobs et a l . (1989) reported evidence of concurrent validi t y for an early version of the PAES which comprises four factors rather than three, with anger-reflection and anger-control standing as two separate factors. Correlations were calculated between the factor scores of this four-factor PAES and (a) self-ratings on the Pediatric Anxiety Scale, the Pediatric Anger Scale (Jacobs, cited in Jacobs et a l . , 1989), and the Hunter-Wolf A-B Rating Scale (HWAB) (Hunter, Wolf, Sklov, Webber, Watson, & Berenson, cited in Jacobs et a l . , 1989); (b) peer-ratings; and (c) teacher ratings of Type A behaviour, personality characteristics, and social s k i l l s . Anger-out was found to correlate positively with the self-measures of state and t r a i t anxiety and anger (with correlations ranging from .28 to .11, E < .001, for females, and correlations ranging from .24, E < '01/ to .74, E < .001, for males). Anger-out also correlated positively with the HWAB rating of Type A (r = .44, p < .001 for females, and r = .52, p < .001 for males). For males, anger-out correlated positively with peer ratings of anger (r = .32, p < .001), and teacher ratings of anxiety (r = .19, p < .05) and impatience (r = .24, p < .01), and negatively with teacher ratings of anger-in (r = -.24, p < -Ol), and s e l f -control (r = -.25, E < .01). Anger-control was found to correlate negatively with self-reports of state and t r a i t anxiety and anger (with correlations ranging from -.23, E < «01, to -.34, E < .001 for females, and correlations ranging from -.27, E < .01, to -.51, E < .001, for males). Anger-control also correlated negatively with the HWAB rating of Type A (r = -.25, p < .001 for females, and r = -.45, p < .001 for males). For both males and females anger-control correlated negatively with the teacher ratings of Type A ( r = -.20, p < .01 for females and r = -.22, p < .01 for males). For females, anger-control correlated positively with peer ratings of shyness (r =.19, p < .05), and negatively with teacher ratings of impatience (r = -.27, p < .001). Anger-reflection was found to correlate negatively with self-ratings of state and t r a i t anger and state anxiety for females (r = -.22, p < .01, to r = -.31, E < .001), and with state anxiety for males (r = -.22, E < .05). Anger-reflection also correlated negatively with the HWAB rating of Type A (r = -.22, p < .01 for females, and r = -.18, p < .05 for males). For females, anger-reflection correlated positively with peer ratings of shyness (r = .19, p < .05), and i t correlated negatively with peer ratings of anger (r = -.24, p < .01), teacher ratings of impatience (r = -.23, p < .01), and teacher ratings of anger (r = -.71, p < .05). For males, anger reflection correlated negatively with teacher ratings of impatience (r = -.21, p < .05), and Type A (r = -.23, p < .01). Anger-suppression was found to correlate negatively with self-reported t r a i t anger (r = -.19, p < .05 for females, and r = -.51, p < .01 for males). For females, anger-suppression also correlated negatively with teacher ratings of impatience (r = -.23, p < .01) and Type A (r = -.18, p < .05), and positively with peer ratings of shyness (r = .19, p < .05). In addition, Jacobs et a l . (1989) found significant item-total correlations for each of the four subscales, providing evidence of their internal consistency. "For anger-out, item-total correlations ranged from 0.44 to 0.58, with an alpha coefficient of 0.74. Anger-control item-correlations ranged from 0.47 to 0.52, with an alpha coefficient of 0.68. Anger-reflection item-total correlations ranged from 0.29 to 0.50, with an alpha coefficient of 0.63. For anger-suppression both item-total correlations were 0.50 and the alpha coefficient was 0.67" (Jacobs et a l . , 1989, p. 61). The State-Trait Anxiety Inventory for Children (STAIC) The STAIC was developed to measure state and t r a i t anxiety in fourth-, f i f t h - , and sixth-grade children. It consists of two 20-item scales. The state scale assesses the intensity of a child's worry at a given time. The t r a i t scale asks children to report how they generally feel by indicating the frequency of the behaviours described in the items (Papay & Spielberger, 1986). Papay and Spielberger (1986) reported that the STAIC was found to correlate with the Children's Manifest Anxiety Scale (Castenada, McCandless, & Palermo, 1956, cited in Papay & Spielberger, 1986) at r = .75, and with the General Anxiety Scale for Children (Sarason, Davidson, Lighthall, Waite, & Ruebush, 1960, cited in Papay & Spielberger, 1986) at r = .63. In a sample of kindergarten, and f i r s t - and second-grade children, Papay and Spielberger (1986) found a median alpha coefficient of .875 for the internal consistency of the t r a i t scale, and alphas ranging from .71 to .76 for the state scale, when administered individually. In a sample of seventh- and eighth-grade students. Cross and Huberty (1993) reported alpha coefficients of .84 and .85 for the state and t r a i t scales, respectively, "which are comparable to the original standardization sample" (p. 238) . Appendix G Anger Management Questionnaire For each statement, circ l e "T" i f you think i t i s true, and "F" i f you think i t is false. Do not spend too much time on any one item. T F 1. Anger is bad. (F)* 2. It isn't okay to feel angry, unless you keep i t to yourself. (F) T F 3. Feeling angry and acting angry are two different things. (T) 4. When someone is very angry, i t ' s better for that person to go for a run than to try and win an argument. (T) 5. When people get angry, their muscles tighten up. (T) 6. When you are angry, i t can help i f you say to yourself "I'm not going to lose my cool." (T) 7. When you are angry, i t can help i f you stop and think about what•s happening and what you are going to do. (T) 8. If two kids are so angry they're shouting at each other, i t can help for one of them to go away and come back later. (T) Jay lent a book to a friend. The friend took i t home and on the way the book f e l l in a mud puddle. When the friend returns the book, a good thing for Jay to say might be, "I feel upset when I lend my things and they get wrecked." (T) *The letter in parentheses following each item indicates the response that best corresponds with program content. 10. When someone has done something to make you mad, i t ' s better not to t e l l them how you feel. (F) F 11. Everyone gets angry sometimes. (T) F 12. People shouldn't be allowed to get angry. (F) T F 13. A person can feel very angry, and s t i l l act in ways that won't hurt anyone. (T) T F 14. When someone is very angry, i t ' s better for that person to get i t a l l out by shouting at the person they're mad at, than to just go away. (F) F 15. When people get angry, i t mostly happens in their minds, not in their bodies. (F) T F 16. When you are angry, i t can help i f you say to yourself, "I'm not going to let that old so-and-so get away with t h i s l " (F) 17. When you are angry, counting to ten can slow you down and help you decide the best thing to do. (T) 18. When you are really angry, you don't have to settle the argument right away, you can leave i t u n t i l later. (T) T F 19. You have been standing quietly in line ever since the teacher said to line up. Suddenly someone pushes in front of you. It w i l l probably make things worse i f you say, "Hey, don't be so pushy. I was here f i r s t ! " (T) T F 20. When someone has done something to make you mad, i t ' s best to t e l l them exactly what upset you. (T) Construction of the Anger Management Questionnaire fAMO) The AMQ was designed to reflect the specific learning outcomes of the anger management program (see Appendix C). Each item on the questionnaire asks a question testing the student's achievement of one of the learning outcomes (as indicated in Appendix C). The two learning outcomes deemed by the researcher prior to program implementation to be the most important ("Understand that i t ' s okay to feel angry," and "Recognize 'I' messages and their components") are weighted with four items each. A l l the other learning outcomes are represented in the questionnaire by two items each. The f i r s t half of the questionnaire (10 items) contains two items for each of the weighted outcomes, and one item each for each of the others; the second half mirrors the f i r s t . The questionnaire was designed without reference to the specific vocabulary and concepts taught in the anger management program so that i t would be possible for the students to answer the questions even prior to exposure to the program content. R e l i a b i l i t y and Validity of the Children's Nowicki-Strickland Internal-External Control Scale (CNSIE) The CNSIE (Nowicki & Strickland, 1973) i s a paper-and-pencil measure consisting of 40 questions that are answered either yes or no. Nowicki and Duke (1983) summarized researchers' reports of internal consistency and test-retest r e l i a b i l i t y for the CNSIE. Almost a l l estimates of internal consistency are above .60. "Because the CNSIE test i s additive and items may not be comparable, split-half r e l i a b i l i t i e s probably tend to underestimate the true internal consistency of the scale " (Nowicki & Duke, 1983, p. 13). Reports of test-retest r e l i a b i l i t y range from r = .52 (over 1 year) to r = .76 (over 5 weeks). Generally the test-retest r e l i a b i l i t i e s are higher over shorter periods. Nowicki and Strickland (1973) reported evidence of construct validity for the CNSIE by showing moderate relationships between this scale and other measures of locus of control. A significant correlation with the Bialer-Cromwell score (Bialer, 1961, cited in Nowicki & Strickland, 1973) was found (r = .41, p < .01). An examination of the Intellectual Achievement Responsibility Questionnaire (Crandall, Crandall & Katkovsky, 1965, cited in Nowicki & Strickland, 1973) showed significant correlations with the 1+ but not with the I- scores (for grade 7 subjects, r = .51, p < .01). In addition, Nowicki and Strickland (1973) found that CNSIE scores were not related to social desirability, and Nowicki (cited in Nowicki & Duke, 1983) found that CNSIE scores are not related to the gender of the subject. Administration of the Self-Report Measures Timing of Administration The pretests were administered a few days before the experimenter was introduced to the students and started the anger management program (Time 1). The intermediate tests were administered in between Lessons 3 and 4, on the same day for each group (Time 2). The posttests were administered in the week following the completion of Lesson 6 for a l l groups (Time 3). Due to student absences, for a few of the subjects the measures were taken at somewhat different times than for the rest of the sample. For 4 subjects - 2 from Group 1 and 2 from Group 4 - the pretests were administered after Lesson 1, not before. The intermediate tests were administered between Lessons 4 and 5 (instead of between Lessons 3 and 4) to a t o t a l of 5 subjects - 2 from Group 1, 1 from Group 2, and 2 from Group 4. Another 3 subjects, 1 each from Groups 1, 2, and 3, did not complete the posttests u n t i l after the researcher had visited the classrooms following the last lesson to explain to the students the research aspect of the anger management program. Order of Administration The measures were administered in the following order: Pediatric Anger Expression Scale (PAES) Anger Management Questionnaire (AMQ) Pediatric Personality Scale - 2 (PPS-2) (trait anger and t r a i t anxiety scales) Children's Nowicki-Strickland Internal-External Locus of Control Scale (CNSIE) Pediatric Personality Scale - 1 (PPS-1) (state anger and state anxiety scales) The CNSIE was administered only at Time 1. Instructions to Teachers Administering the Measures A. Feelings Questionnaires (PAES, PPS-1, and PPS-2): Read the instructions at the top of each questionnaire out loud to the class before they begin that questionnaire. Let them complete each questionnaire on their own. Encourage them to finish each one within 5 - 6 minutes. (That i s , i f anyone i s s t i l l working on the f i r s t questionnaire after about 4 1/2 minutes, give a warning, and then go on to read the instructions for the second questionnaire, etc.) B. Anger Management Questionnaire: Read the instructions at the top of the questionnaire out loud to the class before they begin. Let them complete the questionnaire more or less at their own rate, but encourage them to finish within 10 minutes. C. Internal-External Control Scale: Read the instructions at the top of the questionnaire out loud to the class before they begin. Read each item aloud as the students complete the questionnaire to make sure they a l l understand and to keep them working at the same pace. Read each item twice, pause briefly, then go on to the next one. This scale should take between 10 and 15 minutes to administer. If you are asked, "What should I do i f I can answer both yes and no to a question?" assure the students that this i s not an unusual happening, and t e l l them that i f i t is a l i t t l e more yes than no, answer yes; i f i t i s a l i t t l e more no than yes, answer no. Urge them to pick one or the other response and to try and answer that and a l l items (S. Nowicki Se B. Strickland, personal communication, 1990) . Scoring and Interpretation of Questionnaire Data This appendix describes the procedures used for scoring the Pediatric Anger Expression Scale (PAES) (Jacobs et a l . , 1989), the Pediatric Anger Scale (PANG), and the Pediatric Anxiety Scale (PANX) (Jacobs, cited in Jacobs et a l . , 1989); the Children's Nowicki-Strickland Internal-External Locus of Control Scale (CNSIE) (Nowicki & Strickland, 1973); and the Anger Management Questionnaire (AMQ). It also discusses anomalies in the raw data results from these self-report measures and describes how these anomalies were interpreted for scoring purposes. See Chapter V for a discussion of these anomalous responses. PAES, PANG and PANX The version of the PAES used in the current study (G. A. Jacobs, personal communication, January 15, 1990) i s a 15-item self-report t r a i t measure. This scale forms three factors: anger-out, anger-suppression, and anger-reflection/control. The subscale for each of these factors consists of five items. For each item, subjects are instructed to rate the frequency with which they use the specified manner of expression when they are angry or very angry on the following 3-point scale: (a) hardly-ever. (b) sometimes. (c) often. Numerical values are assigned to the responses as follows: hardly-ever = 1, sometimes = 2, often = 3. Individual scale scores are calculated by summing the numerical values for responses on the items that comprise the scale (G. A. Jacobs, personal communication, January 15. 1990). The PANG and PANX (Jacobs, cited in Jacobs et a l . , 1989) measure anger and anxiety as emotional states and personality t r a i t s . Items from the two scales are alternated on a single form, the Pediatric Personality Scale (PPS), which i s divided into two parts: PPS-1 and PPS-2. Form PPS-1 comprises the state anger and state anxiety scales. Form PPS-2 comprises the t r a i t anger and t r a i t anxiety scales. The state scales ask subjects to assess the intensity of emotion experienced at the moment of responding to each item by rating themselves on the following 3-point scale: (a) very much so. (b) somewhat, (c) not at a l l . Numerical values are assigned to the responses as follows: very much so = 3, somewhat = 2, not at a l l = 1. Scoring for several items on the state anxiety scale are reversed before summing the scale scores. Individual scale scores are calculated by summing the numerical values for responses on the items which comprise the scale (G. A. Jacobs, personal communication, January 15. 1990). The t r a i t scales require subjects to report the frequency with which they experience the specified personality characteristics by rating themselves on the following 3-point scale: (a) hardly-ever. (b) sometimes. (c) often. Numerical values are assigned to the responses as follows: hardly-ever = 1, sometimes = 2, often = 3. Individual scale scores are calculated in the same manner as for the state scales, noted above (G. A. Jacobs, personal communication, January 15. 1990). PAES Anomalies The raw data results from the PAES yielded several kinds of anomalies. The types and frequency of these anomalous responses are itemized below and summarized in Table J-1. 1. Two or more points marked on an item rating scale. One subject marked two adjacent points on the rating scale for one anger-out item on the intermediate test. Another subject marked the two points at the opposite extremes of the scale for one anger-suppression item on the same test. 2. A l l items marked with the same response. One subject responded with sometimes to every item on the intermediate test of the PAES, affecting the scales for a l l three factors. Another subject did the same on the posttest questionnaire. A third subject responded with sometimes to a l l posttest items but one, and a fourth subject marked every posttest item hardly-ever. 3. Items omitted. On the PAES pretest, one subject fai l e d to give a response to one anger-reflection/control item, and another subject omitted an anger-out item. On the intermediate test, one subject did not respond to one anger-reflection/control item. Anomalies in Subject Responses on PAES, PANG, and PANX Type and frequency of anomaly Variable 1 2 3 4 5 Total ANGOUT Time 1 1 1 Time 2 1 1 2 Time 3 3 3 ANGSUPP Time 1 0 Time 2 1 1 2 Time 3 3 3 ANGRXC Time 1 1 1 Time 2 1 1 2 Time 3 3 3 STANX Time 1 1 1 2 4 Time 2 1 1 2 Time 3 2 1 3 STANG Time 1 1 1 2 Time 2 1 1 (table continues) Table J-1 (continued) Type and frequency of anomaly Variable 1 2 3 4 5 Total STANG Time 3 1 1 1 3 TRANX Time 1 2 2 Time 2 1 1 1 3 Time 3 4 4 1 1 10 TRANG Time 1 1 1 Time 2 1 1 Time 3 1 4 1 6 Totals 13 22 10 6 4 55 Note. Types of anomalies are coded as follows: 1 = two or more points marked on an item rating scale; 2 = a l l items marked with the same response; 3 = items omitted; 4 = page missed; 5 = apparent inter-item discrepancies. The frequency of the anomaly refers to the number of subjects' questionnaires affected by i t . See Table 3 for explanation of variable abbreviations. Anomalies on the State Scales The raw data results from Form PPS-1 (the state scales) yielded the following anomalies (see Table J-1): 1. Two or more points marked on an item rating scale. On the state anxiety scale, one subject marked both very much so and not at a l l (opposite extremes on the scale) for one item on the pretest and one item on the posttest. The same subject marked adjacent points on the scale (e.g., somewhat and not at all) once on the pretest and twice on the posttest. In addition, this subject marked two adjacent responses for one item on the state anger posttest, and marked a l l three points on the rating scale for another item on the same test. Two other subjects marked adjacent responses once each on the state anxiety scale, one on the intermediate test and one on the posttest. One other subject did the same for one item of the state anger posttest. 2. A l l items marked with the same response. One subject marked not at a l l for every item on the intermediate test of Form PPS-1, affecting both the state anger and state anxiety scales. 3. Items omitted. On the state anger scale, one subject f a i l e d to mark any response for one item on the pretest, and another subject did the same for one item on the posttest. 4. Page missed. One subject failed to give responses for the last 8 (of a total of 19) items on Form PPS-1, pretest, apparently missing or forgetting the second page of the questionnaire. 5. Apparent inter-item discrepancies. On the state anxiety scale, pretest, two subjects gave apparently discrepant responses on two different items each (e.g., marking very much so for "I feel happy" and not at a l l for "I feel good.") Another subject consistently gave apparently discrepant responses a l l the way through Form PPS-1 posttest (affecting both the state anger and state anxiety scales). Anomalies on the Trait Scales The raw data results from Form PPS-2 (the t r a i t scales) yielded the following anomalies (see Table J-1): 1. Two or more points marked on an item rating scale. On the t r a i t anxiety scale, intermediate test, one subject marked two adjacent points on an item rating scale. Four other subjects did the same on the posttest, two of them twice each. One of the latter subjects also marked adjacent responses on the rating scale for one item of the posttest of the t r a i t anger scale. 2. A l l items marked with the same response. Three subjects consistently gave the same response a l l the way through the posttest of Form PPS-2, affecting both the t r a i t anger and t r a i t anxiety scales. Another subject gave a l l the same responses except one, on the posttest. A f i f t h subject gave a l l the same responses on the f i r s t page of the posttest questionnaire only (on 10 out of 20 items). None of the aforementioned subjects was the same subject as the one who responded in this way on Form PPS-1. 3. Items omitted. On the t r a i t anxiety scale, pretest, two subjects failed to mark any response for one item each. One subject did the same on the t r a i t anxiety intermediate test, and another on the posttest. A different subject omitted an item in this manner on the t r a i t anger pretest. 4. Page missed. On the intermediate-test questionnaire for Form PPS-2, one subject omitted the last 10 of 20 items, apparently missing or forgetting the second page of the questionnaire. Another subject did the same on the posttest. These subjects effectively omitted 5 of 10 items on each of the t r a i t scales. Scoring of Anomalous Responses on PAES. PANG, and PANX The scoring directions for PAES, PPS-1 and PPS-2 did not include guidelines for scoring anomalous results. For scoring purposes, the anomalies described above were interpreted in the following ways: 1. Two or more points marked on an item rating scale. When the responses marked were adj acent (i.e., having numerical values of 1 and 2, or 2 and 3), their values were interpolated to yield one value (1.5 or 2.5), in order to reflect the assumed intention on the part of the subject at least to re s t r i c t the response to one side of the scale. The alternative procedure, to assign a value of 0 to items thus marked, would have carried the score beyond one of the extremes of the scale (i.e., 1), which would have been a less accurate interpretation of the subject's intention. Similarly, when the two points at the opposite extremes of the rating scale were marked (1 and 3), or when a l l three points were marked (1, 2, and 3), i t was assumed that assigning the intermediary score would tend to skew the results less than going beyond one of the extremes by assigning a score of 0 or 4. 2. A l l items marked with the same response. These were interpreted as suspected cases of "malicious compliance" (L. D. Travis, personal communication, 1993). The resulting scores were nevertheless included in the analysis at face value. 3. Items omitted. Single items not marked with any response were interpreted as omitted and assigned a value of 0. Although this procedure may tend to skew the results in one direction, the alternative of assigning the midpoint value was rejected as untenable inference. 4. Page missed. Assigning a value of 0 in cases where responses were missing for up to half of the items would have tended to distort the results unacceptably. Hence, for such cases, the data from the incomplete questionnaires were omitted from the analysis. 5. Apparent inter-item discrepancies. These items were scored at face value. Children's Nowicki-Strickland Internal External Control Scale (CNSIE) The CNSIE (Nowicki & Strickland, 1973) was scored according to the standard directions: The score i s the total number of external responses a subject chooses. To the l e f t of each item [on the scoring key] there i s either a "Y" or "N." That i s the external response for that item. Each time the subject responds in agreement with the keyed "Y" or "N" next to the item, i t i s considered to be one external response. The score i s the total number of agreements with the keyed "Y" or "N." Scores can range from 0 to 40. The higher the score, the more external the score (S. Nowicki & B. Strickland, personal communication, 1990). Ten subjects responded ambiguously on a total of 15 items (by c i r c l i n g both responses or neither, and/or writing in "sometimes" or "maybe"). Ambiguities were deemed not to agree with the keyed responses, and were therefore not counted as t a l l i a b l e contributions to the score. Anger Management Questionnaire (AMQ) The AMQ i s a 20-item true/false questionnaire designed to assess subjects' knowledge of anger coping strategies as taught during the didactic component of the anger management program. One point was scored for each item correctly marked as true (T) or false (F), for a possible total of 20. AMQ Anomalies The raw data results from the AMQ yielded the following anomalies: 1. Neither T nor F marked. On the pretest, one subject responded this way to one item. On the intermediate test, two subjects marked one item each neither T nor F, and another subject gave this response to three items. On the posttest, five subjects marked one item each neither T nor F (one of them drawing a two-headed arrow between the T and the F), and two other subjects gave this response to two items each. There was a total of 14 such responses. 2. Neither T nor F marked, a question mark added. On the pretest, one subject marked two items with a question mark only. On the intermediate test, the same subject responded in the same manner to two different items, and another subject also marked two items (different again) with a question mark only. On the posttest, the latter subject marked four items with a question mark only, and another subject (the same as the one who marked two items with a question mark on the pretest) responded to one item in this way (one of the same items as on the pretest). There was a total of 11 such responses. 3. Both T and F marked. On the pretest, one subject marked both T and F for one item. On the intermediate test, eight subjects marked both T and F for one item each. A ninth subject responded this way to two items. On the posttest, four subjects marked both T and F for one item each, and two additional subjects did the same for two items each. There was a total of 19 such responses. 4. Other ambiguity. On the intermediate test, one subject responded to one item by writing "maybe" beside the T, and marking both the T and the F. 5. Page missed. On the posttest, one subject omitted the last 10 of 20 items, apparently missing or forgetting the second page of the questionnaire. Scoring of Anomalous Responses on the AMO Each occurrence of the anomalous responses numbered 1 to 4, above, was assigned a value of 0 for scoring purposes, in order to best reflect the subjects' knowledge of program content. For item 5, above, assigning a value of 0 where responses were missing for half of the items would have tended to skew the results unacceptably. Hence the data from the one incomplete questionnaire were omitted from the analysis. Development of the Rating Scales for Subjects' Paintings The rating scales were developed after three raters (an ar t i s t , an actor and a health educator) - different individuals than those assisting in the actual study -viewed 102 of the slides taken of the paintings. In this p i l o t rating, the paintings were rated once for degree of anger on a 5-point rating scale: (a) masked (i.e., disguised or defended), (b) none, (c) mild. (d) moderate. (e) extreme. As a result of the pilot i t was concluded that disguised anger cannot be conceived of as one among several degrees of anger, defensiveness against anger being conceptually different from anger i t s e l f . In addition, i t was found that the distinction between mild and moderate degrees of anger was not clear enough to be useful. Hence, two scales, in a compressed form, were developed. The two raters who used the scales in their f i n a l form in the research ratings were both elementary school counsellors from a suburban school d i s t r i c t - a different d i s t r i c t from that from which the study sample was drawn. Both raters had had some experience in art interpretation in a counselling context. Instructions Read to Raters of Subjects' Paintings I'm interested in anger in children, and my study takes a look at that topic from a variety of perspectives, one of these being the expression of anger in artwork, specifically painting. I w i l l be showing you 165 slides of children's paintings. I w i l l be showing you the entire set of slides twice. The f i r s t time, you w i l l use the Degree of Anger rating scale. The numbers in the left-hand column are the slide numbers. As each slide appears on the screen, I w i l l read out the slide number and the t i t l e of the painting. You w i l l have a maximum of 15 seconds to rate the picture on the degree of anger i t expresses, that i s , to answer the question "Does the depictor give evidence of an angry state, and i f so to what degree?" (none, moderate, or extreme). Before we begin I w i l l show you samples of pictures from each category to give some guidance to your necessarily subjective definition of these categories. In some paintings the t i t l e i s clear and obvious, in others i t i s barely legible or does not appear on the face of the painting. I w i l l be reading out a l l the t i t l e s because I would like you to take the t i t l e into account in your rating of each painting. The slide numbers appear on the slides as well as on your rating sheets. I ' l l read them out for c l a r i t y . In a few cases, the slide number on the slide does not match that on the rating sheet. In a l l such cases, the number on the rating sheet i s correct, so you won't need to change anything; I ' l l point out the cases where there are discrepancies just to avoid confusion. You may not need as much as 15 seconds for some of the slides. Please say "ready" i f you are ready to move on to the next slide before the f u l l 15 seconds have elapsed. When you have both said "ready" I w i l l bring up the next slide. There w i l l be two breaks of 5-10 minutes each, each time we go through the slides (in order to counteract the effects of fatigue.) Where we'll take the breaks i s indicated on your rating sheets. If you find yourself fatiguing much before a scheduled break, let me know and we'll stop sooner. Please don't ask any questions about the study or make comments about the paintings until we're finished with both run-throughs. Are there any questions about the procedure? (Slides demonstrating the ratings categories were shown at this point.) (The entire set of slides was then shown the f i r s t time.) (Next the raters were given the Degree of Defensiveness rating sheet.) This time you view the slides you w i l l be looking for evidence that the depictor has attempted to disguise an angry state or inhibit i t s expression; in other words, evidence that the ar t i s t is experiencing or has experienced anger (in relation to the subject of the painting) but for whatever reason is not comfortable overtly expressing i t . I'm calling this a defensive reaction. On the rating sheet the shorthand for i t is "defensiveness." So you'll be rating each painting for degree of defensiveness. None means that there is no evidence of a defensive reaction. Moderate means there i s some evidence of defensiveness, and extreme means an extremely defensive reaction, extreme defensiveness. Are there any questions? (The entire set of slides was then shown for the second time.) (The raters were then debriefed as to the specific nature of the study and the purpose of the painting ratings.) 191Appendix NPaintings of High Scorers on State AngerSubject A, Painting 1: I’m SadSubject A, Painting 2: Mold192Subject A, Painting 3: Ow! Ouch! Insult!193Subject B, Painting 1: The LibrarySubject B, Painting 2: The Boring Weekend194Subject B, Painting 3: The Bedroom Window195Subject C, Painting 1: Boy with BallSubject C, Painting 2a: The Rage196• - : •‘.•,.s • 4’,’.- 9 , I____•:.r•-- .‘‘-.# b‘a’-• ‘-. f,iIl••;-.•--..• \ •••u•. V...’4-- JW?_,•.m1.• •;; . •. ..;4•” cI 41Subject C, Painting 2b: Helping ConfusionSubject C, Painting 3: Death by Medicine Ball

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0054036/manifest

Comment

Related Items