UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The depressive self-schema : its relationship to the anxiety self-schema and to changes in depressed… Cheung, Elsie 1987

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

Item Metadata

Download

Media
831-UBC_1987_A8 C43.pdf [ 7.36MB ]
Metadata
JSON: 831-1.0097257.json
JSON-LD: 831-1.0097257-ld.json
RDF/XML (Pretty): 831-1.0097257-rdf.xml
RDF/JSON: 831-1.0097257-rdf.json
Turtle: 831-1.0097257-turtle.txt
N-Triples: 831-1.0097257-rdf-ntriples.txt
Original Record: 831-1.0097257-source.json
Full Text
831-1.0097257-fulltext.txt
Citation
831-1.0097257.ris

Full Text

THE DEPRESSIVE SELF-SCHEMA: ITS RELATIONSHIP TO THE ANXIETY SELF-SCHEMA AND TO CHANGES IN DEPRESSED MOOD by ELSIE CHEUNG B.A. , The University of British Columbia, 1982 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS in THE FACULTY OF GRADUATE STUDIES Department of Psychology We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA FEBRUARY, 1987 0 Elsie Cheung, 1987 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the University of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y available for reference and study. I further agree that permission for extensive copying of t h i s thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. I t i s understood that copying or publication of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of The University of B r i t i s h Columbia 1956 Main Mall Vancouver, Canada V6T 1Y3 /R-n Abstract The self-referent encoding task (SRET) was employed to investigate the content and stability of the depressive self-schema. Two central hypotheses were proposed: the depressive self-schema is characterized by anxiety, as well as, depressive content; and certain aspects of the depressive self-schema remain stable across time and across remitted depressed mood. Study 1 ensured the appropriateness of the stimulus adjectives for the SRET. Forty-five undergraduates rated adjectives on the properties of anxiety, depression, emotional intensity, social desirability, and imagery. These adjectives formed four word conditions: anxiety-content, depression-content, negative-content, and positive-content. Multivariate analysis suggested that these four word conditions were roughly equated for word length, word frequency, and ratings of imagery, and emotional intensity, but were distinguishable with respect to their ratings of anxiety, depression, and social desirability. Study 2 employed the SRET as a potential method of documenting the relationship between anxiety and depression. Twenty-two moderately depressed, 40 mildly depressed, and 26 nondepressed undergraduates rated randomly presented sets of adjectives (described in Study 1) as to their self-descriptiveness. Both the types of self-descriptors and the rating times were recorded for this SRET. This task was followed by an intentional recall task for the adjectives, and the number of recalled words per adjective set was recorded. Multivariate analyses suggested that the nondepressed subjects showed a schematic bias in their processing of the positive-content words. Although the two depressed groups showed a bias in the processing of the depression-content words, this bias was not found in the processing of the anxiety-content words. These results were discussed in terms of the utility of i i the SRET for the domain of anxiety. Study 3 asked which aspects of depresssive schematic processing remain constant across time and across remitted depressed mood. Subjects from Study 2 were reassessed approximately three months after they had first completed the SRET. Among nondepressed subjects who remained nondepressed and depressed subjects who remained depressed, highly consistent schematic processing was observed across the two testing sessions. Among subjects who were depressed at the first testing session but were no longer depressed at the second testing session, these subjects no longer rated the depression-content adjectives as self-descriptive. However their decision latencies for the four groups of adjectives remained stable across the two testing sessions. Implications of these results were discussed in terms of cognitive structures as causal markers in depression. i i i Table of Contents Page Title page i Abstract i i Table of contents iv List of tables v i i List of figures ix List of appendices x Acknowledgements xii Introduction 1 A cognitive model of depression 1 Self-schemata theory and research 5 The depressive self-schema 10 The SRET as a cognitive assessment device 17 The relationship between depression and anxiety 18 Statement of first problem 27 Cognitions as causal entities in depression 27 Statement of second problem 32 Overview of studies 32 Study 1: Normative ratings of 62 adjectives on the 34 properties of anxiety, depression, social desirability, emotional intensity, and imagery. Method 35 Results 36 Study 2: Is the depression self-schema predictive of 45 the anxiety self-schema? i v Method 47 Results 54 Discussion 69 Study 3: The stability of the depression self-schema across 76 time and across remitted depressed mood Method 78 Results - Time 1 data 80 Discussion - Time 1 data 89 Results - Time 1-Time 2 data 90 Discussion - Time 1-Time 2 data 93 The Utility of the SRET as a cognitive assessment device 96 References 98 Appendix A 110 Appendix B 116 Appendix C 119 Appendix D 122 Appendix E 128 Appendix F 131" Appendix G 140 Appendix H 142 Appendix I 144 Appendix J 146 Appendix K 148 Appendix L 150 Appendix M 152 Appendix N 154 Appendix 0 156 v Appendix P -^ 58 Appendix Q 2g0 Appendix R 6^2 Appendix S 2g4 Appendix T jgg Appendix U 2gg Appendix V 1 ? n Appendix W 2?2 Appendix X 274 v i List of Tables Number Description Page 1 Anxiety-content adjectives. 37 2 Depression-content adjectives. 38 3 Negative-content adjectives. 39 4 Positive-content adjectives. 40 5 Attributes of the four content conditions. 41 6 Classification results from the discriminant 43 analysis using the ratings of social desirability as the discriminating variable. 7 Classification results from the discriminant 46 analysis using the ratings of anxiety and depression as the discriminating variable. 8 Group means for the BDI, DAS, STAI-state, and 55 STAI-trait for Study 2. 9 Group proportions for the endorsement rate 57 index for Study 2. 10 Group averages for the descriptiveness ratings 58 for Study 2. 11 Group reaction time averages for Study 2. 59 12 Group proportions for the recall index 60 for Study 2. 13 Adjusted recall proportions for Study 2 66 14 Non-parametric analysis for the recall data 66 for Study 2. 15 Subject classification system for Study 3. 79 16 Group means for the BDI, DAS, STAI-state, 81 and STAI-trait for Study 3. 17 Group proportions for the endorsement rate 86 rate index for Study 3. 18 Group averaged descriptiveness ratings for 86 Study 3 19 Group averaged reaction time for Study 3. 87 v i i Group proportions for the recall index for Study 3. v i i i List of Figures Number Description Page 1 A plot of the four content conditions using 44 discriminant scores obtained from anxiety and depression ratings.. 2 Ratings times of the yes-responses for the 64 four content conditions. 3 Adjusted recall data as a function of levels 68 of endorsement rate for the depression-content condition. 4 Adjusted recall data as a function of levels 68 of endorsement rate for the positive-content condition. ix List of Appendices Page Appendix A Ratings of Adjectives Questionnaire 110 Appendix B Groupwise multiple comparisons for 116 each attribute of the four content conditions. Appendix C Beck Depression Inventory 119 Appendix D Dysfunctional Attitude Scale 122 Appendix E State-Trait Anxiety Scale 128 Appendix F Program listing for the self-referent 131 encoding task Appendix G 95% confidence intervals for BDI, DAS, 140 STAI-state, and STAI-trait score for Study 2. Appendix H ANOVA summary table for the endorsement 142 rate index for Study 2. Appendix I ANOVA summary table for the descriptiveness 144 ratings index for Study 2. Appendix J ANOVA summary table for the reaction time 146 time index for Study 2. Appendix K ANOVA summary table for the recall index 148 for Study 2. Appendix L ANOVA summary table for BDI scores for 150 Study 3. Appendix M ANOVA summary table for DAS scores for 152 Study 3. Appendix N ANOVA summary table for STAI-state scores 154 for Study 3. Appendix 0 ANOVA summary table for STAI-trait scores 156 for Study 3. Appendix P ANOVA summary table for the endorsement rate 158 index for Study 3 - Time 1. Appendix Q ANOVA summary table for the descriptiveness 160 ratings index for Study 3 - Time 1. x Appendix R ANOVA summary table for the reaction time 162 index for Study 3 - Time 1. Appendix S ANOVA summary table for the recall index 164 for Study 3 - Time 1. Appendix T MANOVA summary table for Study 3. 166 Appendix U ANOVA summary table for the endorsement rate 168 index for Study 3. Appendix V ANOVA summary table for the descriptiveness 170 ratings index for Study 3. Appendix W ANOVA summary table for the reaction time 172 index for Study 3. Appendix X ANOVA summary table for the recall index 174 for Study 3. x i Acknowledgements A number of individuals provided invaluable assistance in the completion of this thesis. I would like to thank Jackie DiGeso for her assistance in data coding. Eric Eich and members of my committee, which included Jennifer Campbell and Dimitri Papageorgis, provided thoughtful comments. Finally, I am greatly indebted to my research supervisor, Keith Dobson, who provided productive discussions, helpful critical comments, and generous encouragement and moral support. xii 1 Introduction A cognitive model of depression Judging by the phethora of publications, cognitive models appear to be one of the most influential and rapidly growing approaches to depression. Although these approaches vary in their implication of cognitive factors, the common premise is that the cognitive system is closely linked to the affective and behavioral systems, and that cognitive factors can serve as potent agents in mediating change in the other two systems (for a review, see Dobson & Block, in press). An exemplar of such cognitive approaches is Beck's model of depression (Beck, 1974; Beck, Rush, Shaw, & Emery, 1979). Through clinical observation, Beck identified certain tendencies or idiosyncracies in depressed individuals' thoughts and. styles of thinking. In his first studies, Beck (Beck & Hurvich, 1959; Beck & Ward, 1961) identified, the theme of loss in depressed patients' dreams. These dreams often depicted the depressed patient as a "loser" with the patient losing his self-esteem, a loved one, or some desired goal. The depressed individual is often protrayed as impotent and incapable, of achieving his goals. In later investigations, Beck (1963, 1964) observed that depressed patients' thoughts, are characteristically and consistently self-depreciative and overly pessimistic. These negative evaluations are often-not substantiated by others and often stand, in stark contrast to reality. The- content of depressive thinking can be summarized by the cognitive triad (Beck, 1976; Beck, Rush, Shaw, & Emery, 1979). The first component of the triad involves the depressed individuals' view of themselves. They see themselves as defective, inadequate, and unworthy. The second component involves the depressed individuals' view of the world. The depressed individuals tend to interpret their experiences in a. negative manner even when more plausible explanations are 2 available. They see the world as making continuous unreasonable demands on themselves. The third component of the triad involves the depressed individuals' pessimistic outlook on the future. They expect their problems to continue indefinitely. When they do attempt to alleviate their depression, they often predict failure. The cognitive model postulates that this triad is self-referential: depressed individuals often can make accurate predictions about others. Beck (1976) maintained that these depressive cognitions form the core of depressive symptomatology. It is the depressed individuals' beliefs and predictions that lead to the other affective, motivational, and behavioral aspects of depression. For example, Beck: maintained that, the- sadness is a direct result of the depressed; individuals' appraisal of themselves and; their situation. Since* the depressed individuals often predict failure, they become lethargic. Since, the depressed individuals believe that they lack the personal resources to overcome their depression, they rarely take the initiative to alleviate their depression. The vegetative signs, of. depression - loss of appetite, loss of libido, and sleep disturbance - are seen to be physiological concommitants of the depressed individuals' cognitions.. These depressive cognitions are. often not substantiated by more objective evidence and stand steadfast to refutation. Beck (1967) postulated that depressives make systematic errors in their thinking that allow them to maintain the belief in their evaluations despite contrary evidence. These errors- are: Arbitrary inference - process of drawing: conclusions based on partial evidence or despite contrary evidence. Selective* abstraction - process of focusing on selective features of the situation despite the presence of perhaps more salient features. Overgeneralization - process* of applying, a- specific rule or conclusion to a wide 3 range- of situations despite the inapplicability of that rule. Magnificaion and minimization - process by which certain (usually negative) features are magnified or distorted, while certain (usually positive) features are downplayed. Personalization - refers to the depressed individuals' tendency to relate external events to themselves. Absolutistic, dichotomous thinking - tendency to make categorical evaluations rather than relativistic evaluations. For example, the depressed individuals, might appraise their accomplishments by the categories "perfect" and "failed", rather than on. a more- adaptive continuum. Together; the content and process of depressive thinking play * primary role in the predisposition, maintenance and eventual alleviation of depression. The mechanism for this depressogenic catalyst is the schemata. Schemata form the basis for "molding data into cognition" (Beck, Rush, Shaw, & Emery, 1979, p. 12). They function like templates in screening out, differentiating;, coding, and, integrating environmental stimuli. Schemata are theorized to be enduring: structures, whose genesis lie in early experiences. From their inception, schemata can; lie dormant until they are activated, by specific environmental stressors whose form resembles the initial experiences that were responsible for embedding the schemata. For example, a marital separation may precipitate a depressive episode and may be analogous to the early loss, felt by the death of a. parent. Once activated, these depressive schemata, can be evoked by an increasingly wider range of stimuli until they become extensive; as seen in the cognitive triad, and involuntary, as seen by the depressives' "automatic thoughts!*. These readily accessible depressive cognitions serve to reinforce- the depressed individuals' beliefs and predictions. This, in- turn, interacts in a reciprocal feedback loop with depressive 4 affect and other symptoms of depression until the depressed individual is caught in a continuous downward spiral of depression. These depressive cognitions eventually become so global that they form a depressive paradigm. Similar to Kuhn's (1970) concept of a scientific paradigm, the depressive paradigm forces a distinctive interpretation on the data while limiting the awareness of alternative explanations. It is in this manner that the depressed individuals' negative cognitions and beliefs appear to themselves to be veridical representations of reality while they may appear to be farfetched to others or even to themselves when they are not depressed. Furthermore, when the depressed individual's "personal paradigm- is reversed and realigned with reality (a kind of 'counter-revolution') his depression starts- to disappear" (Beck, Rush, Shaw, & Emery, 1979, p. 21). The process of "counter-revolution" specifically involves the identification and modification of depressogenic thoughts and assumptions (see- Beck, Rush, Shaw, & Emery, 1979). As mentioned earlier, cognitive approaches represent one of the most influential approaches in the study and treatment of depression. There- are several reasons, for their appeal and popularity: First of all, cognitive depression models are consonant with the Zeitgeist of general psychology instigated by the "cognitive revolution". Secondly, cognitive approaches hold much intuitive appeal, especially for the practicing clinician. But perhaps more- importantly, cognitive- models lack the superficiality of behavioral approaches in looking, "beyond the behavior", while avoiding the mysticism of unconscious motives found in psychoanalytic approaches. In their- relatively explicit theoretical formulations and objectifiable constructs, cognitive theories lend themselves- easily to empirical, verification. In the research literature, much research has supported, cognitive theories. In the laboratory, depressive- cognitions: are measured as biases in perception and recall 5 of performance feedback, perception and recall of interpersonal feedback, and self-report of cognitions. A review of this literature is beyond the scope of this thesis and the interested reader is referred to Coyne and Gotlib's (1983) thorough and critical review. In the therapy outcome literature, the relevance of cognitions is aptly illustrated by the efficacy of Cognitive Therapy over Behavior Therapy (Shaw, 1977) and standard pharmacotherapy (Rush, Beck, Kovacs, & Hollon, 1977). However, support for the cognitive model is not unanimous with some researchers holding opposing viewpoints (e.g. Alloy & Abramson, 1979; Coyne, 1976; Lewinsohn, Mischel, Chaplin, & Barton, 1980). As with the other constructs i n cognitive depression theory, the construct of schemata lends itself to empirical study. Self-schemata theory and research Independent of the cognitive- theory of depression, Markus (Markus, 1977; Markus & Sentis, 1982) developed schema theory which implicates the self as the driving force in the- processing of social information. Schema theory represents the fruition of two lines of research. In the cognitive literature, researchers have-addressed the issue of how coherent cognitions are derived from, the vast and confusing; array of environmental stimuli. Cognitive structures were implicated which selectively attend to, encode, and: integrate the incoming stimuli, and form a framework for representing; that information. Some recent versions of these cognitive structures are frames (Minsky, 1975) and scripts (Schank & Abelson, 1977). In the social psychological literature, numerous theories have implicated the self: as a cognitive structure (see e.g. Epstein. 1973) with some explicitly exploring the role of schemata in. social, information processing: (e.g. Abelson, 1976; Cantor & Mischel, 1979). • 6 Markus conceptualized the self-schemata as the mechanism for perceiving and understanding one's social experiences. Self-schemata are defined as cognitive generalizations about the self for a particular behavioral domain. They are constructed from repeated and consistent evaluations by oneself and by others for that particular behavioral domain. For example, a person who avoids eye contact, feels inept in social situations, and is told by others that he is "quiet" may develop a self-schema of "shyness". These self-schemata are considered to be stable and enduring, in nature. These self-schemata serve to form a framework for selective processing of self-related information. The schema-based processes have been summarized into four categories: selection of self-relevant information from the environment, abstraction: of the meaning of this information, integration of this-meaning, into an already existing knowledge structure, and intepretation in which the information is either enhanced, altered, or distorted (Alba & Hasher, 1983). One implication from such a formulation is- that one's self-schema largely determines one's social world in that it determines what types, of social information one receives. This information serves- to reinforce or alter the existing knowledge- structure which in turn, provides, the framework for making future judgments, decisions, inferences, or predictions, about the self. These structures, in turn, determine what types of social experiences are possible. In the case of the adolescent with a "shyness" self-schema, he would be reluctant to ask a girl for a date since his predictions of being accepted are sufficiently low. Besides its potential heuristic value in explaining behavior; the strength of schema theory lies in its predicted empirical referents. First of all , self-schemata are defined as well articulated knowledge structures of the self with, the result that the schematic individual has a distinct idea; of what type of person he or she- is for a particular behavioral domain. Thus the. schematic person is. likely to endorse 7 schema-congruent adjectives as self-descriptive and to predict future schema-consistent behavior. Secondly, schema theory postulates that self-schemata work as a selective filter in information processing, predicting a bias for schema-congruent stimuli. Thus- a schematic individual is likely to process schema-related information with relative ease and will resist counter-schematic information. Thirdly, self-schemata are cognitive structures and as such, items that can be easily accommodated into the schema can be more easily recalled. Each of these empirical referents has been explored. In these studies, the typical instrument of assessment is the self-referent encoding; task (SRET) which asks the: individual to rate a list of adjectives in degrees of self-descriptiveness. This may be- followed by a recall task. Such an assessment involves the contrast of the- pattern of responses obtained from the processing of schema-relevant and schema-irrelevant adjectives of schematic and aschematic individuals. Markus (1977) provided preliminary evidence for the construct validity of self-schemata by demonstrating that self-schemata can be influential in the types of responses selected and the latencies of self-relevant judgments. Markus reasoned that schema-based processing, results in a particular pattern of responses: a person with a developed self-schema, would process self-relevant information with relative ease, retrieve more behavioral examples of schema-consistent behavior, predict more future schema-consistent behavior, and resist counterschematic information about him-or-herself. To test this, Markus- obtained and classified two groups of schematic individuals (dependent and independent), and one group of aschematic individuals (those who described- themselves neither as dependent or independent) according: to scores, obtained from the- Gough-Heilburn Adjective Checklist (Gough & Heilbrun, 1965). These individuals; were then- asked to rate a list of trait adjectives in terms of self-descriptiveness; cite instances from their past, behaviors, of self-rated trait 8 adjectives, estimate how likely they would engage in future schema-consistent behaviors, and rate the veridicality of counter-schematic information. In every instance, schema theory was predictive of performance. Independent individuals rated schema-consistent adjectives quicker, generated more future occurrences of independent behaviors, and resisted feedback that was counter to their self-schemata. The reversed pattern was obtained for dependent individuals. In contrast, aschematic individuals did not differentiate between independent and dependent stimuli: they had greater difficulty in providing behavioral evidence of independence and dependence, thought they were as likely to engage in independent and dependent behavior, and were relatively accepting, of. feedback about themselves. The self-schemata have been found to have- a facilitative effect for schema-congruent words. Markus, Crane, and Siladi (1978) classified individuals as feminine; masculine, or androgynous by their responses on a self-report inventory on sex roles. After the completion of this inventory, subjects were given an incidental recall task and were asked to recall as many adjectives from the inventory as possible. Again, schema theory was predictive of performance. Masculine subjects recalled significantly more masculine words than feminine words, while feminine subjects recalled significantly more feminine words than masculine words. In contrast, androgynous subjects showed much less differentiation in the number of masculine and' feminine words recalled. The facilitative schematic effect on memory may be suggestive of the potency of the self-schema in organizing personal information. Rogers, Kuiper, and Kirker (1977) contrasted the- encoding of adjectives produced by the self-referent task (describe you?) with the structural, task (big or small letters?), phonemic task (rhymes with?) and the semantic task (means- same as?). The investigators reasoned that the different tasks would, lead to differential "depth-of-processing," where depth refers to 9 the degree of semantic involvement. The deeper or more elaborately that a word is encoded, the more likely it will be later recalled (see Craik & Lockhart, 1972). If the self-schema serves as a potent organizer of personal information, then adjectives processed under the self-referent instructions, as compared with the other instructions, would lead to richer and more elaborate encoding and consequently greater ease of retrieval. The results were supportive of this contention. On a different vein, the construct validity of self-schemata can be established by demonstrating that self-schemata are a particular form of cognitive structures and as such,, should follow- similar principles. Rogers (1977) suggested that the- self can be: conceptualized as a specific set of cognitive structures called prototypes (for a discussion of prototypes, see for example Rosch,. 1973). Prototype theory predicts two results. First of all,, it has been demonstrated that the probability of committing: a false alarm in a recognition task (saying a previously unseen item has been seen before) increases with the item's similarity to the prototype- (Cantor & Mischel, 1977). Thus, schema-related stimuli should be particularly suspectible to this false alarm effect. Rogers, Rogers,, and Kuiper (1979) gathered self ratings on a series of adjectives. On a recognition task conducted 3 1/2 months later, it was found that subjects committed more false alarms on items previously rated as highly self-descriptive. Secondly, prototype theory predicts that the ease of deciding whether an item belongs to a category relates to the item's degree of resemblance of the prototype. Highly prototypical as well as highly unprototypical items are associated with quicker decision latencies while- items, in the mid-range of prototypicality are associated with longer response latencies (Rosch, 1973). This inverted-U function has been demonstrated for self-referent judgments where- highly self-descriptive and highly non-self-descriptive- stimuli showed faster latencies- than moderately self-descriptive 10 items (Kuiper & Rogers, 1978). The argument could be raised that the self-referent effect is due to the social nature of the task rather than to a schematic effect per se. Kuiper and Rogers (1979) compared the recall for items following judgments made about the self, stranger, and a person known for some time. Recall of words judged in reference to the self was enhanced ralative to the recall in the complete stranger condition but not relative to the well-known person condition. Additionally, Kuiper (1978) found enhanced recall for the self condition, as compared to the recall for judgments made for the better known acquaintances. The.' sum of the studies provides converging evidence of the construct validity of: the' self-schema. The- robustness of the schema effect has been demonstrated by a wide range of stimuli, such as personal adjectives (e.g. Markus, 1977), faces (Mueller, Bailis, & Goldstein, 1982), and situations (Kendzierski, 1980); for a wide range of behavioral domains such as, independence (Markus, 1977), sex roles (Markus, Crane; & Siladi, 1978), shyness (Wurf & Markus, 1982) and obesity (Markus, Hamill, & Sentis, 1979); and for both adult (e.g. Markus, 1977) and child samples (Hammen & Zupen, 1984). But more central to this thesis, the applicability of schema theory has been demonstrated in the area of depression. The depressive self-schema The first, studies (Davis, 1979a; 1979b) attempting to find evidence of a depressive self-schema yielded largely negative results. In the first of these studies (Davis, 1979a), depressed psychiatric patients and normal subjects (university undergraduates and members from a clerical, pool) encoded a list of adjectives either through', ar self-referent judgment (describes you?), a structural judgment (capital letters?), a phonemic judgment (rhymes with?), or a semantic judgment (means the same as?). These adjectives were- nonpathological. in nature and were the same used 11 by Rogers, Kuiper, and Kirker (1977), a point that will be important in discussing alternative explanations. Overall, the depressed subjects showed poorer recall than the nondepressed subjects in the SRET. Futhermore, the depressed subjects did not show a facilitation effect of the SRET, and the recall data for words encoded under that task was comparable to those for words encoded under the semantic task. However, this facilitation was observed for the nondepressed group, leading Davis to conclude that the "self-schema is not an active agent in the encoding of personal information as it is with normals (p. 107)". Using a different approach, Davis (1979b) attempted to find evidence of the depressive self-schema by examining the subjective organization of personal adjectives and abstract nouns. In tests of subjective organization, a subject is given a list of words to recall in any order he or she wishes. The subject is given several trials of the same set of words with each trial involving a different random ordering. Over a number of trials, a consistent ordering or clustering of words develops and this is defined as subjective organization. If the self-schema is present and acts as a potent agent in organizing personal information, then subjective organization should be stronger for the self-descriptive adjectives than for the abstract nouns. The results did not support the hypothesis. Although severity level of depression was not related to the subjective organization of abstract nouns, depressed college students displayed lower degrees of subjective organization for their self-referent adjectives than the nondepressed college students. In another study, Davis and Unruh (1981) found evidence for schema-based processing for long-term depressives. Using a similar methodology as Davis (1979b), Davis and Unruh contrasted the degree of subjective organization of self-referent stimuli of long-term depressives, short-term depressives, and nondepressed subjects. Relying solely, on self-report estimates of duration, these researchers found that long 12 term depressives (mean duration of 74.6 months) displayed the same degree of subjective organization as the nondepressed controls. This result was contrasted with the relatively low degree of subjective organization exhibited by short-term depressives (mean duration of 5.33 months). The investigators concluded that there is a developmental aspect to schemata where the depressive self-schemata emerge only after a relatively long period of depression. One of the weaknesses of this study is that it relies solely on self-report of duration. Given that cognitive depression models postulate that depressed individuals negatively distort their perceptions, these self-report estimates are especially suspect. But if this developmental hypothesis is correct, then the negative results of the earlier findings can be explained by the employment of short-term depressives: the Davis (1979a) study employed depressives with an estimated mean duration of depression of 10 months, and the Davis (1979b) study utilized depresssives with an estimated mean duration of 5.8 months. Derry and Kuiper (1981) offered an alternative explanation for the above negative results. They proposed that schemata are "content-specific" and consequently any tests of a particular schema must utilize schema-relevant stimuli. As such, they argue that Davis' use of general nonpathological adjectives was an inappropriate test of the depressive self-schema. Derry and Kuiper (1981) found evidence for their content-specificity model of self-schemata. These investigators specifically used depression-related adjectives (e.g. bleak, dismal) and positive-content adjectives (e.g. amiable, loyal) that were previously rated for their relevance for and unrelatedness to the construct of depression by college students. The investigators asked clinical depressives, nondepressed psychiatric controls and normal nondepressives to encode these adjectives via the self-referent (describe you?), the structural (small letters?), or the semantic (means the same as xxxx?) tasks. This was followed by an incidental recall task. In contrast to the earlier studies, this study found enhanced recall for 13 the self-referent adjectives under specific content conditions. The normal and nondepressed psychiatric control groups displayed superior recall for self-referenced positive-content adjectives while the clinically depressed group displayed enhanced recall only for those depression-content adjectives that were previously encoded via the self-referent task. However this study failed to find any differences in decision latencies for the different content conditions between the depressed and nondepressed groups. The Derry and Kuiper (1981) findings have been replicated by Bradley and Mathews (1983). Using a different set of judgments, the investigators asked clinical depressives and non-psychiatric controls to rate positive and negative adjectives either by a self-referent judgment, a familiar-other judgment, or an unfamiliar-other judgment. To ensure that the negative-content adjectives (e.g. glum, depressed) were relevant to the construct of depression, 15 clinical psychologists rated these adjectives to be descriptive of a depressed person. Examples of positive-content words were "good" and "glad". The results paralleled those obtained by Derry and Kuiper (1981). Clinical depressives, compared to the non-psychiatric controls, recalled more negative than positive self-referent adjectives.' Futhermore this negative bias displayed by the depressed group was present only for the self-referent judgment. For the other two other-referent judgments, the depressed group showed the same positive bias as the nondepressed group. As with the Derry and Kuiper (1981) findings, the two groups did not differ in decision speed. The Derry and Kuiper (1981) and Bradley and Mathews (1983) studies serve to illustrate that evidence for a depressive self-schema in a clinically depressed sample can be obtained if depression-related adjectives are employed. However these studies do not provide unequivocal support for the content-specificity hypothesis. These studies used two types of adjectives - depression-related and positive-content. The 14 failure to include a general negative-content condition means a failure to rule out the alternative hypothesis that the enhanced recall for the depression-content adjectives is due to an increased global accessibility of general negative cognitions. At least two points can be made in support of this alternative hypothesis. First of all, these studies have illustrated that depressed psychiatric patients, unlike their nondepressed counterparts, are characterized by an absence of a positive bias. This absence may be a reflection of the depressed patients' greater accessibility to negative cognitions. Secondly, the alternative hypothesis would be consistent with studies showing that induced dysphoric mood can have a global and pervasive effect on, for example, perceptions and estimations of unrelated negative events (e.g. Johnson & Tversky, 1983). This issue is further addressed by this thesis. Derry and Kuiper (1980) investigated the effects of mild levels of depression on schema-based processing. Using a college sample, the investigators identified depressed subjects as those who obtained a Beck Depression Inventory (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961) score of nine or more, and nondepressed subjects as those who scored less than nine. Using the same experimental procedure as Derry and Kuiper (1981), the investigators found enhanced recall for self-referent words under certain conditions. Similar to the clinical studies, an enhanced recall for the positive-content words was found for the nondepressed group. For the mildly depressed group, there was also a facilitative effect for the self-referent words, although this effect was found for both depression-content words and positive-content words. More simply stated, both the mildly depressed and the nondepressed groups displayed a enhanced recall effect for the positive-content words; what distinguished the two groups was the depressed group's additional bias for the depression-content words. In another study utilizing college students, Kuiper and MacDonald (1982) 15 extended Derry and Kuiper's (1980) study by the inclusion of decision latencies. The recall data were similar: Nondepressed subjects showed an enhanced recall for the positive-content adjectives as compared to the depression-content adjectives, whereas the mildly depressed subjects did not show a differentiation between the two word conditions. The study also showed that the nondepressed group, compared with the mildly depressed group, had shorter decision latencies for both word conditions. The nondepressed group's reaction time advantage can be explained in the following manner. As mentioned before, highly self-decriptive and highly non-self-descriptive words are judged with relative ease (Kuiper & Rogers, 1978). The nondepressed subjects were confident that they saw themselves positively and were confident that they did not see themselves as depressed. Thus they responded quickly to the two types of words. However, the mildly depressed subjects saw themselves as depressed but the subclinical nature of their depression prevented extreme judgments. Similarly, the subclinical depression prevented the mildly depressed subjects from describing themselves in extreme positive terms. This pattern of confidence and uncertainty was reflected in the decision latencies. In another study utilizing a college sample, Pietromonaco and Markus (1985) asked mildly depressed and nondepressed college students to imagine, recall, and make inferences about a variety of events while thinking about themselves or somebody else. The events were either sad or happy, and were either social or nonsocial in nature. The group differences in recall for both happy and sad events were not significant. In imagining and making inferences about the sad events, the mildly depressed subjects gave higher ratings than the nondepressed subjects. Futhermore, this difference was restricted to the self condition. For the happy event, however, group differences were not obtained. These results paralleled Derry and Kuiper's (1980) findings where the mildly depressed and nondepressed subjects were not 16 distinguished by their processing of the positive stimuli, but were distinguished by their processing of the depression-content stimuli. In contrast to the clinically depressed patients, mildly depressed individuals did not show an exclusive bias for the depression-content stimuli. The difference between these two types of depressed groups, Derry and Kuiper (1980) proposed, can be directly attributed to the differing severity levels of depression. The clinically depressed group is characterized by a cohesive depressive self-schema, as reflected by the clear bias for the depression-content stimuli. Although the mildly depressed individuals may experience some of the symptoms relating: to depression, the very mild nature- of the symptoms prohibits precise identification. This state of confusion may lead: the depressed individuals, to include pathological as well as nonpathological content in their self-schemata. Thus mild levels of depression are characterized by confusion and lower- degrees of organization surrounding, the self-prototype. This state, in turn, leads to uncertainty in the applicability of positive- and depression-content stimuli, as reflected by the longer- decision latencies; and to the* accommodation of both types, of words, as reflected by the undifferentiated recall of positive-content and depression-content words: Furthermore, Derry and Kuiper proposed that the consolidation of the depressive self-schema would occur at. more severe levels of depression, as seen in the studies utilizing clinical depressives. Although this severity model of self-schemata explains the discrepancy between the clinically depressed's and mildly depressed's self-schematic processing, this assertion is based on cross-sectional results and lacks the decisive answers provided by longitudinal studies. The^ sum- of a l l these studies provides- converging- evidence for the construct validity of the- depressive, self-schema, although i t appears in an attenuated and less cohesive form: in mildly depressed individuals. With the exception, of two studies (Davis 1979a, 1979b), each of these studies* was able, to discriminate between the 17 depressed or mildly depressed subjects and the nondepressed subjects. Given its sensitivity, the self-referent encoding task may potentially serve as a cognitive assessment device. The SRET as a cognitive assessment device The appeal of using the self-referent encoding task (SRET) as a cognitive assessment device is twofold. The first reason is obvious: the SRET purports to measure the self-schema which has been implicated in mediating the processes of depression. Secondly, the SRET assesses this self-schema by a consistent pattern of results from a variety of measures such as endorsement on self-report inventories, response latencies, recall, predictions, and so on. As such, the SRET does not rely solely on self-report measures which can be contaminated by the favorability of the trait term, the context of the situation, and other experimental demands, and are especially problematic in the context of depression since depressed individuals' responses are considered to be negatively distorted. In considering the utility of any assessment devices, the four purposes of assessment can be considered (Shaw & Dobson, 1981). First, assessment serves to describe the cognitive phenomena of depression. Secondly, assessment can be instrumental in formulating or testing theories of depression. Third, assessment can provide information about cognitive processes. Finally, assessment could be used to assess changes in cognition as a function of therapeutic improvement. In this manner, cognitive assessment could be informative in determining what cognitive changes necessitate change, or what cognitive changes prevent relapse. The SRET could potentially satisfy all four purposes. Although it is beyond the scope of this thesis to comprehensively explore all four conditions, this thesis intended to explore some of these conditions. With the aid of the SRET, this thesis intended to: a) further describe the cognitive phenomenon of depression by exploring the 1 8 relationship between anxiety and depression; b) test certain aspects of Beck's model of depression which postulates that the depressive self-schemas are stable entities, impervious to mood changes. Before specific hypotheses are discussed, some background material on these two issues is presented. The relationship between depression and anxiety The relationship between depression and anxiety has captured the interest of theorists and clinicians. Clinical practice has attested to the consistent and pervasive overlap between the" two syndromes. Clinically, symptoms of anxiety and depression co-exist more often that either is found in the pure state (Roth, Gurney, Garside, & Kerr, 1972). Vet the- two syndromes have' traditionally been viewed as distinct diagnostic entities (e.g. Diagnostic and Statistical Manual - third edition; American Psychiatric Association, 1980). The nosological distinction has foundation. Distinctions between anxiety and depression can be made on a number of levels. Affective level The affect appears, to be similar in depressed and anxious patients but varies in its intensity and course. Whereas- feelings of sadness- and despondency are severe among depressives, these feelings, are more mild and episodic among anxious patients (Derogatis, Lipman, Covi, & Rickels, 1972; Prusoff & Klerman, 1974; Roth, Gurney, Gariside, & Kerr, 1972). Additionally only depressives manifest diurnal variation of sad mood (Roth, Gurney, Garside, & Kerr, 1972). For anxious patients, feelings of tension and panic are severe and persistent while these feelings are milder and more episodic among, depressives- (Prusoff & Klerman, 1974). Behavioral level Many forms of depression are characterized by a reduction of activity. Depressives lack initiative, are less responsive, and often take little interest in their-• - 1 9 environment. They feel that they lack energy and may show psychomotor retardation. On the other hand, anxiety is characterized by hyperactivity. The anxious patient feels agitated, "on edge". He becomes behaviorally and autonomically more responsive, showing the "flight or fight" response. Somatic level Depressives often manifest a reduction or disturbance in the basic bodily functions. They often complain of poor appetite, insomnia and other sleep disturbances, and reduced sexual drives (Diagnostic and Statistical Manual - III, American Psychiatric Association, 1980). Although anxious patients generally lack these symptoms, i t has been noted that, both groups suffer equally from insomnia and other sleep disturbances (Roth, Gurney, Garside, & Kerr; 1972). Cognitive level Accompanying cognitions to depression are hopelessness and feelings of self-blame and self-criticism (Beck, 1967). Such thoughts are less prominent in anxiety. On the other hand, anxiety is- characterized by repetitive thoughts of danger and impending, disaster (Beck, 1976), and by persistent feelings of dread and panic (Kolb, 1973). However other studies have shown that this worry component also occurs in depression (Derogatis, Lipman, Covi, & Rickels, 1972; Prusoff & Klerman, 1974). Beck,. Emery, and Greenberg (1985) postulated that depressed and anxious patients share the same cognitive characteristics. Both groups of patients tend to have negative self-evaluations, give overly pessimistic predictions, and negatively interpret their experiences.. The difference lies in the extent. Whereas depressives* negative appraisals are, pervasive, global, and exclusive, the anxious patients' negative appraisals: are more situationally specific and do not exclude the consideration of positive factors: While the- depressed person- feels that, his or her character is defective^ "through and through", this; self-loathing is often absent in the anxious 20 patient. With regard to future perspective, the anxious patient sees some hope while the depressed patient is unremittedly pessimistic. Both anxiety and depression appear to pivot on the theme of loss. In the case of anxiety, the cognition is one of potential loss of self-esteem, happiness, or ability to cope. The anxious patient anticipates possible damage to his relations with others, his goals, or his ability to cope (Beck, 1976), whereas in depression, this threat is perceived to be imminent or as having already occurred. The depressed patient feels that he has permanently damaged his relations with others or has been defeated in his objectives (Beck, 1976). Theories- of affect In theories of emotions, distinctions between the- two constructs have been made. The- Differential Emotion theory (Izard, 1972; 1977) proposes that anxiety and depression differ from: one another in their composition and pattern of the fundamental emotions. This theory proposes that anxiety is composed of fear (key emotion) and two or more of the fundamental emotions of distress, anger, shame (including shyness and guilt), and the positive emotions of interest-excitement. Depression, on the other hand, is composed of distress (key emotion), anger, disgust, contempt, fear, guilt, and shyness. Empirical support for the hypothesized patterns was obtained for both college samples (Izard, 1977) and psychiatric samples (Marshall & Izard, 1972). According: to the Differential Emotion theory, anxiety and depression differ from one another in two respects. First, depression is considered to be a more complex emotion than anxiety since more fundamental emotions (disgust and contempt) are involved. Secondly, the two differ in their- profiles of fundamental emotions where distress and hostility are more salient in the profile of depression, and fear, guilt, shyness,, and interest are more salient in the profile of anxiety. Circumplex models of affect, on the other hand, argue that an emotion can 21 best be represented by its position on a circular arrangement defined by two orthogonal higher-order dimensions. In one version, Russell (1980) postulated that the two defining, dimensions are pleasure-displeasure and degree of arousal (arousal vs sleep). Anxiety and depression are both represented by the displeasure continuum; what differentiates the two is the degree of arousal dimension where anxiety is represented on the arousal end and depression on the sleepy end. In a highly similar model, Tellegen (in press; Zevon & Tellegen, 1982) proposed a two-dimensional structure whose space is defined by the roughly orthogonal poles of "Positive Affect" and "Negative Affect". High scores on the Positive Affect pole, as. represented by high ratings of such adjectives as "elated", "active", and low ratings for "sleepy", "sluggish", represent a state* of positive engagement. Conversely, low scores: on the Positive Affect, pole (reverse pattern' of scores) represent the absence of pleasurable' engagement, or a state of nonpleasurable disengagement. High Negative Affect,, as represented by high scores on adjectives such as "fearful" and low scores, for "relaxed"', represents a state of unpleasurable engagement. In contrast, low Negative Affect represents non-unpleasant disengagement. This model almost parallels Russell's (1980) circumplex model with Positive and Negative Affect corresponding to the pleasant-unpleasant dimension and engagement-disengagement corresponding, to the degree of arousal dimension. What is advantageous, with this model is that it accentuates the distinctiveness of anxiety and depression with anxiety represented by the Negative Affect pole and depression represented by the (reversed) Positive Affect pole. Factor analytic studies In: a number of factor analytic studies based upon symptom presentation (e.g. Derogatis, Klerman, & Lipman, 1972; Derogatis, Lipman, Covi, & Rickets, 1972), anxiety and depression emerge as separate factors, provided that enough factors are extracted. Furthermore,, in. a review of 40 factor analytic studies employing: patients 22 with affective disorders, these factors emerged as orthogonal or nearly orthogonal to each other (Mullaney, 1984). On the various descriptive and theoretical levels, distinctions between anxiety and depression can be drawn. However, observations drawn from the psychometric and clinical literature illustrate that this distinction may be a purely conceptual one (Dobson, 1985b). Using psychometric measures, separation of anxiety and depression proves to be difficult. In a comparison study using the widely used anxiety and depression scales, Mendels, Weinstein, and Cochrane (1977) found that all correlations among the scales were positive and significant, leading the investigators to conclude that anxiety and depression were highly interrelated. In a more extensive comparison, Dobson (1985a) compared eight scales which included three state depression scales, two trait depression scales, and three trait anxiety scales. Consistent with the earlier study, Dobson found the every correlation among the scales was positive and significant. More importantly, when average correlations among scales were computed, the average correlation among anxiety and depression scales (r = .61) was indistinguishable from those computed from among the depression scales (r = .69) and among the anxiety scales (r = .66) suggesting that the scales have poor discriminant validity (Campbell & Fiske, 1959). Clinically, there is considerable overlap between the two syndromes. Although estimates of this overlap vary from study to study, the general consensus is 25 to 40% (Klerman, 1977). While these estimates are low enough for separation figures to reach statistical significance, it is uncertain that statistical separation represents an adequate degree of clinical separation. One can argue that a 25-40% overlap does not constitute adequate clinical separation. One such study assessing the classification overlap is Prusoff and Klerman 23 (1977). Using a variety of classification methods on 364 depressed women with 364 matched anxious women, the authors reported a misclassification rate of 31.9 to 39.6%. Furthermore, this study indicated that symptoms of anxiety were more common among depressed patients than were symptoms of depression among anxious patients. Although both groups showed an equally high level of anxiety, only the depressed group showed the high level of depression. In the case of the co-occurrence of anxious and depressive symptoms, differential diagnosis appears to be made on the basis of severity or saliency of certain symptoms. The Newcastle group in Britain (Gurney, Roth, Garside, Kerr, & Schapira, 1972; Roth, Gurney, Garside, & Kerr, 1972; Scapira, Roth, Kerry, & Gurney, 1972) argued that the anxiety and depression groups were separable by virtue of the severity of symptoms within their subtypes. This research group reported that the major distinguishing feature was the presence of more symptoms in the depressed group. These symptoms included older age of onset, early morning awakening, and absence of environmental precipitants. Another determining feature was the duration of anxiety and depression episodes: Although both groups had episodic anxiety and depression, more persistent anxiety was present for the anxiety group and more persistent depression was present for the depression group. Downing and Rickels (1972) found that level of depression was a determining factor in the differential diagnosis of anxiety and depression. These investigators attempted to determine why, in a patient pool displaying mixed anxious and depressive symptoms, certain patients received anxiolytics (antianxiety drugs) while others received antidepressants. When contrasted between groups, patients who received antidepressants had higher levels of depression than patients who received anxiolytics. Patients who received anxiolytics, on the other hand, were rated as having more severe levels of anxiety than those who received antidepressants, although this difference was not of statistical significance. Among the patients who 24 received anxiolytics, their anxiety level was judged to be more severe than their depression level, although this difference was not great. However, among those who received antidepressants, their depression level was judged to be more severe that their anxiety level but in this case, the difference was striking. This pattern of results suggested that the diagnosis of depression was made not on the absence of anxious symptoms but was made when the severity of the depressive symptoms was of sufficient magnitude or when the difference between the severity of anxious and depressive symptoms was sufficiently large that the depressive symptoms became the more salient feature. These clinical studies- serve to illustrate three points. The first one is the co-occurrence of anxiety and depression in psychiatric patients. The second point is that although' classsification systems such as the Diagnostic and Statistical Manual -III (American Psychiatric Association, 1980) have- separate diagnostic categories for anxiety and depression,, differential diagnosis, is not made on the basis of the exclusion, of one category but is made on the- basis of the severity or saliency of certain features within a category. The third point is- that symptoms of anxiety appear to be more pervasive than symptoms of depression, a finding that will be important in later discussions- of theories relating, anxiety and depression. The finding that the phenomenon of anxiety is more prevalent than the phenomenon of depression has been replicated in a college sample. Investigators (e.g. Gotlib & Robinson, 1982) who attempted, a 2 x: 2 factorial design of low and high anxious and. depressed subjects had relative ease in filling the high anxious - low depression cell but extreme difficulty in filling, the low anxiety - high depression cell. In addition to classification- studies,, clinical studies on the history and course of affective and anxiety disorders point to a relationship between anxiety and depression. For example, at present, or past history of depression has been reported in 33 to 91%. of patients who met the Diagnostic- and. Statistical Manual • III criteria 25 for agoraphobia with panic attacks (Buglas et al, 1977; Bowen & Kohout, 1979; Dealy et al, 1981). Videbeck (1975) reported that in 23 to 66% of cases, the incidence of obsessions increased during a depressive episode. Using the Diagnostic and Statistical Manual - III criterion for major depression, Unde et al (1985) found that 50% of the panic patients has had a lifetime incidence of depression. Theories relating anxiety and depression The cogent trends in the clinical and psychometric data have led to theories relating anxiety and depression. Garber, Miller, and Abramson's (1980) model stresses that the anxious and depressive responses can be classified by perceived controllability and perceived probability of an outcome. Given a stressful situation, the anxious individual perceives that the probability of an undesirable outcome is independent of his behavior, and that the probability of that outcome lies somewhere between 0 and 1. The depressed individual also believes that he is helpless in avoiding an undesirable outcome, but he perceives the probability of that outcome to be 1. Thus an individual is either in an anxious state or a depressed state depending on his or her perceived probability of an undesirable outcome. And there lies the problem in this model. As discussed earlier, the striking feature of the anxiety and depression literature is the co-existence of the two phenomena. However this model requires that the anxious and depressed individual must make the simultaneous and impossible predictions of uncertainty and certainty for an undesirable event. Dobson (1985b) reasoned that since anxiety and depression tend to co-occur with anxiety as the more prevalent phenomenon, anxiety and depression must be hierarchically arranged with depression as the superordinate response. Within this model, the anxious and depressive responses are an interaction between the person's trait make-up and his or her temporal predictions to a stressful situation. According to this model, three trait configurations are possible: absence of trait anxiety and 26 absence of trait depression; presence of trait anxiety and absence of trait depression; presence of both trait anxiety and trait depression. When presented with a stressful situation, the type of response depends on the trait predisposition and the perceived probability of an undesirable outcome. If the threat of the undesirable outcome is low to moderate or i f the threat is far in the future, only those individuals with the trait anxiety will respond with anxiety. As the threat approaches closer in time, and the likelihood of the event becomes more certain, the anxiety response will become stronger. For those individuals with the trait depression, the anxious response will be admixed with depressive response. From the psychophysiological perspective, Gray (1982) also proposed a developmental sequence in the anxious and depressive responses. Basing his theory on the actions of antianxiety and antidepressant drugs, Gray proposed that both anxiety and depression are mediated by the septo-hippocampal system. Under stressful situations, there occurs an increase in noradrenergic, serotonergic, and cholinergic input to the septo-hippocampal system giving: rise to increased activity in that system. As stress continues, there- is a temporary exhaustion of nonadrenergic neurons, especially in the hypothalamic terminals. This exhaustion is manifest as helplessness, as measured by escape-avoidance behavior. If the increased activity in the septo-hippocampal system corresponds to anxiety, and helplessness is an analogue of depression, then this model, implies that under prolonged stress the initial reaction is anxiety. But' with time and prolonged stress, there will, be an increasing admixture of depression. The strength of. the Dobson (1985b) and Gray (1982) models is that they not only allow anxiety and depression to co-exist, but these' models explicity postulate that anxiety and depression are expected to co-exist under certain specified conditions. In theorizing the passage from anxiety to anxiety-depression, these two models- largely explain the nested* co-existence of the two constructs. However these 27 models do not account for the clinical phenomenon (although relatively rare) where depression can precipitate an anxiety response such as the case where a depressive episode can precipitate a panic attack (Unhe et al. 1985). Statement of first problem Even though the literature cogently argues for a relationship between anxiety and depression, the evidence is largely observational either through self-report inventories or through clinical observation. What is lacking is a body of experimental research that explicity seeks to document the co-existence of anxiety and depression. The SRET was chosen for such a task. Since depression is considered to be superordinate response, a depressive, response is highly predictive of an anxious response. In particular, i t is hypothesized that depressed, individuals- are characterized by a depressive- self-schema- as well as an- anxiety self-schema.. Stated differently, depressed, individuals should show parallel schema-based processing for both anxiety-content and depression-content adjectives. The specific hypotheses are outlined in Study 2. Cognitions as causal entities in depression One of the key features of Beck's model of depression (Beck, Rush, Shaw, & Emery, 1979) is that i t explicitly identifies depressogenic cognitions in the etiology of depression. Fundamental to this proposition is the concept of a traitlike depressive cognitive style that characterizes depression-prone individuals and persists beyond remission unless specifically identified and modified. Consequently, i t is imperative to the validation of Beck's theory to demonstrate the presence of depressogenic cognitions in depression-prone individuals in both the presence and, absence of a depressive episode. Research strategies- exploring: this idea take three' forms: a between-subjects design: comparing; remitted previously depressed patients with nondepressed controls; a within-subjects design comparing cognitive 1 processes, in depressed individuals while in. depressive episodes and: then, in remitted phase; and a 28 between-subjects design comparing depression-prone or "cognitively vulnerable" individuals with normal controls. As an example of the first research strategy, Altman and Wittenborn (1980) undertook a study to determine whether there were certain personality variables that differentiated formerly depressed women from women who had never been depressed. Sixty-two items from the 134-item inventory successfully discriminated these two groups. When these discriminating items were factor analyzed, five factors emerged: low self-esteem, unhappy outlook, narcissistic vulnerability, helplessness, and confidence. Although this study provided preliminary evidence that remitted depressives differ from: normal, controls, firm conclusions: can not be made- in the absence of pre-remission data. The possibility exists that these remitted women developed these characteristics as a scar from their depressive episodes. In a non-supportive study, Wilkinson and Blackburn (1981) administered three self-report cognitive measures to depressed patients, remitted depressed, patients, remitted mixed psychiatric patients, and. normal controls. The three measures were found to differentiate the currently depressed group from the other three groups. However the remitted depressed group did not differ from the remitted mixed psychiatric group nor from the normal controls. Again the lack of a. pre-test limits interpretation: i t is possible that the remitted depressed group initially differed from the currently depressed group and. perhaps this difference constitute a subtype of depressives which allowed the remitted group to alleviate their depression. Additionally two of the three measures used are problematic. The Cognitive Style Test was. devised, by the investigators for this study and had not yet been validated. Additionally the Hopelessness Scale- does not specifically predict stable responses across depressive and remitted, episodes.. An often cited longitudinal study is Lewinsohn, Steinmetz, Larsen, & Franklin 29 (1981). These investigators sought to ascertain whether depressive cognitions precede, accompany, or are a consequence of a depressive episode. Using a large pool of 998 volunteers obtained from the community, these researchers administered five self-report cognitive and self-esteem measures on two testing occasions separated by one year. They found that depressed individuals displayed concomittant depression-related cognitions. However this study failed to find any cognitive markers preceding or after a depressive episode. When the depressed individual were no longer in a depressed mood, they displayed cognitions similar to those in the nondepressed group. Furthermore, the participants who were to become depressed at the second testing session did not differ from the control group at time 1 on these measures.. This study suffers, from at least two limitations. As the researchers have pointed out, the generalizability of the results is questionable since only 5% of the original sample volunteered to participate. Secondly, this study used a mixed variety of cognitive measures with only one of the five measures relating to Beck's theory. Futhermore this- one relevant measure, "Expectancies of positive and negative outcomes", was not formulated with explicit consideration of Beck's theory and is difficult to evaluate since i t has unpublished psychometric properties. So far i t is difficult to draw implications from the studies cited since they either used measures that are only loosely tied to Beck's theory, or used measures that are theoretically reactive to remission. However the Dysfunctional Attitude Scale (DAS; Weissman, 1980; Weissman & Beck, 1978) has both desirable properties. Based explicitly on Beck's (1967) theoretical formulation of depression, the DAS purports to measure- those attitudes or beliefs constituting predispositions to depression; In other words, the DAS should be able to identify those individuals who are "cognitively vulnerable" or at risk for depression. The DAS also purports to assess attitudes that are enduring, in- nature and may not necessarily be moderated by the presence or 30 absence of a depressive episode. The research utilizing the DAS has yielded mixed results. There is some initial positive evidence for the ability of the DAS to identify depression-prone individuals. Kuiper, Olinger, & Air (1986; cited in Kuiper & Olinger, 1986) found that elevated DAS scores are predictive of future depressive episodes. These investigators found that among individuals who were previously identified as nondepressed and had elevated DAS scores, 27% of these individuals became depressed three months later, as compared to only 5% of those individuals who previously had low BDI - low DAS scores. Furthermore, Simons, Murphy, Levine, and Wetzel (1984) found that elevated DAS scores were predictive of relapse in a clinical sample. In: terms' of DAS' claims of tapping; stable aspects of depression, the evidence is generally negative. In longitudinal studies, DAS scores have been assessed first when the patient: is depressed and again when the patient had remitted. Four studies of this; kind (Eaves & Rush, 1984; Hamilton & Abransom, 1983; Silverman, Silverman, & Eardly, 1984; Simons, Garfield, & Murphy, 1984) have found that DAS scores have lowered significantly with the abatement of depressive symptoms. Two studies; (Hamilton & Abramson,. 1983; Silverman, Silverman, & Eardly, 1984) suggested that the DAS returned to normal levels while one study (Eaves & Rush, 1984) suggested that the remitted DAS scores are st i l l , higher than, that of comparison group. The one exception is Dobson and Shaw (1986) who found non-significant changes in DAS scores despite the remission of depresssion. However that study did find a lowering of 10 DAS points in the remitted, depressed group, as compared to a lowering of two points in the- two comparison; groups. The general; lack of support may be due to the presence- of therapeutic intervention in all of these studies. Beck, Rush, Shaw, and Emery (1979) theorized that the depressogenic schemata may be altered by direct cognitiver intervention. It is. noted that although a l l five studies involved 31 pharmacotherapy, it is possible that the physiological-affective-cognitive channels are interrelated and intervention via one channel affects the others. Kuiper, Olinger, MacDonald, and Shaw (1984) attempted to find evidence of schema-based processing in cognitively vulnerable but not currently depressed college students. At risk subjects were identified as those who scored in the upper half of a median split on the DAS. Subjects rated depression-related and positive-content adjectives either on a semantic judgment or a self-referent judgment. This was followed by an incidental recall task. The results corroborated previous findings: nondepressed-nonvulnerable subjects showed an enhanced recall for the self-referent positive words whereas the mildly depressed-vulnerable individuals showed an enhanced recall for self-referenced depression-content and self-referenced positive words. Vulnerable-nondepressed individuals, however, did not show any evidence of depressive self-schema processing. In fact, the vulnerable-nondepressed results were highly consistent with the nonvulnerable-nondepressed results. There are, however, at least two limitations in this study. First of all, the classification system used is probably too lax. According to the investigators' formulation, the cognitively vulnerable individuals were those who were at risk for depression. The use of a classification system based on the median split implies that these researchers suspected that half of their sample were prone to depression. This estimate is widely discrepant with the actual incidence of depression estimated to be 8 to 23% (American Psychiatric Association, 1980). In addition, the possibility exists that the cognitively vulnerable-nondepressed individuals have never been depressed. According to Markus' (1977) formulation of self-schemata, a self-schema develops after repeated experience in a particular behavioral domain. Thus an individual who has never been depressed, although he or she may be at risk for future depression, would not be expected to be characterized by a depressive self-schema. However, this does not mean that 32 predisposing factors can not be identified prior to any experience of depression. But it does mean that the application of schema theory in this context is inappropriate and more appropriate measures may examine, for example, certain assumptions about the self and the world. Statement of second probem The largely non-supportive findings from studies seeking stable depressogenic cognitive characteristics are limited by at least three problems: the use of measures that are not theoretically tied to Beck's theory or those that are reactive, in theory, to remission; the confound of therapeutic intervention; and inappropriate application of the self-schema paradigm to potentially never-been-depressed individuals. This thesis intended to curtail these problems in a longitudinal study seeking stable, cognitive markers in depression. This thesis used the SRET which, as discussed earlier, has theoretical ties to cognitive depression theory. Additionally, the SRET is a multimodal assessment device and as such, certain indexes are predicted to be reactive to remission (e.g. endorsement of adjectives) while other indexes are predicted to be robust across remitted mood (e.g recall). A list of these predictions is provided in Study 3. To avoid the potential modification of the depressive self-schema as a function of therapeutic intervention, this thesis assessed individuals in their natural fluctuating depressed and nondepressed moods. Specifically, this thesis intended to determine which aspects of schematic processing are reactive and robust across remitted depressive mood by assessing individuals when they are first mildly depressed and when they are no longer depressed. Overview of studies Study 1 : Study 1 provided normative ratings for a pool of adjectives on the properties of anxiety, depression, social desirability, emotional intensity, and imagery. 33 Out of this pool, four groups of adjectives were selected based on certain properties. These adjectives provided the stimuli for the SRET for Studies 2 and 3. Study 2 : Study 2 seeked to document the relationship between the anxiety self-schema and the depressive self-schema; Study 3 : Study 3 explored which aspects of schematic processing are sensitive to and resilient to remitted depressive mood. 34 Study 1 Normative ratings on 62 adjectives on the properties of anxiety, depression, social desirability, emotional intensity, and imagery. The goal of Study 1 was to select adjectives for the SRET based on a number of criteria. These adjectives formed four groups: depression-content, anxiety-content, negative-content, and positive-content. The first criterion was in consideration of the content-specificity approach to self-schemata research which underscored the necessity of using schema-relevant adjectives (see Derry & Kuiper, 1981). Accordingly, the anxiety-content and depression-content adjectives were selected for their relevance to the constructs of anxiety and depression, respectively. Since this thesis intended to study the relationship between anxiety and depression, any definitional overlap between the two constructs would have obscured the issue. Consequently, the anxiety-content and depression-content adjectives were required to be distinguishable with respect to their ratings of anxiety and depression. The question of whether the self-schema is specific to anxiety and depression or due to a more general negative set was addressed by the inclusion of a negative-content condition. These negative-content adjectives were required to be socially undesirable in content and not related to the constructs of anxiety and depression. Finally, a positive-content condition was included for comparison purposes. The adjectives in the four content conditions were equated on four properties. Since the SRET is a reading task, the adjectives were equated on word length (i.e., the number of letters in a word). At least one study (e.g. Scubert, Spoehr, & Lane, 1981) has shown that decision latency of a lexical decision task increases with word length. Secondly, these words were equated for word frequency since frequent words 35 tend to be perceived quickly and recalled readily (see for example, Broadbent, 1967; Broadbent & Broadbent, 1973; Catlin, 1969; Frederiksen, 1971; Nakatene, 1973). In addition, the imagery values of the words were considered since highly imaginable words are more easily recalled that words of low imagery value (Paivio, 1965; 1971; Paivio, Yuille, & Madigan, 1968). Finally, it has been proposed that intense affect intensifies schematic processing (Rogers, 1977) and consequently, the adjectives were equated for their emotional intensity ratings. Summarizing, the four groups of adjectives were required to be roughly equated for word length, word frequency, ratings of imagery, and ratings of emotional intensity. In addition, the positive-content adjectives were required to be judged to be more socially desirable than the depression-content, anxiety-content, and the negative-content adjectives, while the latter three content conditions were required to be roughly equivalent with each other on this property. Furthermore, the ratings of anxiety for the anxiety-content adjectives were required to be distinguishable from those obtained for the depression-content, negative-content, and positive-content adjectives. Likewise, the ratings of depression for the depression-content adjectives were required to be distinguishable from those obtained for the anxiety-content, negative-content, and postive-content adjectives. Method A pool of 62 adjectives was generated from a rational consideration of the anxiety and depression literature and was aided by a thesaurus (Laird, 1974). These adjectives were randomly presented in a questionnaire (see Appendix A) and were rated by 45 UBC undergraduates, 22 males and 23 females (mean age = 20.0), who received course credit for participation. These subjects were given definitions of the properties of anxiety, depression, social desirability, imagery, and emotional intensity 36 (see Appendix A). They then rated each adjective on these five properties using a 5-point Likert scale ranging from "1" or "not at all or very slightly" to "5" or "very strongly". Results Fourteen adjectives for each content condition were chosen and are presented in Tables 1 to 4 following the criteria discussed above. The desired pattern of data was confirmed by two analyses. First of all, a one-way multivariate analysis of variance for word length, word frequency, and ratings of anxiety, depression, emotional intensity, and. imagery, compared across content conditions yielded a significiant main effect for conditions (F(21,133) - 55.58, £ < .001). Pairwise confidence intervals, were computed for each variable using the OWMAR program (Hakstian, 1980) and are presented in Appendix B. The results of these analyses, as well as group means for each variable, are summarized in Table 5. The post-hoc analyses revealed that the four groups of adjectives did not differ significantly with respect to word length, word frequency, ratings of emotional intensity, nor ratings of imagery. Consequently any biases* in the self-referent processing for any content condition can not be the result of the above factors. In the case of the positive-content adjectives, these were judged to be more socially desirable than the depression-content, anxiety-content, and the negative-content conditions while the latter three content conditions were roughly equivalent in their social desirability ratings. Additionally, the anxiety-content adjectives were judged to e more significantly related* to> the construct of anxiety than the other three content conditions while the other three content conditions did not differ significantly with respect to their anxiety status. For the depression ratings, the depression-content adjectives were judged to be more closely related to the construct 37 Table 1. Anxiety-content adjectives Adjective WL WF A D EI SU I afraid 6 607 3.91 2.24 3.51 2.78 3.53 apprehensive 12 6 3.64 2.22 2.60 2.80 2.44 anxious 7 88 4.56 2.07 3.16 2.80 3.27 enraged 7 13 2.69 1.69 4.27 3.76 4.22 excited 7 270 2.62 1.00 3.87 1.56 3.42 fearful 7 72 3.67 1.67 3.71 2.91 3.20 hyperactive 11 1 3.91 1.22 2.87 2.84 3.60 jittery 7 2 4.16 1.42 2.82 3.24 3.22 jumpy 5 6 4.27 1.49 2.47 3.07 3.00 nervous 7 150 4.38 2.02 2.98 2.82 3.29 panicky 7 6 4.33 2.07 3.16 3.47 3.44 restless 8 71 3.83 2.44 2.93 3.83 3.29 tense 5 194 4.44 2.20 2.80 2.84 3.16 worried 7 176 4.29 3.22 3.24 2.80 3.07 WL = word length (number of letters) WF - word frequency according to Carroll, Davies, & Richman (1971) A = ratings of anxiety D - ratings of depression EI = ratings of emotional intensity SU - ratings of social undesirability I - ratings of imagery 38 Table 2. Depression-content adjectives. Adjective WL WF A D EI SU I blue 4 1071 1.87 4.13 3.31 3.18 3.27 defeated 8 96 2.18 4.24 3.29 3.40 3.11 dejected 8 4 2.40 4.56 3.44 3.78 3.27 depressed 9 14 2.31 4.96 3.78 4.02 3.56 desolate 8 25 2.27 4.31 3.20 3.62 3.42 despondent 10 2 2.16 4.13 2.62 3.80 2.73 dismal 6 22 2.13 4.09 3.00 3.87 3.22 downcast 8 6 1.82 4.27 3.29 3.22 2.98 dreary 6 19 1.53 3.93 2.53 3.93 3.22 lifeless 8 17 1.47 4.16 2.29 4.22 3.49 mournful 8 16 1.76 4.20 3.89 2.87 3.33 sad 3 309 1.89 4.58 3.82 3.04 3.36 sluggish 8 6 1.47 3.67 2.13 3.87 3.31 sorrowful 9 13 1.84 4.16 3.58 3.02 3.24 WL = word length (number of letters) WF = word frequency according to Carroll, Davies, & Richman (1971) A = ratings of anxiety D =• ratings of depression EI = ratings of emotional intensity SU = ratings of social undesirability I •= ratings of imagery 39 Table 3. Negative-content adjectives. Adjective WL WF A D EI SU I conceited 9 6 1.42 1.31 2.69 4.02 3.16 dishonest 9 7 1.91 1.53 2.62 4.51 2.93 greedy 6 28 1.80 1.33 2.53 4.40 3.24 homely 6 11 1.20 1.60 1.80 2.36 3.07 immature 8 10 1.82 1.42 2.29 3.64 3.24 immoral 7 6 1.44 1.33 2.36 3.78 2.51 mean 4 1266 1.64 1.80 3.00 4.47 3.29 messy 5 48 1.44 1.89 1.64 3.04 3.98 naive 5 6 1.46 1.34 2.00 3.12 2.73 repulsive 9 8 1.76 1.98 2.96 4.29 3.73 shallow 7 14 1.31 1.76 2.11 4.07 2.47 sneaky 6 25 1.84 1.33 2.33 3.58 3.07 stubborn 8 51 1.62 1.62 2.80 3.31 3.13 vain 4 53 1.64 1.31 2.22 3.78 2.93 WL = word length (number of letters) WF = word frequency according to Carroll, Davies & Richman (1971) A *= ratings of anxiety D - ratings of depression EI = ratings of emotional intensity SU = ratings of social undesirability I - ratings of imagery 40 Table 4. Positive-content adjectives. Adjective WL WF A D EI SU I assertive 9 7 1.60 1.20 2.33 1.53 2.69 attentive 9 10 1.91 1.09 1.78 1.36 2.58 cheerful 8 87 1.38 1.04 3.56 1.11 3.64 confident 9 37 1.13 1.07 2.36 1.44 3.02 dynamic 7 28 1.67 1.09 2.42 1.38 2.93 gifted 6 16 1.02 1.09 1.69 1.20 2.24 happy 5 774 1.33 1.02 3.67 1.24 3.80 industrious 11 8 1.47 1.11 1.78 1.20 2.73 loved 5 347 1.18 1.13 4.20 1.09 3.47 productive 10 45 1.29 1.09 1.67 1.13 2.44 rested 6 88 1.07 1.38 1.70 1.23 2.44 secure 6 75 1.04 1.16 2.51 1.09 2.76 spirited 8 13 1.82 1.04 2.82 1.18 3.02 thriving 5 29 1.71 1.12 2.40 1.31 3.03 WL = word length (number of letters) WF = word frequency according to Carroll, Davies, & Richman (1971) A = ratings of anxiety D = ratings of depression EI = ratings of emotional intensity SU = ratings of social undesirability I •= ratings of imagery Table 5. Attributes of the four content conditions Content condition Depression Anxiety Negative Positive Results of M M M M multiple Attribute (SD) (SD) (SD) (SD) comparison Word 7.36 7.36 6.64 7.43 length (1.95) (1.95) (1.78) (1.99) (PDAN) Word 116.43 118.71 109.93 111.71 frequency (286.20) (164.88) (333.21) (209.64) (ADPN) Ratings of 4.24 1.93 1.54 1.12 D>(ANP) * depression ( .31) ( .57) ( .24) ( .09) A>P * Rating of 1.93 3.91 1.59 1.40 anxiety ( .31) ( .60) ( .22) ( .14) A>(DNP) * Ratings of social 3.56 2.97 3.74 1.25 undesirability (.43) ( .55) ( .62) ( .14) (NDA)>P * Ratings of 3.25 3.30 3.11 2.92 imagery ( .21) ( .39) ( .41) (.46) (ADNP) Ratings of emotional 3.16 3.17 2.38 2.49 intensity ( .57) ( .51) ( .41) ( .81) (ADPN) * 2 < .05 A = anxiety-content condition D •= depression-content condition N = negative-content condition P = positive-content condition 42 of depression than the other three content conditions. Group-wise comparisons showed that the other three content condtions were equivalent in their depression ratings with the exception that the anxiety-content adjectives had a significantly higher depression rating that the positive-content condition. However this significant result may have been a statistical artifact arising from the small variability associated with the positive-content condition (sd - .09). The fact that the mean depression ratings for the anxiety-content condition (M •= 1.93) and for the positive-content condition (M = 1.12) do not represent separate judgment categories suggested that the result was not of any practical significance. Secondly, the data were submitted to two discriminant function analyses to assure that the groups of words could be adequately discriminated from one another. Using the social desirability ratings as the discriminating variable, the resulting discriminant scores were used to predict group membership. The results are summarized in Table 6. As the results indicated, the social desirability ratings did not lead to accurate discrimination among the anxiety-content, depression-content, and negative-content words. However in the case of the positive-content words, the analysis resulted in 100% correct classification. One anxiety-content word was misclassified as a positive-content adjective. However this misclassification was not considered to be crucial since the contrast-of-interest was between the negative-content and the positive-content words. The second discriminant analysis asked whether the anxiety-content and depression-content words could be distinguished on the basis of their anxiety and depression ratings. Using the anxiety and depression ratings as the discriminating variables, a plot of the analysis (see Figure 1) shows that the anxiety-content and depression-content adjectives were clearly distinguishable with respect to one another, while the negative-content and the positive-content adjectives were clustered 43 Table 6. Classification results from discriminant analysis using the ratings of social  desirability as the discriminating variable. Predicted group membership Actual Content condition group membership Anxiety Depression Negative Positive Anxiety- 10 1 2 1 content (71.4%) (7.1%) (14.3%) (7.1%) Depression- 5 2 7 0 content (35.7%) (14.3%) (50.0%) (0.0%) Negative- 3 3 8 0 content (21.4%) (21.4%) (57.1%) (0.0%) Positive- 0 0 0 14 content (0.0%) (0.0%) (0.0%) (100.0%) 44 9-8-7-6-Z 5" g Z ZD ^ 3 l— Z c i 1 z - 1 -o z < -2H o -3-A A A A A A Legend A ANXIETY CONTENT ADJECTIVES X DEPRESSION-CONTENT ADJECTIVES • NEGATIVE -CONTENT ADJECTIVES 8 POSITIVE-CONTENT AOJECTIVES x x X X X X - 4 --6 -6 - 4 -2 0 2 4 6 8 CANONICAL DISCRIMINANT FUNCTION 1 Figure 1. A plot of the four content conditions using discriminant scores  obtained from the anxiety and depression ratings. 45 together. In terms of predicting group membership on the basis of anxiety and depression ratings, the results are summarized in Table 7. As the results show, none of the non-anxiety-content words was misclassified into the anxiety-content group and none of the non-depression-content words was misclassified into the depression-content group. This further- supported the notion that the depression-content and anxiety-content adjectives were reliably distinguishable on the basis of their anxiety and depression ratings. Study 2 Is the- depression self-schema predictive of an anxiety self-schema? The goal of this- study was: to document the- relationship between the results of the self-referent encoding: of anxiety and depression related adjectives. Since anxiety and depression are often co-existing phenomena, i t was hypothesized that the presence of a depressive self-schema: would be predictive of the presence of the anxiety self-schema. Two pheripheral issues were also addressed by this study. The issue of a depressive bias, per se, versus a global negative bias was addressed by the inclusion of both a depression-content condition and a negative-content condition. Additionally, the: issue of the severity of depression level as a determinant of the self-schema: was addressed by comparing, schema-based processing of mildly depressed individuals with moderately depressed individuals. The- specific hypotheses were: a.) Depressed individuals are characterized by a depressive self-schema and display a schematic bias in their processing, of depression-content adjectives, as seen. in. a. higher endorsement rate, higher- ratings, of descriptiveness, shorter decision latencies, and greater- recall for these adjectives. 46 Table 7. Classification results from discriminant analysis using the ratings of anxiety  and depression as the discriminating variable. Predicted group membership Actual Content condition group membership Anxiety Depression Negative Positive Anxiety- 14 0 0 0 content (100.0%) (0.0%) (0.0%) (0.0%) Depression- 0 14 0 0 content (0.0%) (100.0%) (0.0%) (0.0%) Negative- 0 0 12 2 content (0.0%) (0.0%) (85.7%) (14.3%) Positive- 0 0 1 13 content (0.0%) (0.0%) (7.1%) (92.9%) 47 b. ) The depressive bias displayed by the depressed individuals is not a reflection of a global negative bias and consequently, distinctive schematic processing patterns would be observed between the depression-content and the negative-content adjectives. c. ) Severity of depression level is not a contributory factor to schematic processing and highly consistent schematic processing would be observed among the mildly depressed and moderately depressed individuals. d. ) Evidence for the co-existence of anxiety and depression can be obtained by the observation of parallel schematic processing of depression-content and anxiety-content adjectives by depressed individuals. Method Measures Beck Depression Inventory The Beck Depression Inventory (BDI; Beck, Ward, Mendelson, Mock, & Erbaugh, 1961; see Appendix C) was used in the classification of depressed and nondepressed individuals. The BDI is a 21-item set of ordered statements intended to provide a comprehensive survey of depression symptomalogy. It is a self-report inventory which asks the individual to describe his or her current state of depression. The BDI is intended to provide some indication of the severity of depressive symptoms and yields a potential range of scores from 0 to 63. Higher BDI scores reflect more severe levels of depression. Psychometric properties of the BDI are considered to be good. Beck (1972) reported satisfactory internal consistency: Spearman-Brown corrected split-half reliability of .93 and Kruskal-Wallis item-total correlations ranging from .31 to .68. The BDI has been shown to correlate significantly with other self-report measures of 48 depression (including the MMPI-D scale, Zung's Self-Rating Depression Scale, and Lubin's Depression Adjective Checklist), with clinician rating scales, and with a behavioral observation scale (as reported by Rehm, 1981). The BDI has good test-retest reliability of .78 over three weeks for college students (Oliver & Burkham, 1979) but is also sensitive to clinical change, as demonstrated by its frequent use in therapy outcome studies (Rehm, 1981). Although the BDI was originally designed for psychiatric populations, it has since been validated on subclinical samples. Using a college sample, Blumberry, Oliver, and McClure (1978) obtained a correlation of .77 between the BDI scores and overall psychiatric interview ratings. In agreement, Hammen (1980) found a correlation of .80 between the Hamilton Rating Scale for Depression and BDI scores. In Hammen's sample, half of the students' depression was considered to be transitory over the 2-3 week period. But of those who remained depressed, many had diagnosable major or minor depressive disorders and displayed symptomatology similar to clinically depressed persons. Dysfunctional Attitude Scale The Dysfunctional Attitude Scale (DAS; Weissman, 1980; Weissman & Beck. 1978; Appendix D) is a 40-item attitude scale yielding a potential range of scores from 40 to 280. Based on Beck's (1967) theoretical formulation of depression, the DAS contains a series of attitudes or beliefs thought to be reflective of depressogenic assumptions. The psychometric properties of the DAS are considered to be good. Cronbach's alpha coefficients of .88 for college students (Dobson & Breiter, 1983), .90 for an adult population (Oliver & Baumgart, 1985), and .91 for depressed psychiatric patients (Dobson & Shaw, 1986) consistently suggest a high level of internal consistency. For an adult sample, Oliver and Baumgart (1985) obtained a test-retest reliability 49 coefficient of .73 over a 6-week period. A moderate correlation of .41 for adults (Oliver & Baumgart, 1985), a moderate correlation of .30 for college students (Dobson & Breiter, 1983), and a correlation of .64 for psychiatric patients (Dobson & Shaw, 1986) between the DAS and the BDI confirm the notion that the DAS is related to measures of depression. However the relationship is only a moderate one and this suggests that the DAS is conceptually different from measures of depression. Stronger relationships have been found with more conceptually similar measures such as the Automatic Thoughts Questionnaire (r - .78; Dobson & Shaw, 1986) and the Measures of Distorted and Depressed Cognitions (r = .62; Krantz & Hammen, 1979). State-Trait Anxiety Inventory The State-Trait. Anxiety Inventory (STAI; Spielberger; Gorsuch, cY Lushene, 1970; Appenidx E) was; used to assess- state and trait anxiety levels. The STAI-trait scale consists of 20 statements that ask: individuals; how they generally feel. Likewise the STAI-state scale also consists of 20 statements but asks subjects to describe how they feel "right: now, that is, at this moment". Both scales have a potential range of score 20 to 80 with higher scores suggestive of more intense anxiety levels. Psychometric properties of the STAI are considered to be good. Both the Trait and State subscales have a high degree of internal consistency with Kruder-Richardson-20 coefficients ranging from .83 to .92. Additionally, the STAI trait subscale correlates significantly with IPAT Anxiety Scale, the Taylor Manifest Anxiety Scale, and the Zuckerman Affect Adjective Checklist (Spielberger, Gorsuch, & Lushene, 1970). Classification system For this study, classification of subjects was- based on two depression measures. Although the STAI was administered and analyzed, STAI scores were not included in the classification system. Classification of depression, lev e l was made in 50 accordance to the criteria suggested by Burns (1980). Moderately depressed subjects were those who scored 16 or greater on the BDI while the mildly depressed subjects were those who scored between 10 to 15 inclusively on the BDI. Nondepressed subjects scored below 10 on the BDI. In addition to current levels of depression, depressed subjects were required to be depression-prone, as defined by elevated scores on the DAS. Subjects in the two depressed groups were required to be "cognitively vulnerable", as defined by DAS scores of 120 or more. In contrast, the nondepressed subjects were not "cognitively vulnerable" and met the criterion of DAS scores less than 120. Using norms from Dobson and Breiter (1983), the cutoff level of 120 was chosen because it represented one standard deviation above the mean. Consequently, subjects who scored 120 or more represented the upper 15th percentile. Since Dobson and Breiter (1983) found an absence of sex difference for the DAS, the same classification system was used for both males and females. Subjects Subjects were recruited from introductory psychology classes and were given course credit for participation. Interested students were asked to volunteer by completing the BDI. On the basis of their BDI scores, subjects were selected and contacted. Any subject who participated in Study 1 was excluded from Study 2. A total of 146 subjects participated. Fifty-two subjects were excluded from the data analyses because they failed to meet the DAS criteria. Five nondepressed subjects were excluded because they were classified as depressed at the second session. One depressed subject was excluded because she repeatedly failed to attend the second session. The remaining 88 subjects comprised of 31 males and 57 females whose ages ranged from 16 to 31 with a mean age of 18.94 (sd = 2.16). For Study 2, these 88 subjects formed three groups: moderately depressed group (MO; n = 22), the mildly depressed group (MI; n= 40), and the nondepressed group (ND; n = 26). These same subjects returned to participate in Study 3, and were reclassified according to their BDI scores at the second session (for more details, see Study 3). The Self-Referent Encoding Task The self-referent encoding task (SRET) involved presenting via Osborne microcomputer the four groups of adjectives described in Study 1 (see Appendix F for program listing). Prior to the presentation of the adjectives, subjects read the following instructions: This study consists of two sessions, each lasting one hour long. Session one consists of two tasks that will be described to you shortly. Session two will be the same as session one but will be conducted three months from now. Session 1: Encoding Task You will be shown a list of words, one at a time. You are to respond as quickly as you can stating whether or not the word describes you. You will then be asked how you rate the word as a description of yourself. For each word, the screen will be blanked for a few seconds, the word "READY" will then appear on the screen. When it does, place your finger on the spot marked READY on the keyboard. Then the word will be displayed on the screen. Read the word and decide if it describes you or not. Press one of the following keys, once you have decided: 1: if you decide "yes, the word describes me" 2: if you decide "no, the word does not describe me" Once you have decided if the word describes you, you will be asked to rate the word on a scale of 1 to 5, indicating how strongly the word describes you: 1 not at all 2 slightly 3 somewhat 4 well 5 extremely These choices will be displayed for each word that you are to rate. Try doing this with the following sample word. The process will be the same for the other words. (The sample word "SHORT" appears here) Good. Now you are ready to start. You will see 60 words, one at a time. You are to respond just as you have to the sample word. Remember you are to respond as quickly and as accurately as you can. After the list, the experimenter will come in, and will ask you to recall as many of the 60 words as you can. If you have any questions, please ask the experimenter now. Otherwise, press RETURN to begin the test. Each adjective trial began with the word "READY" presented on the center of the screen for 1 1/2 seconds followed by a blank screen for 1/2 second. The stimulus word then appeared. Subjects made their first judgment by pressing the "1" or "2" key from a resting position calculated to be an equal distance from the "1" and "2" keys. The reaction time was recorded by the computer as the time from the presentation of the stimulus word until a response was made. Subjects then made their second judgment and this also was recorded by the computer. The presentation of the adjectives was as follows: a practice item, two buffer items, the 56 experimental stimuli, and two buffer items. The buffer items were not included in any of the data anayles and were included to compensate for the facilitative primacy-recency effect of the subsequent recall task. The experimental stimuli were presented in sets of fours such that every content condition was represented in each set. The specific adjective and its position within the set were randomly selected. This method of presentation was selected to avoid any serial effect in the subsequent recall task. Following the computer task, subjects were asked to "Remember as many words from the task as possible. Remember them in any order. Take your time and please be thorough". Subjects were given as much time as they needed and their responses were recorded by the experimenter. Procedure Subjects completed the SRET followed by the completion of the BDI, DAS, STAI-state, and STAI-trait. These same subjects all returned to participate in Study 3. Self-schema indexes The self-referent encoding task produced four indexes which are considered to reflect operations of the self-schema. These indexes are: a) endorsement rate - the proportion of words that a subject rated as self-descriptive. b) descriptiveness ratings - a one to five rating of how well the adjective described the subject c) reaction time (RT) - the time spanning from when the stimulus appeared on the computer screen until the subject rated whether the word was self-descriptive or not. d) recall - the proportion of words recalled as a function of the content condition. 54 Results Subject characteristics Group means for the BDI, DAS, STAI-state, and STAI-trait are tabulated in Table 8. A one-way multivariate analysis of variance for these measures compared across the three groups of subjects revealed a significant main effect (F(8,164) = 42.38, £ < .01). Groupwise 95% confidence intervals were computed using the OWMAR program (Hakstian, 1980) and are presented in Appendix G. The results are summarized in Table 8. The three groups had significantly different BDI and DAS scores. As expected, the moderately depressed group had the highest BDI and DAS scores, followed by the mildly depressed group and the nondepressed group. The two depressed groups had significantly higher state and trait STAI scores than the nondepressed group, but the two depressed groups did not differ from one another on these measures. SRET results The design was a three groups (between groups factor) x four content conditions (within groups factor) multivariate split plot factorial (Kirk, 1982). First, the multivariate analysis revealed three significant effects: a significant main effect for groups (Wilks' lambda - .55, F(14,158) = 3.97, £ < .001), a significant main effect for content condition (Wilks' lambda = .38, F(21,716) = 13.70, £ < .001), and a significant group x content interaction (Wilks' lambda - .44, F(42,1171) = 5.41, £ < .001). The multivariate analysis was followed by univariate ANOVAs for each SRET index. The results are summarized in Appendices H - K. The significant main effect for groups was examined more closely by univariate ANOVAs for each SRET variable. These analyses revealed two significant univariate group effects. The significant group effect F(2,85) - 8.31, £ < .001, for the endorsement rate index was followed by Tukey-Kramer multiple comparisons for 55 Table 8. Group means for the BDI. DAS. STAI-state. and STAI-trait for Study 2. Moderately Mildly Nondepressed depressed depressed group group group Results Mean Mean Mean of multiple Questionnaire (SD) (SD) (SD) comparisons Beck Depression 20.18 12.13 3.00 Inventory (4.08) (1.80) (2.30) MO > MI > ND * Dysfunctional 163.64 143.95 98.00 Attitude Scale (15.96) (19.22) (14.09) MO > MI > ND * STAI-state 47.36 43.05 34.15 (10.94) (9.10) (6.12) (MO.MI) > ND * STAI-trait 55.41 48.77 33.15 (8.17) (7.70) (7.36) (MO.MI) > ND * * £ < .05 MO - moderately depressed group MI - mildly depressed group ND = nondepressed group unequal groups (Kirk, 1982). Computing averages across the four content conditions, the multiple comparisons suggested that the two depressed groups endorsed significantly more items that the nondepressed group (MO vs ND: .55 vs .41, g = 5.70, £ < .05; MI vs ND: .49 vs .41.g = 3.73, £ < .05) but did not differ from the one another (.55 vs .49, g = 2.67, £ > .05) For the degree of descriptiveness index, the ANOVA revealed a significant univariate group effect (F(2,85) = 10.63, £ < .001). The Tukey-Kramer multiple comparison results parallel the endorsement results. When averaged across the four content conditions, the two depressed groups rated the items as significantly more self-descriptive (MO vs ND: 2.58 vs 2.23, g - 6.17, £ < .05; MI vs ND: 2.48 vs 2.23, g = 5.05, £ < .05) but the two depressed groups did not differ significantly from one another (2.58 vs 2.48, g - 1.94, £ > .05). The depressed groups' willingness to endorse more items and to rate them as more self-descriptive merely reflects the choice of the stimuli (i.e., 1/4 of the items specificially related to depression and 2/4 of the items potentially related to depression). The univariate analyses for overall groups effect, more importantly, yielded nonsignificant effects for the RT index (F(2,85) = 2.91, £ > .05) and for the recall index (F(2,85) = .40, £ > .05) ruling out the confound that the depressed groups suffered from psychomotor retardation, memory deficits, or motivational deficits (see Miller, 1975). The significant multivariate group x content interaction specifically addressed the self-schema effect and was examined more closely. Group means for each SRET index are presented in Tables 9 - 12. Endorsement rate The univariate ANOVA revealed a significant group x content interaction (F(6,255) - 31.22, £ < .001). Consequently, group-wise Tukey-Kramer multiple comparisons were conducted within each content condition and the results are summarized in Table 9. As predicted, the three groups differed from one another in 57 Table 9. Group proportions for the endorsement rate index for Study 2. Moderately Mildly depressed depressed Nondepressed group group group Results Content- Mean Mean Mean of multiple Condition (SD) (SD) (SD) comparisons Depression .59 .31 .06 (.27) (.27) (.14) MO > MI > ND * Anxiety .74 .60 .40 (.18) (.19) (.30) MO > MI > ND * Negative .45 .35 .27 (.18) (.15) (.19) (MO (MI) ND) * Positive .42 .69 .91 (.20) (.19) (.12) ND > MI > MO * • £ < .05 MO - moderately depressed group MI •= mildly depressed group ND = nondepressed group Table 10. Group averages for descriptiveness ratings for Study 2 58 Moderately Mildly depressed depressed Nondepressed group group group Results Content- Mean Mean Mean of multiple Condition (SD) (SD) (SD) comparisons Depression 2.57 2.04 1.38 (.65) (.47) (.28) MO > MI > ND * Anxiety 2.99 2.63 2.24 (.68) (.45) (.61) MO > MI > ND * Negative 2.32 2.17 1.75 (.43) (.46) (.37) (MO.MI) > ND * Positive 2.43 3.07 3.53 (.55) (.47) (.40) ND > MI > MO * * E < .05 MO — moderately depressed group MI - mildly depressed group ND •= nondepressed group Table 11. Group reaction time averages (in seconds) for Study 2 Moderately Mildly depressed depressed Nondepressed group group group Results Content- Mean Mean Mean of multiple condition (SD) (SD) (SD) comparisons Depression 2.21 2.08 1.67 (.80) (.59) (.48) <MO,MI) > ND Anxiety 1.81 1.87 1.75 (.61) (.50) (.50) (MI.MO.ND) Negative 1.98 1.83 1.64 (.66) (.57) (.55) (MO.MI.ND) Positive 2.07 1.84 1.65 (.64) (.51) (.43) (MO(MI)ND) * £ < .05 MO = moderately depressed group MI = mildly depressed group ND = nondepressed group Table 12. Group proportions for the recall index for Study 2 60 Moderately depressed group Mildly depressed group Nondepressed group Results Content- Mean Mean Mean of multiple condition (SD) (SD) (SD) comparisons Depression .25 .21 .23 (.09) (.09) (.10) Anxiety .21 .27 .25 (.14) (.25) (.16) Not Negative .31 .22 .25 a p p l i c a b l e (.16) (.12) (.18) Positive .27 .25 .25 (.14) (.14) (.11) MO = moderately depressed group MI = mildly depressed group ND - nondepressed group 61 their endorsement rate of the depression-content items. Also as predicted, the moderately depressed group endorsed more of these items than the other two groups (MO vs ND: 3 -12.51, £ < .05; MO vs MI: a - 7.49, £ < .05) with the mildly depressed group endorsing more of these items than the nondepressed group (a -6.89, £ < .05). With the anxiety-content words, again the three groups differed from one another with the highest endorsement rate obtained by the moderately depressed group (MO vs ND: a - 7.92, £ < .05; MO vs MI: a - 3.60, £ < .05) followed by the mildly depressed group (MI vs ND: a °* 5.54, £ < .05) and followed by the nondepressed group. A reversed pattern was obtained for the positive-content words with the nondepressed group endorsing significantly more of these items than the two depressed groups: (ND vs MI: a - 6.18,, £ < .05; ND vs MO: a ~ 11-77, £ < .05) and the mildly depressed group endorsing- more of these items than the moderately depressed group (3 - 7.34, £ < .05). For the negative-content adjectives, the only significant: difference- was found between the moderately depressed and the nondepressed groups (MO vs ND: a - 4.33, £ < .05; MI vs ND: a ~ 2.38, £ > .05; MO vs MI: a - 2.60, £ > 05). Within-group Tukey-Kramer comparisons were conducted comparing the three groups' endorsement rates for the depression-content and the negative-content adjectives. For the moderately depressed group, these subjects endorsed significantly more' depression-content adjectives, that the negative-content adjectives (a -6.74, £ < .05). In contrast, the nondepressed subjects endorsed more negative-content adjectives, when contrasted to the depression-content adjectives (a - 10.71, £ < .05). However, the mildly depressed group did not differentiate between these two content conditions in their- endorsement rate (a - 2.70, £ > .05). Descriptiveness ratings The significant' univariate' ANOVA revealed a: significant group x content 62 interaction (F(6,255) - 27.85, £ < .001). The subsequent Tukey-Kramer multiple comparison results are summarized in Table 10. As expected, the pattern of results parallels the results obtained for the endorsement rate index. For the anxiety-content condition, the three groups differed significantly from one another (MO vs ND: a = 7.84, £ < .05; MI vs ND: 3 - 4.66, £ < .05; MO vs MI: a - 4.13, £ < .05). Likewise, the three groups were distinguishable with respect to their ratings of the depression-content adjectives (MO vs ND: a " 12.45, £ < .05; MI vs ND: a ** 7.94, £ < .05; MO vs MI: a = 6.05, £ < .05), and the positive-content adjectives (ND vs MO: 3 - 11.58, £ < .05; ND vs MI: a - 5.60, £ < .05; MO vs MI: a -7.31, £ < .05). For the negative-content adjectives, the two depressed groups rated these items as significantly more self-descriptive than the nondepressed group (MO vs ND: 3 = 6.01, E < .05; MI vs ND: a ~ 5.05, £ < -05) but did not differ from one another (a = 1.76, E > .05). In comparing the descriptiveness ratings between the depression-content and negative-content adjectives, the moderately depressed group rated the depression-content adjectives as more self-descriptive that the negative-content adjectives (3 = 4.99, £ < -05). In contrast, nondepressed subjects rated the negative-content adjectives as more self-descriptive that the depression-content adjectives ( 3 = 8.03, £ < -05). Although there was a tendency for the mildly depressed subjects to endorse more negative-content words as more self-descriptive than the depression-content words, this trend did not reach statistical significance (a = 3.47, £ > -05) Reaction time One particularly striking pattern in the RT data was the longer averaged reaction times for the two depressed groups processing the depression-content adjectives. At first glance, the results appeared to be contrary to the predicted 63 pattern. However this "reverse" pattern" may be the result of the subclinical nature of the student sample and the resulting ambiguity of their depression. As Kuiper and Rogers (1981) have noted, faster RTs were found for both extremely self-descriptive and extremely non-self-descriptive adjectives. In other words, longer RTs tended to occur for uncertain items. For the nondepressed sample, these subjects confidently saw themselves as not depressed and consequently rated the depression-content items quickly. In contrast, the depressed subjects admitted to some depressive symptoms, but the subclinical status of their depression may have resulted in ambiguity of the applicability of the depression-content items. Furthermore, this ambiguity may have resulted in longer reaction times. It is noteworthy that the predicted pattern of RTs can be obtained when the RTs of only the yes-rated words were considered (see Figure 2). Regardless of direction of RT patterns, the crucial issue still involved the group differences in the processing of the different content conditions. For this variable, the univariate ANOVA suggested a significant group x content interaction (F(6,255) = 3.23, £ < .01). Tukey-Kramer multiple comparisons for the anxiety-content condition, summarized in Table 11, revealed no significant group differences (MO vs ND: 3 = .51, E > .05: MI vs ND: 3 = 1.23, £ > .05; MO vs MI: g = .61, £ > .05). Likewise, there were no group differences in the processing of the negative-content adjectives (MO vs ND: 3 - 2.94, E > .05; MI vs ND: 3 - 1.86, E > .05; MO vs MI: 3 = 1.54, £ > .05). With respect to the depression-content adjectives, the two depressed groups differed significantly from the nondepressed group (MO vs ND: 3 = 4.63, £ < .05; MI vs ND: 3 = 4.12, £ < .05) but did not differ significantly from one another (3 •= 1.26, £ > .05). For the positive-content adjectives, only the moderately depressed group differed significantly from the nondepressed group (MO vs ND: 3 = 3.54, £ < .05; MI vs ND: 3 •= 1.90, £ > .05; MO vs MI: q= 1.54, £ > .05). 64 3 H 2 . 8 -O Moderately depressed group • Mildly depressed group O Nondepressed group 2 .6 -2 . 4 -Content condition Figure 2. Rating times (seconds) of the yes-responses, f o r the four content c o n d i t i o n s . 65 In comparing the RT pattern across the depression-content and negative-content adjectives, the two depressed groups differentiate^ processed the two types of words (MO: a - 6.84, £ < .05; MI: 3 - 10.40, £ < .05) while the nondepressed group did not differentiate between the two types of stimuli (a =* .91, E > .05). Recall The univariate group x content effect was nonsignificant for this variable (F(6,255) - 1.52, E > .05). Since it is known, however, that previously endorsed words tend to be- better recalled (see for- example, Craik & Tulving, 1975), the yes-rated recalled data were considered in a separate: analysis. In line with other researchers such as Derry and Kuiper (1981), the recall data were adjusted in the following; manner: To ensure that the adjusted recall data were not purely a function of the groups' differential endorsement rates in the various content conditions, each subjects' number of yes-rated recalled words was divided, by his or her number of endorsements, for that content condition. • In other words, the adjusted recall score represents the proportion of recalled words that the person had previously endorsed. In order not to artifically inflate, these estimates, any cell that had a "0" endorsement rate was considered as missing data and was excluded from the analysis. The group means and standard deviations for this adjusted score are presented in Table 13. The univariate- ANOVA for. these adjusted scores yielded a nonsignificant group x content effect (F(2,85) 1.74, £ > .05). In addition, the trends in the data ran contrary to predictions. Specifically, the nondepressed group appeared to recall more depression-content stimuli, and the two depressed groups appeared, to recall more positive-content adjectives, than the nondepressed group; These aberrant results: may have arisen from the groups' widely varying endorsement rates. Since endorsement rates formed the denominator- for the adjusted: recall data, the depressed Table 13. Adjusted recall proportions for Study 2 Moderately Mildly depressed depressed Nondepressed group group group Content- Mean Mean Mean condition (SD) (SD) (SD) Depression .28 .25 .37 (.19) (.32) (.30) Anxiety .22 .24 .18 (.15) (.21) (.18) Negative .28 .27 .23 (.23) (.26) (.22) Positive .31 .28 .25 (.19) (.17) (.14) Table 14. Non-parametric analysis for the recall data for Study 2. Moderately Mildly Content- depressed depressed Nondepressed condition group group group Depression .29 .19 .39 Anxiety .21 .26 .23 Negative .23 .25 .23 Positive .31 .28 .25 67 groups' relatively high endorsement rates for the depression-content adjectives, as compared to the nondepressed group, formed a negative bias for these groups' adjusted recall data. A similar argument applies to the nondepressed group's relatively large endorsement rate for the positive-content adjectives. The inverse relationship between the rate of endorsements and the adjusted recall data for the depression-content and the positive-content conditions (r =» -.68, £ < .05) supports the above argument. Another factor limiting the interpretability of the adjusted recall data lies in the number of missing, cells. In particular, only 10/26 of the nondepressed subjects endorsed any of. the depression-content adjectives: One potential solution to the missing data problem- was to use a non-parametric analysis. Such an analysis: involved tabulating, the number of yes-rated recalled words within a content condition across all. subjects, within a. group. This total was. divided by the total number of endorsements within a content condition, summed across all subjects within a group. The results of this analysis- (summarized in Table 14) did not lead to more interpretable results but yielded, a pattern that is identical to the pattern found for the adjusted recall data.. A. solution to the differential endorsement rates was. attempted by yoking subjects on endorsement rates. Specificially, the adjusted recall data for the three groups were compared across the three levels of endorsement rates. Since the- crux of the depressive schematic effect lies in the depression-content and positive-content conditions, these two results are presented in Figures 3 and 4. The patterns are difficult to interpret since- many cells contain either no or- small, numbers of observations. For the depression-content condition, adequate number of observations were found, for the "1-4" endorsement level. These results were contrary to • I , —1 1-4 5-9 10-14 "Endorsement rate Figure 3. Adjusted r e c a l l data as a function of l e v e l s of endorsement rate for the depression-content condition. 1^ 4 5 ^ '. T o H T Endorsement rate Figure 4. Adjusted r e c a l l data as a function of l e v e l s of endorsement rate for the positive-content condition. - -Mildly depressed group -•- • - Nondepressed group — — — Moderately depressed group 69 predictions: Nondepressed subjects had a higher proportion of recall than the nondepressed subjects. Likewise, in the positive-content condition, depressed subjects tended to have a higher proportion of recall than the nondepressed subjects. Discussion Before the possibility of an anxiety self-schema can be examined, evidence of a depressive self-schema must be established. Some studies (e.g Derry & Kuiper, 1981) have provided evidence of the depressive self-schema tangentially by establishing depressed-nondepressed group differences in the processing of the positive-content adjectives. In other words, a subclinical depressive self-schema was conceptualized as an absence of a positive bias which was considered to be prevalent in nondepressed individuals. In this investigation, evidence of a positive self-schema was found on three of the four SRET indexes. Nondepressed individuals consistently showed an advantage for the positive-content words as measured by: a higher endorsement rate, higher descriptiveness ratings, and quicker RTs (when compared with the moderately depressed subjects). More direct evidence of the depressive self-schema can be established by examining depressed-nondepressed differences in the processing of the depression-content adjectives. Again, consistent support was found on three of the four SRET indexes. The two depressed groups, compared with the nondepressed group, endorsed more depression-content words, rated these adjectives as more self-descriptive, and rated these more slowly. Summarizing, this study has replicated the depressive schematic effect in three of the four SRET indexes. Even though there are limitations in the interpretability of the recall data, the trends in the data are contrary to previous studies. Specifically, Derry and Kuiper (1981), using an adjusted recall measure of the number of recalled yes-rated words divided by the endorsement rate, found that depressed patients 70 recalled more self-referenced depressed words than nondepressed patients, while nondepressed patients recalled more self-referenced nondepressed content words. However; this study's data were in the opposite direction. Although there were differences in methodology, for example, the Derry and Kuiper study used a structural and semantic tasks in addition to the self-referent task, there is no compelling reason to believe that these differences would lead to a discrepancy in results. A clue may lie in the missing data cells. The reader is reminded that this study excluded missing data in its analyses so that i f the subject failed to endorse any adjectives for a content condition, that subject's recall data would constitute a missing data point. However Derry and: Kuiper did not employ this; policy. However these investigators justified their practice by stating that the actual number of missing data points represented only "1.3% of the TOTAL data points" (p 293, emphasis mine). In other words, the number of missing, cells was divided by the total number of data points produced by the three tasks (structural, semantic, and self-referent) and the two content conditions (depressed, nondepressed). Since the struqtural and semantic tasks were manipulated to produce equal proportions of yes and no ratings, missing data could not have been a problem for these two tasks. The missing, data must lie in the data, generated by the self-referent task. Although. Derry and Kuiper do not indicate precisely where the majority of the missing data points were, the endorsement rate of .15 for the nondepressed. group for the depression-content adjectives makes this condition suspect. Since these investigators included missing data, as represented by "0" points, in their calculation of the adjusted recall data, the nondepressed group's recall result for the depression-content condition would, be vastly underestimated. Although these investigators stated that exclusion of the missing, data points produced highly similar results, i t is noteworthy that the representation of missing, data as "0" in this study's data produce more comparable results to Derry and Kuiper's results. ! 71 For example, the pattern of adjusted recall for the depression-content words for this study would be: .28 (MO), .25 (MI), and .14 (ND). The series of analyses in this study conducted on the recall data illustrate the difficulty of using subsequent recall as an index of the self-schema. The heart of the controversy lies in the difficulty, if not impossibility, of "deconfounding" endorsement rate with the recall index. Future self-schema research could avoid this difficulty by employing recognition as a measure of the self-schema. Specifically, Rogers, Rogers, and Kuiper (1979) have demonstrated a high rate of false positives for schema-related stimuli on a recognition task. Since false positives involve "never-been-seen" stimuli, use of such a measure in future research would totally circumvent the problem of differing endorsement rates. The severity of depression level appears to be a determinant of schematic processing for two self-description indexes. Specifically, the moderately depressed group, compared to the mildly depressed group, endorsed more depression-content adjectives and rated these as more self-descriptive. On the other hand, the mildly depressed group, compared to the moderately depressed group, endorsed more positive-content items and rated these as more self-descriptive. However these self-description results could have been predicted from initial BDI differences. For the RT index, severity of depression level does not appear to be instrumental in schematic processing. For each of the four content conditions, the two depressed groups did not reliably differ from one another. Another issue was whether depressive biases for the depression-content adjectives were schema-content-specific biases or were reflective of a more general negative bias. A comparison between the depressed subjects' processing of the depression-content and negative-content adjectives addressed this isssue. For the self-description indexes, the moderately depressed group reliably endorsed more 72 depression-content adjectives and rated them as more self-descriptive. In contrast, the mildly depressed group failed to differentiate between the two conditions in their self-descriptions. However, in their rating times of these types of adjectives, both depressed groups significantly differentiated between the two content conditions. On a different vein, the specificity of the depressive schematic bias can be investigated by examining group differences of schematic processing of the depression- and negative-content adjectives. As discussed before, depressed subjects displayed a bias for the depression-content adjectives on three of the four SRET indexes. Such a bias can be seen for the negative-content adjectives on the two self-description SRET indexes: depressed subjects tended to endorse more negative-content adjectives and tended to rate these adjectives as more self-descriptive. This more general negative bias, however, appeared to be in an attenuated form. For the depression-content adjectives, as well as for the anxiety-content and positive-content adjectives, reliable differences existed among all three groups. For the negative-content adjectives, only the moderately depressed group differed significantly from the nondepressed group on the endorsement rate index, and the two depressed groups did not differ from one another on the descriptiveness index. A more profound difference was found in the RT index. Although the moderately depressed group displayed a bias for the depression-content adjectives on this index, depressed-nondepressed differences were not found on this index for the negative-content words. The argument could be raised that the negative-content adjectives constituted an imperfect control condition since negative adjectives such as "lazy" and "ugly" could be considered as descriptive of "personality traits" while the depression-content words were descriptive of negative affect. Although it was not the intent of this thesis, the anxiety-content adjectives provided the appropriate contrast since these adjectives were also descriptive of negative affect. For the two self-description 73 indexes, parallel schematic processing was observed for the depression-content and anxiety-content adjectives: the two depressed groups, compared with the nondepressed group, endorsed more depression-content and anxiety-content words and rated these two sets of words as significantly more self-descriptive. As with the negative-content adjectives, differential schematic processing between the depression-content and anxiety-content adjectives can be found in the RT index. Whereas group differences existed for the depression-content condition on this index, no group differences were found for the anxiety-content words. In summary, this study provided evidence for the positive and depressive self-schemata. The nondepressed group appeared: to be characterized by a positive self-schema, as seen as a schematic advantage for the positive-content adjectives, on three of the four SRET indexes.. On the other hand, the two depressed groups displayed a bias for the- depression-content words, providing support for the depressive self-schema. This, depressive schematic bias does not appear to be reflective of. a global negative bias since differential schematic processing for. the RT index was obtained between the depression-content with the anxiety-content and negative-content conditions. Given the presence of a depressive self-schema in depressed individuals, this study asked whether an anxiety self-schema also exists in these individuals. The data, did not support the hypothesis. At first glance, the two self-description SRET indexes, appeared to support the depression-anxiety relationship in that the anxiety-content self-description data paralleled the depression-content self-description data. Specifically, depressed individuals tended to endorse the depression-content adjectives as well as the anxiety-content adjectives, and they rated both sets of: adjectives as highly self-descriptive. However-this, pattern, does not provide unequivocal support for the depression-anxiety relationship since a similar pattern was also obtained for the 74 negative-content adjectives. Furthermore, these results are considered to be redundant with the self-report inventories results: depressed individuals obtained high scores on the BDI, as well as on the STAI-state and STAI-trait, and so it is not surprising that depressed individuals endorsed depression-content as well as anxiety-content adjectives. While these results do confirm the finding that depressed individuals also describe themselves as anxious, this finding does not contribute significantly to the literature since there is already ample research documenting a high degree of correlation between anxiety and depression self-report inventories (e.g. Dobson, 1985a). For the less^  redundant. SRET index - the RT index - there was no evidence of an anxiety self-schema in the depressed subjects. Neither of the depressed, groups showed any bias in. the processing, of: the anxiety-content adjectives. Furthermore, their- performance was indistinguishable: from that of the nondepressed group: The argument could be raised that the RT index was an insensitive self-schema measure. But this can not be the case since- this index provided clear evidence of the depressive self-schema (see- above discussion). As for the recall data, the assessment of the relationship between the' anxiety self-schema and the depressive self-schema with this index: is very problematic. Given the- differential endorsement rates of the adjectives across the three groups of subjects, reliable conclusions cannot be based upon these data. Since this study has failed to find evidence of an anxiety self-schema in depressed individuals, it is possible that the SRET is an inappropriate and insensitive measure of: the- anxiety self-schema.. A study was. conducted to investigate this possibility. Forty-four UBC undergraduates formed two groups of subjects. The high anxiety' group- (n-22) constituted those subjects who scored 50 or above on the STAI-state. Using: college norms, a raw score of 50 represented the 88th centile for 75 both males and females. The low anxiety group (n=22) were those who scored 34 or below on the STAI-state. Using college norms, a raw score of 34 represented the 24th percentile for males and the 29th percentile for females. Additionally, these two groups were roughly equated on depression level (high anxiety group: BDI — 13.27; low anxiety group: BDI - 8.45). These two groups completed the SRET using the same procedure as in Study 2. The results were not supportive of the presence of an anxiety self-schema. On all four SRET indexes, group differences could not be found for the processing of the anxiety-content adjectives. Methodological limitations could have led to this absence of results. As Derry and Kuiper (1981) have pointed, out, schematic effects can be reliably demonstrated-only when, schema-specific-content stimuli are used. However this condition was satisfied since Study 1 ensured the relevance of the anxiety-content words to the construct of anxiety. Another methodological limitation may have been the subject classification system. The college sample used was an analogue sample and as such, the question, of generalizability is an. issue (see for example, Kazdin, 1980). Perhaps the anxiety self-schema exists and. the- SRET is an appropriate assessment instrument but the anxiety level must be of clinical magnitude to produce the schematic effect.. Aside, from methodological issues, the possibility exists- that a self-schema does not exist for the^ behavioral domain of anxiety. While Beck, Emery, and Greenberg. (1985) have proposed that anxiety and depression share the same cognitive characteristics, they also proposed that the difference between the two lies in the pervasiveness of these cognitions. While the depressed individual applies negative cognitions impartially across- situations,, the anxious individual's negative- appraisals are more situationally specific. In. other words, depressed individuals provide repeated and; consistent depressive evaluations" while these evaluations are evoked only by certain situations for the anxious individuals. According to Markus' (1977) formulation, 76 a particular self-schema develops only after repeated and consistent evaluations for a particular behavioral domain. If the anxious individual provides anxious evaluations only for certain situations and these anxious evaluations are countered by more adaptive evaluations in other situations, it is possible that an anxiety self-schema may not develop. If this is the case, then the presence/absence of the depression/anxiety self-schemata may be one method of distinguishing between anxiety and depression. Returning to the original intent of this study, direct implications can be drawn for Study 2. Since this study has failed to find evidence of an anxiety self-schema in an anxious sample, it is not surprising that Study 2 also failed to find evidence of an anxiety self-schema in a depressed sample. Consequently, the failure of Study 2 to document a relationship between the depressive self-schema and the anxiety self-schema does not necessarily mean a failure to substantiate the relationship between the two constructs - it merely illustrates the possible inappropriateness (at / least in a college sample) of studying such a relationship using the self referent encoding task. Study 3 The stability of the depression self-schema across time and across remitted depressed mood Study 3 investigated whether there were stable aspects of the depressive self-schema. The stability issue was investigated across time and across remitted mood. An additional issue of interest was whether the SRET was predictive of remission. With this goal in mind, a comparison between the schematic processing of depressed subjects who remained depressed and depressed subjects whose depression had been alleviated was made. Specific hypotheses were: a. ) The SRET would demonstrate test-retest consistency. Subjects with nondepressed moods at both testing intervals would have demonstrated highly consistent schematic biases for the positive-content adjectives at both testing sessions. Similarly, subjects with depressed moods at both testing intervals would have demonstrated highly consistent schematic bias for the depression-content adjectives at both testing sessions. b. ) Certain indexes of the SRET would be sensitive to mood shifts while other indexes would be robust across mood shifts. For those subjects who were depressed at the first testing session- but were no longer depressed at the second testing, session, the two self-description indexes (the endorsement rate index and descriptiveness indexes) would reflect the mood shifts. For the RT index, it is debatable whether this index would be sensitive to mood changes. Kuiper and Rogers- (1981) have demonstrated that response latencies are a function of the descriptiveness; of the adjectives. Since the depression-content adjectives would no longer be applicable to the remitted depressed subjects, the decision latencies for these depression-content adjectives should be inconsistent across the two testing sessions. On the other hand, Markus (1977) postulated that decision latencies are a reflection of self-schema- operations. Since the depressive self-schemata are hypothesized to be stable entities, decision latencies for schema-relevant adjectives should be consistent across the two testing sessions despite the varying, moods. For the; recall data, it was hypothesized that this index would be insensitive to mood changes. The recall index is considered to be- reflective of schematic operations. Since the- self-schemata are hypothesized to be stable structures; 78 the recall index should be impervious to mood shifts, c.) It is debatable whether the SRET would be predictive of remission. On one hand, Beck (Beck, 1976; Beck, Rush, Shaw, & Emery, 1979) does not make a distinction among depressed individuals who remain depressed and those who spontaneously remit. In this respect, the cognitive structure of these two types of individuals should be indistinguishable. On the other hand, it is puzzling that certain individuals manage to alleviate their depression. It is reasonable to assume that individuals who can alleviate their depression have distinguishable cognitive structures, such as they respond more positively to the positive-content adjectives than the unremitted depressed individuals. Method Subjects Subjects who participated in Study 2 also participated in this study. A different classification system was used since this study asked whether the SRET could distinguish the remitted subjects from the unremitted subjects. Classification criteria Three groups of subjects were involved. The stable depressed group (SDG; n = 43) consisted of those individuals who were depressed at time 1 and remained depressed at time 2. The remitted depressed group (RDG; n = 19) consisted of subjects who were classified as depressed at time 1 but were no longer depressed at time 2. Finally, the stable nondepressed group (SNDG; n = 26) consisted of subjects who were not depressed at time 1 and remained not depressed at time 2. The classification system is summarized in Table 15. At time 1, classification was made on the two depression measures. First, depressed subjects were required to be currently in a depressed mood, as defined by a BDI score of 10 or more. Table 15. Subject classification system for Study 3 Time 1 Time 2 Stable BDI >= 10 BDI >= 10 depressed group DAS >- 120 Remitted BDI >- 10 BDI < 10 depressed group DAS >- 120 Stable BDI < 10 BDI < 10 nondepressed group DAS < 120 80 Additionally, depressed subjects had to be "cognitively vulnerable" to depression, as defined by a DAS score of 120 or more. Nondepressed subjects were defined as those who were not in a depressed mood (BDI score < 10) and who were not "cognitively vulnerable" (DAS score < 120). Classification of subjects at time 2 was made solely on the basis of BDI scores. Procedure Subjects completed the SRET, as participants in Study 2, followed by the completion of the BDI, DAS, STAI-state inventory, and STAI-trait inventory. This session constituted time 1. Approximately three months later, the same subjects returned and repeated the session (time 2). Subjects were then thanked and debriefed. The attrition rate for the time 2 sessions was minimal. Only one subject was dropped, as she repeatedly failed to show up for her second session. Results Subject characteristics Group means for the BDI, DAS, STAI-state, and STAI-trait are tabulated in Table 16. The design was a multivariate three groups (between-subjects factor) x two times (within-subjects factor) split-plot factorial (Kirk, 1982). The multivariate analysis yielded three significant effects: an overall main effect for groups (Wilks' lambda = .24, F(8,164) - 21.68, 2 < .001), an overall main effect for time (Wilks' lambda - .71, F(4,82) = 8.38, £ < .001) and a significant group x time interaction (Wilks' lambda = .61, F(8,164) - 5.85, £ < .001). Since the multivariate analysis yielded significant results, follow-up univariate ANOVAs were conducted for each measure. The ANOVA results are summarized in Appendices L - O. BDI scores The univariate ANOVA yielded a significant group x time interaction (F(2,85) = 25.63, £ < .001) and subsequently Tukey-Kramer multiple comparisons were conducted Table 16. Group means for the BDI, DAS. STAI-state. and STAI-trait for Study 3 8 ! Stable depressed Remitted depressed Stable nondepressed group group group Measures TI M (SD) T2 TI M (SD) T2 TI M (SD) T2 BDI 15.49 16.19 13.84 6.63 3.00 3.04 (4.98) (6.32) (4.20) (2.50) (2.30) (2.23) DAS 150.14 155.33 152.74 142.68 98.00 102.31 (19.10) (21.84) (23.40) (19.22) (14.09) (20.21) STAI-state 44.51 44.21 44.74 38.32 33.15 34.31 (9.67) (11.76) (10.74) (9.20) (7.36) (9.68) STAI-trait 54.44 52.19 50.89 43.26 38.77 32.69 (13.05) (8.78) (14.10) (5.27) (18.70) (6.01) 82 (Kirk, 1982). At time 1, the two depressed groups had significantly higher BDI scores than the nondepressed group (SDG vs SNDG: g - 16.37, £ < .05; RDG vs SNDG: g = 11.47, £ < .05). However, the two depressed groups did not differ in their depression level (g = 2.05, £ < .05). When the means were compared across time, the expected pattern was obtained. The two stable groups' BDI scores did not significantly change across time (SDG: a = 1-56, £ > .05; SNDG: a = 07, £ > .05) while the remitted group's BDI score was significantly lowered at time 2 (a - 10.72, £ < .05). At time 2, groupwise comparisons were conducted. As expected, the SDG had significantly higher BDI scores than the RDG (a - 11.87, £ < .05) and the SNDG (g = 17.23, £ < .05). The RDG had slightly higher BDI scores than the SNDG (g = 3.80, £ < .05) DAS scores The univariate ANOVA yielded a significant group x time interaction (F(2,85) = 5.77, £ < .001) and subsequent Tukey-Kramer multiple comparisons were conducted. At time 1, the two depressed groups had higher DAS scores than the nondepressed group (SDG vs SNDG: g - 16.87, £ < .05; RDG vs SNDG: g - 14.29, £ < .05). When compared with each other, the two depressed groups had roughly equivalent DAS scores (g «= .80, £ > .05). When compared across time, the pattern of the DAS results echoed those of the BDI scores. As expected, the two stable groups had stable DAS scores (SDG: g = 2.84, £ > .05; SNDG: g - 1.83, £ > .05). While the remitted depressed group had slightly lowered DAS scores at time 2 (g •= 3.65, £ < .05), the time 2 DAS scores remained significantly higher than those of the stable nondepressed group (g = 10.54, £ < .05) and slightly lower than those of the stable depressed group (g •= 3.57, £ < .05). Furthermore the stable depressed group had significantly higher DAS scores than 83 the stable nondepressed group at time 2 (a = 17.16, £ < .05). STAI-state scores The univariate ANOVA revealed a significant overall main effect for groups (F(2,85) = 15.24, £ < .001). Since there was a nonsignificant main effect for time (F(l,85) = 1.72, E > .05) and a nonsignificant group x time interaction (F(2,85) = 2.21, £ > .05), it was appropriate to collapse across time to make groupwise multiple comparisons. Averaging across time, the two depressed groups had significantly higher STAI-state anxiety scores than the nondepressed group (SDG vs SNDG: 44.36 vs 33.73, a - 5.65, £ < .05; RDG vs SNDG: 41.53 vs 33.73, a = 3.36, £ < .05) but the two depressed groups did not differ from one another (SDG vs RDG: 44.36 vs 41.53, a - 1.51, £ > .05). STAI-trait scores The univariate ANOVA revealed a significant overall main effect for groups (F(2,85) = 29.99, £ < .001). Since there was a nonsignificant group x time interaction (F(2,85) *- 1.00, £ > .05), the means were averaged across time. Groupwise Tukey-Kramer multiple comparisons suggested that the two depressed groups had significantly higher STAI-trait scores than the nondepressed group (SDG vs SNDG: 53.31 vs 35.73, a - 7.99, £ < .05; RDG vs SNDG: 47.08 vs 35.73, a *= 4.16 £ < .05) while the two depressed groups had roughly equivalent trait scores (SDG vs RDG: 53.31 vs 47.08, a - 2.69, £ > .05) Summary of subject characteristics results The analyses provided confirmation of the expected pattern of subject characteristics. At time 1, the two depressed groups were differentiable from the nondepressed group on the BDI, DAS, STAI-state, STAI-trait scores. However, the two depressed groups did not differ from one another on any of these measures at time 1, ruling out the hypothesis that any effect found in the self-schema data was 84 due to initial group differences. When compared across time, both the stable depressed group and the stable nondepressed group had stable BDI and DAS scores. In contrast, the remitted depressed group had significantly lowered BDI scores, which were just slightly higher than the stable nondepressed group's time 2 BDI scores. Although there was some tendency for the remitted depressed group's DAS score to decrease with remitted depressed mood, the remitted depressed group's DAS score remained significantly higher than the SNDG's time 2 DAS scores. SRET data For reasons of simplicity, the analyses were divided into two sections. The first set of analyses considered the time 1 data and asked whether there were any group differences between the stable depressed group and the remitted depressed group. The second set of analyses compared the time 2 data with the time 1 data and addressed the stability issue. Time 1 The design was a multivariate three groups (between-subjects factor) x four content conditions (within-subjects factor) split-plot factorial (Kirk, 1982). The multivariate analysis yielded three significant effects: a main effect for groups (Wilks* lambda = .58, F(12,160) •= 4.24, £ < .001), a main effect for content condition (Wilks' lambda = .41, F(18,708) - 14.64, E < .001), and a group x content interaction effect (Wilks* lambda - .57, F(36,1100) - 4.16, £ < -001). For the significant main effect for groups, the univariate ANOVA yielded two significant effects for the endorsement rate (F(2.85) = 6.50, £ < .01) and the descriptiveness ratings indexes (F(2,85) =10.14, £ < .001). Averaging across content conditions, groupwise Tukey-Kramer multiple comparisons suggested that the two depressed groups endorsed more items than the nondepressed group (SDG vs SNDG: 85 .52 vs .41, a = 5.25, £ < .05; RDG vs SNDG: .50 vs .41, 3 - 3.39, £ < .05) but did not differ significantly from one another (SDG vs RDG: .52 vs .50, a = -98, £ > .05). Additionally, the two depressed groups gave higher overall ratings that the nondepressed group (SDG vs SNDG: 2.54 vs 2.23, a - 6.60 £ < .05; RDG vs SNDG : 2.46 vs 2.23, a - 3.90, £ < .05) but did not differ from one another (2.54 vs 2.46, a - 1.60, £ > .05). These results merely reflected the nature of the stimuli: 1/4 of the stimuli (the depression-content adjectives) specifically related to the construct of depression and 2/4 of the stimuli (the anxiety-content and negative-content) potentially related to the construct of depression. The: univariate main effect for groups, more importantly, yielded non-significant results for both the RT (F(2,85) - 2.68, £ > .05) and the recall (F(2.85) - 1.10, £ > .05) indexes. This pattern suggested that the. groups; did not differ in overall RT nor overall recall, ruling out the? confound that the depressed groups, exhibited psychomotor retardation, memory deficits, or motivational deficits (see Miller, 1975). The multivariate analysis was followed by separate ANOVAs f o r each SRET variable. These results are summarized in Appendices P - S. The groups means for each SRET variable are presented in Tables 17-20. Endorsement rate The' univariate ANOVA suggested a significant group x content interaction (F(6,255) - 18.11, £ < .001). Groupwise Tukey-Kramer multiple comparisons conducted within the depression-content condition suggested that, the two depressed groups endorsed more of these items than the nondepressed group (SDG vs ND: a - 10.15, £ < .05; RDG vs SNDG: a " 5.63, £ <• -05) but: did not differ significantly from one another (a. " 2.97, £ > .05). Likewise, the two depressed groups endorsed more anxiety-content words than the nondepressed- group (SDG vs SNDG: a °" 6.55, E < -05; RDG vs. SNDG: a - 5.23, E < .05) but did not differ from one another (a - 04, £ > Table 17. Group proportions for the endorsement rate index for Study 3 86 Content condition Depression Anxiety Negative Positive M (SD) M (SD) M (SD) M (SD) TI T2 TI T2 TI T2 TI T2 SDG .45 .41 .65 .62 .40 .41 .57 .63 (.28) (.34) (.20) (.22) (.17) (.20) (.23) (.28) RDG .33 .20 .65 .48 .36 .36 .64 .77 (.32) (.19) (.21) (.27) (.17) (.19) (.23) (.18) SNDG .06 .09 .40 .41 .27 .22 .91 .90 (.14) (.11) (.30) (.27) (.19) (.21) (.12) (.11) Table 18. Group averaged descriptiveness ratings for Study 3. Content condition Depression Anxiety Negative Positive M (SD) M (SD) M (SD) M (SD) TI T2 TI T2 TI T2 TI T2 SDG 2.32- 2.32 2.78 2.66 2.23 2.37 2.81 2.74 (.58) (.76) (.55) (.61) (.46) (.53) (.53) (.63) RDG 2.01 1.64 2.70 2.28 2.20 2.03 2.91 3.08 (.58) (.39) (.61) (.47) (.43) (.37) (-70) (.53) SNDG 1.38 1.49 2.24 2.09 1.75 1.71 3.53 3.52 (.28) (.39) (.61) (.51) (.37) (.39) (.40) (.43) SDG = stable depressed group RDG = remitted depressed group SNDG = stable nondepressed group 87 Table 19. Group averaged reaction time (in seconds) for Study 3. . Content condition Depression Anxiety Negative Positive M (SD) M (SD) M (SD) M (SD) TI T2 T l T2 T l T2 T l T2 SDG 2.09 1.97 1.82 1.78 1.89 1.74 1.90 1.74 (.68) (.95) (.54) (.77) (.57) (.66) (.56) (.60) RDG 2.20 1.89 1.93 1.72 1.87 1.58 1.98 1.69 (.65) (.80) (.54) (.48) (.68) (.46) (.57) (.63) SNDG 1.67 1.54 1.75 1.59 1.64 1.39 1.65 1.45 (.48) (.50) (.50) (.39) (.55) (.28) (.43) (.39) Table 20. Group proportions for the recall index for Study 3. Content condition Depression Anxiety Negative Positive M (SD) M (SD) M (SD) M (SD) TI T2 T l T2 T l T2 T l T2 SDG .23 .28 .23 .27 .25 .31 .25 .28 (.09) (.11) (.22) (.15) (.15) (.14) (.14) (.15) RDG .22 .31 .29 .31 .26 .27 .27 .33 (.10) (.12) (.21) (.17) (.13) (.13) (.14) (.17) SNDG .23 .28 .25 .27 .25 .29 .25 .28 (.10) (.12) (.16) (.15) (.18) (.13) (.12) (.11) SDG = stabie depressed group RDG •= remitted depressed group SNDG - stable nondepressed group 88 .05). In contrast, the nondepressed group endorsed significantly more positive-content adjectives than the two depressed groups (SNDG vs SDG: g = 9.04, £ < .05; SNDG vs RDG: g 5.73, £ < .05) while the two depressed groups did not differ significantly from one another (g = 1.80, £ > .05). For the negative-content adjectives, the three groups did not differ from one another (SDG vs SNDG: 3 = 3.48, £ > .05; SDG vs RDG: g = .93, £ > .05; RDG vs SNDG: g = 2.01, £ > .05). Descriptiveness ratings The univariate ANOVA revealed a significant group x content interaction (F(6,255) •= 16.93, £ < .001) and was followed by Tukey-Kramer multiple comparisons. As expected, the two depressed groups rated the depression-content words as more self-descriptive than the nondepressed group (SDG vs SNDG: g = 10.67, £ < .05; RDG vs SNDG: 3 = 5.75, £ < .05) but the two depressed groups did not differ significantly from one another (3 = 3.36, £ > .05). The two depressed groups also rated the anxiety-content adjectives as more self-descriptive than the nondepressed group (SDG vs SNDG: 3 = 6.13, £ < .05; RDG vs SNDG: 3 - 4.17, £ < .05) but were not differentiate from one another (3 = .91, £ > .05). As expected, the nondepressed group described the positive-content words as more self-descriptive than the two depressed groups (SNDG vs SDG: 3 = 8.21, £ < .05; SNDG vs RDG: 3 - 5.65, £ < .05) while the two depressed groups did not differ from one another (3 = 1.15, £ > .05). As for the negative-content adjectives, the two depressed groups rated these as more self-descriptive (SDG vs SNDG: 3 = 5.46, £ < .05; RDG vs SNDG: g = 4.16, £ < .05) while the two depressed groups were not discernible from one another (g — .30, £ > .05). Reaction time The univariate ANOVA for the group x content condition interaction was significant (F(6,255) - 2.31, £ < .05). Tukey-Kramer multiple comparisons suggested 89 that the two depressed groups differed from the nondepressed group in their RTs to the depression-content words (SDG vs SNDG: g = 4.31, 2 < .05; RDG vs SNDG: g = 4.45, £ < .05) but did not differ significantly from one another (g = 1.13, £ > .05). No group differences were found for the anxiety-content condition (SDG vs SNDG: g = .67, £ > .05; SDG vs RDG: g - 1.18, £ > .05; RDG vs SNDG: g - 1.49, £ > .05) nor the negative-content condition (SDG vs SNDG: g = 2.53, £ > .05; SDG vs RDG: g - .17, £ > .05; RDG vs SNDG: g = 1.90, £ > .05) nor the positive-content condition (SDG vs SNDG: g = 2.50, £ > .05; SDG vs RDG: g = .78, £ > .05; RDG vs SNDG: g -.93, £ > .05). Recall Although the recall index was hypothesized to be a potential cognitive marker in the depressive self-schema, methodological difficulties with this measure (as discussed in Study 2) preclude interpretation. Consequently, this analysis was dropped and is not discussed. Discussion of time 1 SRET data As in Study 2, these results provided evidence of a depressive self-schema. Consistent evidence was present for a content-specific depression self-schema. On three self-schema indexes - the endorsement rate, descriptiveness ratings and RT -significant group differences were found between the two depressed groups and the nondepressed group in the processing of the depression-content adjectives. This depressive bias displayed by the depressed groups could not have been seen as a subset of a general negative bias since no group differences were found in two of the three interpreted SRET indexes (the exception is the descriptiveness ratings) for the negative-content condition. Additionally, this study provided some evidence of a positive self-schema, although the evidence was not as strong as for Study 2. Although group differences were not found on the RT index for the positive-content 90 adjectives, they were found for the other two SRET indexes. The issue of interest was to determine whether any differences existed between the- stable depressed, group and remitted depressed group at time 1. Stated differently, were there any aspects of the SRET which were predictive of remission? In every content condition of the endorsement rate, descriptiveness ratings, and RT indexes, the stable depressed group was indistinguishable from the remitted depressed group. Two implications follow. Theoretically, self-schemata, as measured by the SRET, provide little clue to the process of remission. Methodologically, the lack of distinctiveness in schematic processing of these two depressed groups implies that any group differences found in the time 2 analysis can not be attributable to initial group differences: SRET time 1-time 2 analyses The- design was a- multivariate three groups (between-subjects factor) x four content conditions (within-subjects factor) x two times- (within-subjects factor) split-plot factorial (Kirk, 1982). The results, of the multivariate analysis, are summarized in Appendix T. The crux: of the analysis lay in the group x content condition x time interaction which was significant (Wilks* lambda = .82, F(36,1100) — 1.45, £ < .05). The- multivariate analysis was followed by separate univariate ANOVAs for each SRET variable. The results are summarized in: Appendices U - X. Endorsement rate The follow-up univariate ANOVA yielded a: significant group x content condition x time interaction (F(6,255) - 3.52, £ < .01). Tukey-Kramer multiple comparisons across time and within group and. within content condition were computed following, the^  procedures outlined by Kirk (1968). As predicted, the stable depressed group had. stable endorsement rate across- time1 in all content conditions: depression-content condition, a ~ 1-50, ^  >- .05; anxiety-content condition, a 1«32, £ > .05; 91 negative-content condition, 3 = .80, £ > .05; and positive-content condition, 3 •= 3.02, £ > .05. Likewise, the stable nondepressed group had stable endorsements rates for all content conditions: depression-content condition, 3 = 1.13, £ > .05; anxiety-content condition, 3 = .07, £ > .05; negative-content condition, 3 = 1.63, E > .05; and positive-content condition, 3 •= .40, £ > .05. As for the remitted depressed group, the endorsement rate for the depression-content adjectives was significantly lowered (3 « 3.89, £ < .05). Additionally, the remitted depressed group had a significantly lowered endorsement rate for the anxiety-content adjectives at time 2 (3 - 5.43, £ < .05) and a significantly higher endorsement rate for the positive-content adjectives (3 = 3.98, E < .05). The endorsement rate for the negative-content adjectives remained constant (3 = .12, £ > -05) for the remitted depressed group. For those content conditions in which the remitted depressed group exhibited change, multiple comparisons were then conducted to determine the remitted depressed group's status with respect to the other two groups at time 2. For the depression-content condition, time 2 groupwise multiple comparisons indicated that the stable depressed group endorsed significantly more of the depression-content adjectives than the remitted depressed group (3 = 8.60, E < -05) and the stable nondepressed group (3 *» 13.84, E < .05). Additionally, the remitted depressed group continued to endorse more depression-content adjectives than the stable nondepressed group (3 = 3.84, £ < -OS). With respect to the anxiety-content condition, the remitted depressed group's endorsement rate was, at time 2, similar to the stable nondepressed group (3 = 2.51, £ > .05) and was significantly lower than the stable depressed group (3 - 5.94, E < .05). The stable depressed group continued to endorse significantly more of the anxiety-content adjectives than the stable nondepressed group (3 = 9.37, E k -05). Finally for the positive-content adjectives, the stable nondepressed group continued to endorse more 92 of these adjectives than the stable depressed group (3 = 11.49, £ < .05) and the remitted depressed group (a = 4.56, £ < .05), and the remitted depressed group endorsed more of these items than the stable depressed group (3 =» 5.54, £ < .05). Descriptiveness ratings The follow-up univariate ANOVA yielded a significant group x content condition x time effect (F(6,255) - 3.05, £ < .01). Tukey-Kramer multiple comparisons conducted across time and within group and within content condition suggested that the stable depressed group descriptiveness ratings remained constant for all four content conditions: depression-content condition, a = -08, £ > .05; anxiety-content condition, a = 2.97, £ > .05; positive-content condition, a ~ 1-62, £ > .05; and negative-content condition, g = 2.14, £ > .05. Likewise, the stable nondepressed group's descriptiveness ratings remained stable across the two testing sessions: depression-content condition, 3 = 1.98, £ > .05; anxiety-content condition, a = 2.90, £ > .05; negative-content condition, a = -71. £ > .05; and positive-content condition, 3 = .20, £ > .05. As expected, the remitted depressed group gave significantly lower descriptiveness ratings for the depression-content adjectives at time 2 (3 = 5.91, £ < .05) as well as for the anxiety-content adjectives (3 = 6.63, £ < .05). However, the remitted depressed group's descriptiveness ratings for the negative-content adjectives (3 = .00, £ > .05) and for the positive-content adjectives (3 •= 2.57, £ > .05) remained unchanged. For those content conditions in which the remitted depressed group exhibited change, multiple comparisons were then conducted using the time 2 data to determine the remitted depressed group's time 2 status with resepect to the other two groups. For the depression-content adjectives, the stable depressed group continued to rate these as more self-descriptive than the stable nondepressed group (3 = 14.72, £ < .05). The remitted depressed group's descriptiveness ratings of the depression-content adjectives were now significantly lower than the stable depressed 93 group (g = 11.47, p. < .05) and were now similar to the stable nondepressed group (a = 2.11, £ > .05). A similar pattern was obtained for the anxiety-content adjectives. The remitted depressed group gave significantly lower descriptiveness ratings to the anxiety-content adjectives than the stable depressed group (a = 6.36, £ < .05) and these ratings were similar to the stable nondepressed group (a = 2.78, £ > .05). The stable depressed group continued to see the anxiety-content words as more self-descriptive than the stable nondepressed group (a = 10.15, £ < .05). Reaction time The follow-up univariate ANOVA yielded a nonsignificant group x content condition x time interaction (F(6,255) =» .21, E > .05) and consequently multiple comparisons were not conducted. Recall This analysis was dropped (see previous discussion). Discussion The data generally supported the hypotheses. First of all, the two SRET self-description indexes have demonstrated stability across time and instability across remitted mood. When retested in consistent moods, the endorsement rate and descriptiveness ratings remained stable across the two testing sessions for the stable depressed group and the stable nondepressed group. However, when retested in inconsistent moods, these two self-description indexes have demonstrated sensitivity to mood changes. With the remission of depressed mood, the remitted depressed group endorsed fewer anxiety-content and depression-content items and endorsed more positive-content items. Additionally, these remitted subjects, when compared to their time 1 ratings, gave lower descriptiveness ratings to the anxiety-content and depression-content adjectives. Although the remitted depressed group gave stable responses to the negative-content adjectives, interpretation of this is difficult since 94 reliable group differences were not found for this condition at time 1. With the remitted mood, the remitted depressed group's status with respect to the nondepressed group is of interest. With respect to the time 2 anxiety-content data, group differences no longer existed on either the endorsement rate or the descriptiveness index. Although the remitted depressed group continued to endorse more depression-content items than the stable nondepressed group, the descriptiveness ratings were comparable for the two groups. However in the case of the positive-content words, dissimilarities still existed. Although there was a tendency for the remitted depressed group to endorse more positive-content words and to rate these as more descriptive at time 2, these ratings remained significantly lower than the stable nondepressed group's ratings. In summary, remitted subjects were roughly equivalent with the nondepressed subjects in their self-descriptions of the anxiety and depression items. However despite a remitted mood, remitted subjects still continued to describe themselves less positively. The second set of hypotheses predicted stability for the RT index. Group differences were found in the RT data at time 1 between the two depressed groups and the nondepressed group in the processing of the depression-content condition. Despite this initial difference, the remitted depressed group did not demonstrate change in this index with remitted mood. The argument could be raised that the sample used is an analogue one and as such, generalization to a clinically population may prove to be tenuous. Dobson and Shaw (in press) have conducted a similar study on a psychiatric sample. These investigators also used the SRET indexes of endorsement rate, descriptiveness ratings, RT, and recall on three groups of depressed patients, nondepressed psychiatric controls, and nonpsychiatric control patients. The endorsement rate and descriptiveness ratings were able to discriminate between the currently depressed 95 groups and the nondepressed group. However, evidence of schematic processing was not found for the RT and recall indexes. When the SRET was again administered to the depressed patients when their depressive symptoms had remitted, their endorsement rate and descriptiveness ratings of the depression-content adjectives had shifted to the point where the results were comparable to the nondepressed group. Thus the shift in the self-report indexes documented in a college sample has been replicated in a clinical sample. However the replication of the stability in the RT index still remains to be seen. Given the stability of the RT index, the issue can be raised: How stable is stable? The possibility exists that all SRET indexes are reactive to remitted mood but there is a desynchrony in the four indexes. It is possible that the two self-report measures are highly volatile indexes and any changes in mood are immediately reflected in these indexes. The RT index may also be sensitive to mood changes but this index changes at a much slower rate. This might be analogous to the desynchrony among the cognitive, behavioral, and psychophysiological channels in the reduction of fear (see Lang, 1967; Rachman & Hodgson, 1974). It would be interesting to reassess the remitted depressed subjects at some later time, assuming that they continued to remain nondepressed, to detect whether this index will change. If the RT index is truly stable, then these results suggest that the depressive self-schema is not merely a corollary to depressed mood. In other words, these results offer preliminary evidence in implicating cognitive structures as causal markers in depression. But the evidence is only preliminary. Since this study assessed individuals only after they had become depressed, the possibility exists that the depressive self-schema is a result of the depression. But even if this scar hypothesis is correct, it does not negate the possibility that the depressive self-schemata are markers to depression. Once the depressive self-schema is formed, according to 96 schema theory, it provides the basis for social information processing. If an individual with a depressive self-schema selectively attends to depression-related stimili, than that individual has a distinctive, perhaps depressogenic, manner of interpreting his or her social experiences. In this respect, it is possible that the depressive self-schema constitutes a cognitive vulnerability to future depressive episodes. It would be interesting to determine whether the presence of a depressive self-schema, independent of the presence of a depressed mood, would be predictive of relapse. The utility of the SRET as a cognitive assessment device As discussed in the introduction, the utility of the SRET as a cognitive assessment device was explored under two guidelines proposed by Shaw and Dobson (1981). This thesis attempted to further describe the cognitive phenomenon of depression by documenting the relationship between the depression self-schema and the anxiety self-schema. Study 2 suggests the potential inappropriateness of applying the SRET to the domain of anxiety in a college sample. This thesis further attempted to document stable cognitive aspects in the depression self-schema by applying the SRET across time and across remitted depressed mood. Study 3 has demonstrated that the test-retest stability of the SRET in two cases. With consistent moods, schematic processing on three SRET indexes remained unchanged over a three month period. With inconsistent moods, the RT index remained constant over the three month interval. One logical progression would be to apply the SRET to the other conditions proposed by Shaw and Dobson (1981). One possibility may be to ask whether the SRET would be instructive of differential change across therapies. Beck and his colleagues (Beck, Rush, Shaw, & Emery, 1979) theorized that cognitive therapy would produce permanent schematic changes. If this theory is correct, the SRET might 97 provide a method of documenting that change. In contrast, other forms of therapies such as pharmacotherapy, do not explicity theorize schematic changes. If the SRET could be used to document schematic changes via cognitive therapy, and if the SRET could be used to document schematic stability across pharmacotherapy, self-schemata research could potentially address the differential effects of various therapies. 98 References Abelson, R. P. (1976). Script processing in attitude formation and decision making. In J. S. Carroll & J. W. Payne (Eds.), Cognition and social behavior. Hillsdale, N.J.: Lawrence Erlbaum Associates. Alba, J. W. & Hasher, L. (1983). Is memory schematic? Psychological Bulletin. 93, 203-31. Alloy, L. B. & Abramson, L. Y. (1979). Judgment of contingency in depressed and nondepressed students: Sadder but wiser? Journal of Experimental Psychology:  General. 108. 441-85. Altman, J. H. & Wittenborn, J. R. (1980). Depression-prone personality in women. Journal of Abnormal Psychology. 89, 303-8. American Psychiatric Association (1980). Diagnostic and statistical manual of mental disorders - third edition. Washington, D.C: American Psychiatric Association. Beck, A. T. (1963). Thinking and depression. I. Idiosyncratic content and cognitive distortions. Archives of General Psychiatry. 9, 324-35. Beck, A. T. (1964). Thinking and depression. II. Theory and therapy. Archives of General Psychiatry. 10, 561-71. Beck, A. T. (1967). Depression: Clinical, experimental, and theoretical aspects. New York: Harper & Row. Beck, A. T. (1972). Depression: Causes and treatment. Philadelphia: University of Pennsylvania Press. Beck, A. T. (1976). Cognitive therapy and the emotional disorders. New York: International Universities Press. Beck, A. T., Emery, G. & Greenberg, R. L. (1985). Anxiety disorders and phobias: A Cognitive perspective. New York: Basic Books. Beck. A. T. & Hurvich, M. S. (1959). Psychological correlates of depression. I. Frequency of "masochistic" dream content in a private practice sample. Psychosomatic Medicine, 21, 50-55. Beck, A. T., Rush, A. J. , Shaw, B.F., & Emery, G. (1979). Cognitive therapy of depression. New York: Guilford Press. Beck, A. T. & Ward, C. H. (1961). Dreams of depressed patients: Characteristic themes in manifest content. Archives of General Psychiatry. 5, 462-67. Beck, A. T., Ward, C. H., Mendelson, M., Mock, J. , & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561-71. Bowen, R. C. & Kohout, J. (1979). The relationship between agoraphobia and primary affective disorders. Canadian Journal of Psychiatry. 24. 317-22. Broadbent, D. E. (1967). Word-frequency effect and response bias. Psychological Review. 74, 1-15. Broadbent, D. E. & Broadbent, M. H. P. (1975). Some further data concerning the word frequency effect. Journal of Experimental Psychology: General. 104, 297-308. Bradley, B. & Mathews, A. (1983). Negative self-schemata in clinical depression. British Journal of Clinical Psychology, 22, 173-81. Buglass, D., Clarke, J., Henderson, A. S., Kreitman, N., & Presley, A. S. (1977). A study of agoraphobic housewives. Psychological Medicine, 7, 73-86. Bumberry, W., Oliver, J. M., & McClure, J. C. (1978). Validation of the Beck Depression Inventory in a university population using psychiatric estimate as the criterion. Journal of Consulting and Clinical Psychology, 46(1), 150-55. Burns, D. D. (1980). Feeling Good. New York: Signet. Campbell, D. T. & Fiske, D. W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 55, 81-105. Cantor N. & Mischel, W. (1977). Traits as prototypes: Effects on recognition memory. Journal of Personality and Social Psychology. 35, 38-48. Cantor N. & Mischel, W. (1979). Prototypes in person perception. In L. Berkowitz (Ed.), Advances in experimental social psychology. New York: Academic Press. Carroll, J. B., Davies, P., & Richman, B. (1971). The American heritage word frequency book. Palo Alto: Houghton Mifflin Company. Catlin, J. (1969). On the word-frequency effect. Psychological Review. 76, 504-6. Coyne, J. C. (1976). Depression and the response of others. Journal of Abnormal Psychology, 85(2), 186-93. Coyne, J. C. & Gotlib, I. H. (1983). The role of cognition in depression: A critical appraisal. Psychological Bulletin. 94(3), 472-505. Craik, F. I. M. & Lockhart, R. S. (1972). Levels of processing: A framework for memory research. Journal of Verbal Learning and Verbal Behavior. 11, 671-84. Davis, H. (1979a). Self-reference and the encoding of personal information in depression. Cognitive Therapy and Research. 3, 98-111. Davis, H. (1979b). The self-schema and subjective organization of personal information in depression. Cognitive Therapy and Research. 3, 415-426. Davis, H. & Unruh, W. R. (1981). The development of self-schema in adult depression. Journal of Abnormal Psychology. 90, 125-33. Dealy, R. S., Ishiki, D. M., Avery, D. H., Wilson, L. G., & Dunner, D. L. (1981). Secondary depression in anxiety disorders. Comparative Psychiatry. 22, 612-18. Derogatis, L. R., Klerman, G. L., & Lipman, R. S. (1972). Anxiety states and depressive neurosis: Issues in nosological discrimination. Journal of Nervous and Mental Disease. 155. 392-403. Derogatis, L. R., Lipman, R. S., Covi, L., & Rickels, K. (1972). Factorial invariance of symptom dimensions in anxious and depressive neuosis. General Psychiatry. 27, 659-65. Derry, P. A. & Kuiper, N. A. (1980). Schematic processing and self-reference in clinical depression. Unpublished manuscript, University of Western Ontario, London, Ontario. Derry, P. A., & Kuiper, N. A. (1981). Schematic processing and self-reference in clinical depression. Journal of Abnormal Psychology. 90, 286-97. Dobson, K. S. (1985a). An analysis of anxiety and depression scales. Journal of Personality Assessment. 49(5),522-27. Dobson, K. S. (1985b). The relationship between anxiety and depression. Clinical Psychology Review. 5, 307-24. Dobson, K. S. & Block, L. (in press). Historical and philosophical bases of the cognitive - behavioral therapies. in K. S. Dobson (Ed.). Handbook of cognitive-behavioral therapies. New York: Guilford Press. Dobson, K. S. & Breiter, H. S. (1983). Cognitive assessment of depression: Reliability and validity of three measures. Journal of Abnormal Psychology. 92(1), 107-09. Dobson, K. S. & Shaw, B. F. (1986). Cognitive assessment with major depressive disorders. Cognitive Therapy and Research. 10(1). 13-29. Dobson, K. S. & Shaw, B. F. (in press). The specificity and stability of self-referent encoding in clinical depression. Journal of Abnormal Psychology. Downings, R. W. & Rickels, K. (1974). Mixed anxiety depression: Fact or myth? Archives of General Psychiatry. 30. 312-17. Eaves, G. & Rush, A. J. (1984). Cognitive patterns in symptomatic and remitted unipolar major depression. Journal of Abnormal Psychology. 93(1). 31-40. Edwards, A. L. (1957). The social desirability variable in personality assessment and research. New York: Dryden. Epstein, S. (1973). The self-concept revisited: Or a theory of a theory. American Psychologist. 28, 404-16. Frederiksen, N. (1965). Response set scores as predictors of performance. Personnel Psychology. 18, 225-44. Frederiksen, J. R. (1971). Statistical decision model for auditory word recognition. Psychological Review. 78, 409-19. Ganellen, R. J . & Carver, C. S. (in press). Why does self-reference promote incidental encoding? Journal of Experimental Social Psychology. Garber, J. , Miller, S. M. & Abramson, L. Y. (1980). On the distinction between anxiety and depression: Perceived control, certainty and probability of goal attainment. In J. Garber & M. E. P. Seligman (Eds.). Human helplessness: Theory and application. Toronto: Academic Press. Gotlib, I. H. & Robinson, L. A. (1982). Responses to depressed individuals: Discrepancies between self-report and observer-rated behavior. Journal of Abnormal Psychology. 91(4), 231-40. Gough, H. G. & Heilbrun, A. B. (1965). The Adjective Checklist manual. Palo Alto: Consulting Psychologist Press. Gray, J. A. (1982). The neuropsychology of anxiety: An enquiry into the functions of the septo-hippocampal system. New York: Clarendon Press. Gurney, C , Roth. M., Garside, R. F., Kerr, T. A., & Schapira, K. (1972). Studies in the classification of affective disorders: The relationship between anxiety states and depressive illness: II. British Journal of Psychiatry. 11, 162-66. Hakstian, A. R. (1980). One-way multivariate analysis of variance. (Computer program). University of British Columbia. Hamilton, E. W. & Abramson, L. Y. (1983). Cognitive patterns and major depressive disorders: A longitudinal study in a hospital setting. Journal of Abnormal Psychology. 92(2), 173-84. 103 Hammen, C. L. (1980). Depression in college students: Beyond the Beck Depression Inventory. Journal of Consulting and Clincial Psychology, 48(1), 126-8. Hammen, C. L. & Zupen, B. A. (1984). Self-schemas, depression, and the processing of personal information in children. Journal of Experimental Child Psychology. 37. 598-608. Izard, C. E. (1972). Patterns of emotions: A new analysis of anxiety and depression. New York: Academic Press. Izard, C. E. (1977). Human emotions. New York: Plenum Press. Johnson, E. J. & Tversky, A. (1983). Affect, generalization, and the perception of risk. Journal of Personality and Social Psychology, 45. 20-31. Kazdin, A. E. (1980). Research design in clinical psychology. New York: Harper & Row. Kendziersk, D. (1980). Self-schemata and scripts: The recall of self-referent and scriptal information. Personality and Social Psychology Bulletin. 16, 23-29. Kirk, R. E. (1968). Experimental design: Procedures for the behavioral sciences. (1st ed.) Montery, California: Brooks. Kirk, R. E. (1982). Experimental design: Procedures for the behavioral sciences. (2nd edition) Belmont, California: Wadsworth. Klerman, G. L. (1977). Anxiety and depression. In G. D. Burrows (Ed.), Handbook of studies on depression. New York: Excerpta Medica. Kolb, L. C. (1973). Modern clinical psychiatry. Philadelphia: W. B. Saunders. Krantz, S. & Hammen, C. (1979). Assessment of cognitive bias in depression. Journal of Abnormal Psychology. 88, 611-19. Kuhn, T. S. (1970). The structure of scientific revolution (2nd ed.). Chicago: The University of Chicago Press. Kuiper, N. A. (1978). The self as an agent in the processing of personal information about others. Unpublished doctoral dissertion. The University of Calgary. Kuiper, N. A. & Derry, P. A. (1980). Content-specificity in the processing of personal information of depressives and nondepressives. Unpublished manuscript. The University of Western Ontario, London, Ontario. Kuiper, N. A. & MacDonald, M. R. (1983). Schematic processing in depression: The self-based consensus bias. Cognitive Therapy and Research. 7(6), 469-84. Kuiper, N. A. & Olinger, L. J. (1986). Stress and cognitive vulnerability to depression: A self-worth contengency model. In R. W. J. Neuf eld (Ed.), Advances in the investigation of psychological stress. New York: Wiley. Kuiper, N. A., Olinger, L. J. , & MacDonald, M. R. & Shaw, B. F. (1984). Self-schema processing of depressed and nondepressed content: The effects of vulnerability to depression. Social Cognition. 3(1), 77-93. Kuiper, N. A. & Rogers, T. B. (1979). Encoding of personal information: Self-other differences. Journal of Personality and Social Psychology. 13. 119-24. Kuiper, N. A. & Rogers, T. B. (1981). Convergent envidence for the self as a prototype: The "inversted-U RT effect" for self and other judgments. Personality and Social Psychology Bulletin. 7, 438-43. Laird, C. (1974). New world thesaurus. New York: Popular Library. Lang, P. J. (1967). Fear reduction and fear behavior: Problems in treating a construct. In J. M. Schlien (Ed.), Research in Psychotherapy(Vol III), Washington: American Psychiatric Association. Lewinshohn, P. M., Mischel, W., Chaplin, W., & Barton, R. (1980). Social competence and depression: The role of illusory self-perceptions. Journal of Abnormal Psychology. 89(2), 203-12. Markus, H. (1977). Self-schemata and processing information about the self. Journal of Personality and Social Psychology. 35(2). 63-78. Markus, H., Crane, M., & Siladi, M. (1978). Cognitive consequences of androgyny. Paper presented at the meeting of the Midwestern Psychological Association. Markus, H., Hamill, R., & Sentis, K. (1979). Thinking fat: Self-schemas for body weight and the processing of weight-relevant information. Paper presented at the meeting of the American Psychological Association, New York. Markus, H. & Sentis, K. (1982). The self in social information processing. In J. Suis (Ed.), Psychological perspectives on the self. (Vol 1). Hillsdale, N.J.: Lawrence Erlbaum Associates. Markus, H. & Smith, J. (1981). The influence of self-schemata on the perception of others. In N. Cantor & J. I. Kihlstrom (Eds.), Personality, cognition, and social  interaction. Hillsdale, N.J.: Lawrence Erlbaum Associates. Marshall, A. G. & Izard,. C. E. (1972). Depression as a pattern of emotions and feelings: Factor-analytic investigations. In C. E. Izard (Ed.), Patterns of  emotions: A new analysis of anxiety and depression. New York: Academic Press. Mendels, J . Weinstein, N., & Cochrane, C. (1972). The relationship between depression and anxiety. Archives of General Psychiatry. 27, 649-53. Miller. W. R. (1975). Psychological deficits in depression. Psychological Bulletin. 82, 238-60. Minsky, M. (1975). A framework for representing knowledge. In P. Winston (Ed.). The psychology of computer vision. New York: McGraw-Hill. Mueller, J. H., Bailis, K. L., & Goldstein, A. G. (1982). Depth of processing and anxiety in facial recognition. British Journal of Psychology. Mullaney, J. A. (1984). The relationship between anxiety and depression: A review of some principal component analytic studies. Journal of Affective Disorders. 7, 139-48. Nakatani, L. H. (1973). On the evaluation of models for the word-frequency effect. Psychological Review. 80, 195-202. Oliver, J. M. & Baumgart, E. P. (1985). The Dysfunctional Attitude Scale: Psychometric properties and relation to depression in an unselected adult popuation. Cognitive Therapy and Research. 9(2), 161-67. Oliver, J. M. & Burkham, J. (1979). Depression in university students: Duration, relation to calender time, prevalence, and demographic correlates. Journal of Abnormal Psychology. 88(6), 667-78. Paivio, A. (1965). Abstractness, imagery, and meaningfulness in paired-associate learning. Journal of Verbal Learning and Verbal Behavior. 4, 32-38. Paivio, A. (1971). Imagery and verbal processes. New York: Holt, Rinehart & Winston. Paivio, A., Yuille, J. C , & Madigan, S. A. (1968). Concreteness, imagery, and meaningfulness values for 925 nouns. Journal of Experimental Psychology. Monograph Supplement. 76 (1, Part 2). Pietromonaco, P. R. & Markus, H. (1985). The nature of negative thoughts in depression. Journal of Personality and Social Psychology. 48(3), 799-807. Prusoff, B. & Klerman, G. L. (1974). Differentiating depressed and anxious neurotic outpatients. Archives of General Psychiatry. 30, 302-17. Rachman, S. J. & Hodgson, R. (1974). Synchrony and desynchrony in fear and avoidance. Behavior Research and Therapy. 12, 311-18. < Rehm, L. P. (1981). Assessment of depression. In M. Hersen & A. S. Bellack (Eds.), Behavioral assessment: A practical handbook (2nd ed.). Toronto: Pergamon Press. Rogers, T. B. (1977). Self-reference in memory: Recognition of personality items. Journal of Research in Personality. 11. 295-305. Rogers, T. B. (1981). A model of the self as an aspect of the human information processing system. In N. Cantor & J. I. Kihlstrom, (Eds.), Personality. 107 cognition, and social interaction. Hillsdale, N.J.: Lawrence Erlbaum Associates. Rogers, T. B., Kuiper, N. A., & Kirker, W. S. (1977). Self-reference and the encoding of personal information. Journal of Personality and Social Psychology, 35, 677-88. Rogers, T. B., Rogers, P. J . , & Kuiper, N. A. (1979). Evidence for the self as a cognitive prototype: The "false alarms effect". Personality and Social Psychology Bulletin. 5, 53-56. Rosch, E. (1973). On the internal structure of perceptual and semantic categories. In T. E. More (Ed.), Cognitive development and acquisition of language. New York: Academic Press. Roth, M., Gurney, C , Garside, R. F., & Kerr, T. A. (1972). Studies in the classification of affective disorders: The relationship between anxiety states and depressive illness - I. British Journal of Psychiatry, 121, 147-161. Rush, A. J . , Beck, A. T., Kovacs, M., & Hollon, S. (1977). Comparative efficacy of cognitive therapy and pharmacotherapy in the treatment of depressed outpatients. Cognitve Therapy and Research. 1(1), 17-37. Russell, J. A. (1980). A circumplex model of affect. Journal of Personality and Social Psychology. 39(6), 1161-78. Schank, R. C. & Abelson, R. (1977). Scripts, plans, goals, and understanding. Hillsdale, N.J. Lawrence Erlbaum Associates. Schapira, K., Roth, M., Kerry, T. A., & Gurney, C. (1972). The prognosis of affective disorders: The differentiation of anxiety states from depressive illness. British Journal of Psychiatry. 121. 175-81. Schuberth, R. E., Spoehr, K. T., & Lane, D. M. (1981). Effects of stimulus and contextual information on the lexical decision process. Memory and Cognition. 9, 68-77. 108 Shaw, B. F. (1977). Comparison of cognitive therapy and behavior therapy in the treatment of depression. Journal of Consulting and Clinical Psychology. 45, 543-51. Shaw, B. F. & Dobson, K. S. (1981). Cognitive assessment of depression. In T. V. Merluzzi, C. R. Glass, & M. Genest (Eds.), Cognitive assessment. New York: Guilford Press. Silverman, J. S., Silverman, J. A., & Eardley, D. A. (1984). Do maladaptive attitudes cause depression? Archives of General Psychiatry. 41. 28-30. Simons, A. D., Garfield, S. L., & Murphy, G. E. (1984). The process of change in cognitive therapy and pharmacotherapy for depression. Archives of General  Psychiatry. 41, 45-51. Simons, A. D., Murphy, G. E., Levine, J. L., & Wetzel, R. D. (1984). Substained  improvement one year after cognitive and/or pharmacotherapy of depression. Paper presented at the Meeting of the Society for Psychotherapy Research, Lake Louise, Alberta. Spielberger, C. D., Gorsuch, R. R., & Lushence, R. E. (1970). State-trait anxiety inventory manual for form X. Palo Alto: Consulting Psychologists Press. Tellegen, A. (1985). Structures of mood and personality and their relevance to assessing anxiety with an emphasis on self-report. In A. H. Tuma & J. D. Maser (Eds.), Anxiety and the anxiety disorders. Hillsdale, N.J.: Lawrence Erlbaum Associates. Unde, T. W., Boulenger, J. P., Roy-Byrne, P. P., Geraci, M. F., Vittone. B. J. , & Post, R. M. (1985). Longitudinal course of panic disorders: Clinical and biological considerations. Progress in Neuro-Psychopharmacology and Biological  Psychiatry. 9, 39-51. Videbech, T. (1975). The psychopathology of endogeneous depression. Acta Psychiatrica Scandinavica. 52, 336-73. Weissman, A. N. (1980). Assessing depressionogenic attitudes: A validation study. Paper presented at the 51st annual meeting at the Eastern Psychological Association, Hartford, Connecticut. Weissman, A. N. & Beck, A. T. (1978). Development and validation of the Dysfunctional Attitude Scale: A preliminary investigation. Paper presented at the annual meeting of the American Educational Research Association, Toronto. Wilkinson, I. M. & Blackburn, I. M. (1981). Cognitive style in depressed and recovered depressed patients. The British Journal of Clinical Psychology, 20, 283-91. Wurf, E. & Markus, H. (1982). Cognitive consequences of the negative self. Unpublished manuscipt. University of Michigan. Zevon, M. A. & Tellegen, A. (1982). The structure of mood change: an idiographic/nomothetic analysis. Journal of Personality and Social Psychology. 43, 111-22. Appendix A Ratings of Adjectives Questionnaire Ratings of Adjectives Study 111 We are interested in studying a number of properties of 62 adjectives. To do so, we are asking you to rate each word on five properties: anxiety, depression, emotional intensity, social undesirability, and imagery. Following is an explanation of the task and a description of each of the property. I. Anxiety For this property, you will rate the degree to which each adjective describes anxiety. Anxiety is an emotion characterized by a feeling of vague, diffuse fear. The anxious individual often feels jittery, impatient, and "on edge". Accompanying physical signs are varied but include trembling, sweating, rapid heart rate, lightheadedness, clammy hands, discomfort in the pit of the stomach, and/or a lump in the throat. II. Depression For this property, you will rate the degree to which each adjective describes depression. Depression is an emotion characterized by feelings of sadness. The depressed individual may have a loss of interest in his/her usual activities. The depressed individual may suffer from a loss of appetitie, decreased energy levels, disturbances in sleeping and eating patterns, difficulties in concentration and/or feelings of worthlessness or guilt. III. Emotional intensity The meaning of certain words may carry emotional overtones. These emotional overtones may be positive or negative, and may or may not include anxiety and depression. Also different words arouse different intensities of emotions. So to rate this property, you will rate how intense these emotional overtones are. IV. Social desirability Each of these adjectives in this task is descriptive of a characteristic found in others. For this property, you will rate the characteristics in terms of how undesirable you consider these to be in others. V. Imagery Some words are capable of arousing a mental image - a mental picture, or sound, or other sensory experience. Rating this property involves rating the ease with which the word arouses mental images. With the above descriptions in mind, your ratings on each of the five properties will be on a 5-point scale as follows: 1 2 3 4 5 very slightly slightly moderately considerably very strongly or not at all So for the first judgement of anxiety, if you think that the adjective is very descriptive of the construct anxiety, then you would rate that a "5". If you think that it is only moderately descriptive of anxiety, then you would assign it a "3", and so on. 112 Feel free to use the entire range of numbers, from 1 to 5; at the same time don't be concerned about how often you use a particular number, as long as it reflects your true judgement. Work fairly quickly but do not be careless in your ratings. To summarize, you will rate 62 adjectives on five properties, each property on a 5-point scale. 1 2 3 4 5 very slightly slightly moderately considerably very strongly Judgement i: To what extent does the word describe anxiety? Judgement ii: To what extent does the word describe depression? Judgement iii: To what extent does the word arouse emotional overtones? Judgement iv: To what extent is this characteristic undesirable in others? Judgement v: To what extent does this word arouse a mental image? If necessary, refer back to these instructions when rating the words on the following pages. If there are any questions, ask them now. Otherwise, turn the page and begin. Thank you in advance for your cooperation. E. Cheung & K. Dobson 113 Subject number: Age: Sex: Please complete the following judgments: i. To what extent does the word describe anxiety? ii. To what extent does the word describe depression? iii. To what extent does the word arouse emotional overtones? iv. To what extent is this characteristic undesirable in others? v. To what extent does this word arouse a mental image? 1 2 3 4 5 very slightly slightly moderately considerably very strongly or not at all Word Judgment i ii iii iv v 1. fearful 2. excited 3. jittery 4. gifted 5. sluggish 6. ugly 7. lifeless 8. despondent 9. dreary 10. alert 11. thriving 12. downbeat 13. dishonest 14. defeated 15. stubborn 16. depressed 17. hyperactive 18. attentive 19. immature 20. assertive / / 114 Please complete the following judgments: i. To what extent does the word describe anxiety? ii. To what extent does the word describe depression? iii. To what extent does the word arouse emotional overtones? iv. To what extent is this characteristic undesirable in others? v. To what extent does this word arouse a mental image? 1 2 3 4 5 very slightly slightly moderately considerably very strongly or not at all Word Judgment i ii iii iv v 21. loved 22. dismal 23. dejected 24. thriving 25. enraged 26. anxious 27. spirited 28. blue 29. greedy 30. jumpy 31. apprehensive 32. messy 33. vain 34. tense 35. mean 36. shallow 37. industrious 38. happy 39. nervous 40. sensitive 115 Please complete the following judgments: i. To what extent does the word describe anxiety? ii. To what extent does the word describe depression? iii. To what extent does the word arouse emotional overtones? iv; To what extent is this characteristic undesirable in others? v. To what extent does this word arouse a mental image? 1 2 3 4 5 very slightly slightly moderately considerably very strongly or not at all Word Judgment i ii iii iv v 41. desolate 42. confident 43. panicky 44. amorous 45. rested 46. immoral 47. productive 48. mournful 49. cheerful 50. sad 51. downcast 52. afraid 53. secure 54. homely 55. rested 56. sneaky 57. conceited 58. repulsive 59. dynamic 60. worried 61. hostile 62. sorrowful Appendix B Groupwise multiple comparisons for each attribute of the four content conditions 117 Groupwise multiple comparisons for each attribute of the four content condition. Variable Groupwise comparison 95% confidence interval Word length A vs D -3.63 3.63 A vs N -2.91 4.34 A vs p -3.70 3.55 D vs N -2.91 4.34 D vs p -3.70 3.55 N vs p -4.41 2.84 Word frequency A vs D -483.65 488.22 A vs N -477.14 494.72 A vs p -478.93 492.93 D vs N -479.43 492.43 D vs p -481.22 490.64 N vs p -487.72 484.14 Ratings of anxiety A vs D 1.24 2.70 * A vs N 1.58 3.04 * A vs P 1.77 3.24 » D vs N -0.39 1.07 D vs P -0.20 1.26 N vs P -0.54 0.92 Ratings of depression A vs D -2.97 -1.66 * A vs N -0.27 1.04 A vs P 0.16 1.46 * D vs N 2.05 3.36 * D vs P 2.47 3.78 . * N vs P -0.23 1.08 Ratings of emotional intensity A vs D -1.11 1.14 A vs N -0.34 1.91 A vs P -0.44 1.80 D vs N -0.35 1.90 D vs P -0.46 1.79 N vs P -1.23 1.01 Ratings of social undesirability A vs D -1.49 0.30 A vs N -1.67 0.12 A vs P 0.82 2.61 * D vs N -1.07 0.71 D vs P 1.42 3.20 * N vs P 1.60 3.38 » Ratings of imagery A vs D -0.67 0.76 A vs N -0.53 0.91 A vs P -0.34 1.10 D vs N -0.57 0.86 D vs P -0.38 1.05 N vs P -0.53 0.91 * E < .05 A - anxiety-content adjectives D - depression-content adjectives N - negative-content adjectives P - positive-content adjectives Appendix C Beck Depression Inventory Beck Inventory 120 On this questionnaire are groups of statements. Please reed each group of statements careful ly . Then pick out the one statement in each group which best describes the way you have been feeling the PAST WEEK. INCLUDING TODAY; C i r c l e the number beside the statement you picked. If several statements in the group seem to apply equally well , c i r c l e each one. Be sure to read a l l  the statements in each group before making your choice. 1. 0 I do not feel sad. 1 I feel sad. 2 I am sad a l l the time and I can't snap out of i t . 3 I am so sad or unhappy that I can't stand i t . 2. 0 I am not part icularly discouraged about the future. 1 I fee l discouraged about the future. 2 I feel I have nothing to look forward to. 3 I feel that the future i s hopeless and that things cannot Improve. 3. 0 I do not feel l i k e a fa i lure . 1 I feel I have fa i led more than the average person. 2 As I look back on my l i f e , a l l I can see i s a lot of fa i lures . 3 I fee l I am a complete fa i lure as a person. 4. 0 I get as much sat isfact ion out of things as I used to. 1 I don't enjoy things the way I used to. 2 I don't get real satisfaction out of anything anymore. 3 I am dissat i s f ied or bored with everything. 5. 0« I don't feel part icularly gui l ty . 1 I fee l guilty a good part of the time. 2 I feel quite gui l ty most of the time. 3 I feel guilty a l l of the time. 6. 0 I don't feel I am being punished. 1 I feel I may be punished. 2 I expect to be punished. 3 1 feel I am being, punished. 7. 0 I don't feel disappointed in myself. 1 X am disappointed:in myself. 2 I am disgusted with myself. 3 I hate myself. 8. 0 I don't feel I am any worse than anybody else. 1 I am c r i t i c a l of myself for my weaknesses or mistakes. 2 I blame myself a l l the time for my faul ts . 3 I blair.3 myself for everything bad that happens. 9. 0 I don't have any. thoughts of k i l l i n g myself. 1 I have thoughts of k i l l i n g myself, but I would not carry them out. 2 I would l ike to k i l l myself. 3 I would k i l l myself i f I had the chance. 10. 0 I don't cry anymore than usual. 1 I cry more now than I used to. 2 I cry a l l the time now. 3 I used to be able to cry, but now I can't cry even though I want to. 121 11. 0 I eon no more Irri tated now than I ever am. 1 I get annoyed or i r r i t a t e d more easi ly than I used to. 2 I feel i r r i t a t e d a l l the time nov. 3 I don't get i rr i ta ted at a l l by the things that used to i r r i t a t e ne. 12. 0 I have not lost interest in other people. 1 I so less interested in other people than I used to be. 2 I have lost most of my interest in other people. 3 I have lost a l l of my interest in other people. 13. 0 I make decisions about as well as I ever could. 1 I put off making decisions more than I used to. 2 I have greater d i f f i cu l ty In making decisions than before. 3 I can't make decisions at a l l anymore. 14. 0 I don't feel I look any worse than I used to. 1 I am worried that I am looking old or unattractive. 2 I feel that there are permanent changes in my appearance that make me look unattractive. 3 I believe that I look ugly. 15. 0 I can work about as well as before. 1 It takes an extra effort to get started at doing something. 2 I have to push myself very hard to do anything. 3 I can't do any work at a l l . 16. 0 I can sleep as well as usual. 1 I don't sleep as well as I used to. 2 I wake up 1-2 hours ear l ier than usual and find i t hard to get back to sleep. 3 I wake up several hours ear l ier than I used to and cannot get back to sleep, 17. 0 I don't get more t ired than usual. 1 I get t i red more easi ly than I used to. 2 I get t i red from doing almost anything. 3 I am too t ired to do anything. 18. 0 My appetite i s no worse than usual 1 My appetite i s not as good as i t used to be. 2 My appetite i s much worse now. .3. I have no appetite at a l l anymore. 19. 0 I haven't lost much weight, i f any late ly . 1 I have lost more than 5 pounds. I am purposely trying to 2 I have lost more than 10 pounds. lose weight by eating less . 3 I have lost more than 15 pounds. Yes No 20. 0 , I am no more worried about my health than usual. 1 I am worried about physical problems such as aches and pains; or upset stomach; or constipation. 2 I am very worried about, physical, problems and i t ' s hard to think of much else. ~ 3 I am so worried about my physical problems, that I cannot think about anything else. 21. 0 I have not noticed any recent change in my interest in sex. 1 I am less interested in sex than I used to be. 2 I am much less interested in sex now. 3 I have lost interest in sex completely. 122 Appendix D Dysfunctional Attitude Scale n&<3 123 FORM A This Inventory l i s t s different attitudes or beliefs which people sometimes hold. Read EACH statement carefully and decide how much you agree or disagree with the statement. For each of the attitudes, show your answer by placing a chectaaarJc (/) under the column that BEST DESCRIBES HOW YOU THINK. Be sure to choose only one answer for each attitude. Because people are different, there is no right answer or wrong answer to these statements. To decide whether a given attitude i s typical of your way of looking at things, slaply keep in mind what you are like MOST OF THE TIME. EXAMPLE: a * o s i ATTITUDES j AGREE VE! ! MUCH OS 1 I l o sa 01 CO M Q if CO 3 o i l 1. Moat people are O.K. once you gat te know them. * • took at the exawple above. To show how much a sentence describee your attitude, you can ehedt any point from totally agree to totally disagree. Zn the above exaasl©, the oaieckxarljc at "agree slightly" indicatea that this statement i s seaewhat typical of the attitudes held by the person ooapleting the inventory. SeMMber that your answer should describe the way you think HOST OF THE TIMS. MOW TURN THE PAGE AND BEGIN Copyright 1978 by Arlene H. Weissman DAS 124 1: 1 ATTITUDES TOTALLY AGREE AGREE VERY MUCH I X a < i 1 1 i 3 ' •1 1 £ i M I * d C n <* o M o \ M a REMEMBER, ANSWER BACH STATEMENT ACCORDING TO THE titt 700 THINK MOST OF THE TIME. 1. It i s d i f f i c u l t to be happy unless one i s good looking, intelligent, r i d ! and creative. • f 2. Bappinsss is sore & natter of *y attitude towards ay self than the way other people feel aJbottt mm. • 3. People will probably think lees of aw i f 4. If I do not do well a l l the tins, people w i l l not reopect ate. * 4 5. Taking even a ssall risk is foolish because ' the loss is likely to be a disaster. \ • 6. It i s possible to gain another person's respect without being especially talented at anything. 7. I cannot be happy unless most people I know adaire ae. 8. If a person asks for help, i t is a sign of weakness. 1 125 ATTITUDES I M s AGREE SLIGHTLY | DISAGREE SLIGHTLY DISAGREE VERY MUCH TOTALLY DISAGREE 9. If I do not do as wall as other people, i t mane I am en inferior hunan being. 10. Zf Z f a i l at ay work, then l a i failure aa a person. i U . Zf you cannot do .something well, there is l i t t l e point in doing i t at a l l . *• 12. Making mistakes i s fine because I can learn fresi tfafav i • • • 13. Zf someone disagrees with a*, i t probably indicates be does not like me. 14. Zf I f a l l partly, i t is as bad as being a complete failure. f IS. Zf other people know what you are really like, they w i l l think less of you. V • 16. Z am nothing i f a person I love doesn't love • float e 17. On* can get pleasure from an activity regardless of the end. result. • 18. People should have a reasonable likelihood of success before undertaking'anything. • 126 ATTITUDES TOTALLY AGREE I i CO I *9 1 s SB 13 OH M 1 < tq M e i ca w a • •" OT a 1 19. My valu« as a parson depends greatly on what others think of at. 30. If I don't set the highest standards for myself, I an likely to end up a second-rats person. 21. If I am to be a worthwhile person, I oust bo truly outstanding in at least one major .respect. 22. People who have good ideas are sore worthy than those who do not. • • 23. I should be upset i f I make a mistake. 24. My own opinions of myself are more important than other's opinions of me. _ 25. To be a good, moral, worthwhile person, I must help everyone who needs i t . !' 28. If I ask a question, i t makes me look inferior. 27. It i s awful to be disapproved of by people important to you. i 28. If you don't have other people to lean on, you are bound to be sad. • 127-ATTITUDES i AGBEE VERY MUCH AGREE SLIGHTLY NEUTRAL DISAGREE SLIGHTLY DISAGREE VERY MUCH TOTALLY DISAGREE 29. I can reach important .goaiu without slave driving myself. 30. It is possible for a person to be scolded and not get upset. 31. I cannot trust other people because they night be cruel to me. 32. -If others dislike you, you cannot be happy. 33. It is best to give up your own interests . in order to please other people. 34. My happiness depends more on other people than i t does on ne. 35. Z do not need the approval of other people in order to be happy. * 36.. If a person avoids problems, the probiests tend to go away. 37. X can be happy even i f I miss out on many of the good things in l i f e . 38. What other.people think about me is very Important. 39. Being isolated from others l a bound to lead to unhappiness. 40. Z can find happiness without being loved by another person. Appendix E State-Trait Anxiety Inventory SELF-EVALUATION QUESTIONNAIRE Developed by C. D, Spielberger, R. L. Gorsuch and R. Lushene STAI FORM X-1 NAME DATE DIRECTIONS: A number of statements which people have used to describe themselves are given below. Read each state- • ment and then blacken in the appropriate circle to the right of the statement to indicate how you feel right now, that is, at this moment. There are no right or wrong answers. Do not spend too much time on any one statement but give the answer which seems to describe your present feelings best. NOT AT ALL SOMEWHAT MODERATELY SO VERY MUCH SO © © ® © © © ® © © © © © © © © © © © © © © © © © © © ® © © © © © © © © © © © © © © © ® © © ® ® © © © ® © © ® ® © © ® ® © © ® © © © ® © © © ® © © © ® ® © © ® ® © 20 S E L F - E V A L U A T I O N Q U E S T I O N N A I R E STAI FORM X-2 130 NAME DATE DIRECTIONS: A number of statements which people have used to describe themselves are given below. Read each state- > r ment and then blacken in the appropriate circle to the right of . g § the statement to indicate how you generally feel. There are no | o 3 right or wrong answers. Do not spend too much time on any a: 3 o * one statement but give the answer which seems to describe 3 S | > how you generally feel. » 5 5 21. I feel pleasant © © ® © 22. I tire quickly © © . © © 23. I feel like crying © ® © © 24. I wish I could be as happy as others seem to be © ® ® © 25. I am losing out on things because I can't make up my mind soon enough.... © ® ® © 26. I feel rested © © © © 27. I am "calm, cool, and collected" © © ® © 28. I feel that difficulties are piling up so that I cannot overcome them © ® ® © 29. I worry too much over something that really doesn't matter © © ® © 30. I am happy © © ® © 31. I am inclined to take things hard © ® ® © 32. I lack self-confidence © ® © © 33. I feel secure © ® ® © 34. I try to avoid facing a crisis or difficulty <p ® ® © 35. I feel blue :. © ® © © 36. I am content , © © ® © 37. Some unimportant thought runs through my mind and bothers me-. © ® ® © 38. I take disappointments so keenly that I can't put thorn out of my mind .... © ® ® ® 39. I am a steady person © ® ® © 40. I get in a state of tension or turmoil as I think over my recent concerns and interests © © © ® 21 Appendix F Program listing for the self-referent encoding task Program listing for the self-referent encoding task 70 words% - 56 80 insec = 146! 90 onesecond% - 1580 100 seconds3% = 3 * onesecond% / 2! 110 halfsecond% = cint(onesecond% / 2) 120 dim reactime<words%), etansS(words%), etrateS(words%), prmpt$(words%) 130 dim ignor$(4), order%(words%) 140 dim ir(4) 150 dim r2(56), a%(14), d%(14), nc%(14),pc%(14),b%(4) 160 dim etreact(4),etrating%(4), etones(4), type%(words%) 170 for i% - 1 to 4 180 etreact(i%) - 0: etrating%(i%) - 0; etones(I%) - 0 190 next i% 200 for i% = 1 to words%: order%(i%): next i% 220 fileS - "" 230 for i% - 1 to 14 240 a%(i%) - i% + 28 250 d%(i%) - i% + 14 260 nc%(i%) - i% 270 pc%(i%) - i% + 42 280 next i% 290 input "enter name of data file;", fileS 300 input " enter random seed.", seed! 310 randomize seed! 320 for i% - 14 to 2 step -1 330 j% - int(rnd*i%): if j% - 0 then goto 330 340 temp% - a%(i%) 350 a%(i%) - a%(j%) 360 a%(j%) - temp% 370 next i% 380 for i% = 14 to 2 step -1 390 i% - int(rnd*i%): if j% - 0 then goto 390 400 temp% = d%(i%) 410 d%(i%) = d%(j%) 420 d%(j%) - temp% 430 next i% 440 for i% - 14 to 2 step -1 450 j% = int(rnd*i%): if j% - 0 then goto 450 460 temp% = nc%(i%) 470 nc%(i%) - nc%(j%) 480 nc%(J%) - temp% 490 next % 500 for i% - 14 to 2 step -1 510 j% - int(rnd*i%): if j%*=0 then goto 510 520 temp% = pc%(i%) 530 pc%(i%) - pc%(J%) 540 pc%(j%) = temp% 550 next i% 133 560 c% - 0 570 for i% - 1 to 14 580 for j% - 1 to 4 590 b%(j%) - j% 600 next j% 610 for i2% - 4 to 2 step -1 620 j% - int(rnd*i2%): if j% - 0 then goto 620 630 temp% - b%(i2%) 640 b%(i2%) - b%(j%) 650 b%(j%) = temp% 660 next i2% 670 for j% - 1 to 4 680 c% - c% + 1 690 if b%(j%) - 1 then order%(c%) - nc%(I%) 700 if b%(j%) - 2 then order%(c%) = d%(I%) 710 if b%(j%) - 3 then order%(c%) - a%(I%) 720 if b%(j%) - 4 then order%(c%) - Pc%(i%) 730 next j% 740 next i% 805 print chr$(27)+ chrS(90) 810 print chr$(27)+chr$(41); "University of British Columbia - Department of Psychology" 820 print chrS(27)+chr$(40) 830 print " Encoding Task" 840 print 850 print 860 print chr$(27)+chr$(41);" This study consists of two sessions, each lasting one hour long" 870 print"Session one consists of two tasks that will be described to you shortly." 880 print"Session two will be the same as session one but will be conducted three weeks 890 print" from now." 900 print 910 print 920 print 930 input "Press RETURN to continue with the instructions.",Dummy$ 940 print chr$(27)+chrS(90) 950 print chr$(27)+chr$(41) 960 print chr$(27)+chr$(40) 970 print" Session 1: ENCODING TASK" 980 print 990 print" You will be shown a list of words, one at a time. You are to respond" 1000 print"as quickly as you can stating whether or not the word describes you. You" 1010 print "will then be asked how you rate the word as a description of yourself. 1020 print" For each word, the screen will be blanked for a few seconds, the word 1030 print"READY will then appear on the screen. When it does, place your finger on 1040 print"the spot marked READY on the keyboard. Then the word will be displayed on 1050 print"the screen. Read the word and decide if it describes you or not. Press 1060 print "one of the following keys, once your have decided: 1070 print 1080 print" 1: if you decide 'yes, the word describes me'" 134 1090 1100 1110 1120 1130 1140 1150 1160 1170 1180 1190 1200 1210 1220 1230 1240 1250 1260 same' 1270 1280 1290 1300 1310 1320 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1570 1580 1590 1600 print" print print 2: if you decide 'no, the word does not describe me" dummyS input "press RETURN to continue with the instructions, print chr$(27)+chr${90) print print " Once you have decided if the word describes you, you will be asked to" print"rate the word on a scale of 1 to 5, indicating how stronly the word print"describe you:" print not at all" slightly" somewhat" well" extremely" print" print" print" print" print" print print" 1 2 3 4 5 These choices will be displayed for each word that you are to rate." print"Try doing this with the following sample word. The process will be the print"for the other words", print input" Press RETURN to try a sample woard. ", DummyS print chr$(27)+chr$(40) print chr$(27)+chrS(90) for i% = 1 to second3%: next print chr$(27)+"=-"+chr$(42)+chrS(60); " READY" for i% - 1 to seconds3%: next print char$(27)+chr$(90) for i% = 1 to halfsecond%: next print chr$(27)+"="+chr$(42)+chr<60);" short" print chr$(27)+chr$(41) print chr$(27)+"-"+chr$(47)+chr$(75):"(l) yes" print chrS(27)+"=-"+chr$(48)+chr$(75):"(2) no" print chr$(27)+chr$(40) respS - "" dummyS - inkeyS while respS <> "1" and respS <> "2" respS = inkeyS wend print chr$(27)+"»"+chr$(44)+chr$(52):"How much does this describe you?" print chr$(27)+chrS(41) print chr$(27)+"~"+chr$(47)+chr$(75):"(l) not at all" print chr$(27)+"="+chr$(48)+chr$(75):"(2) slightly" print chr$(27)+"-"+chr$(49)+chr$(75):"(3) somewhat" print chr$(27)+"="+chr$(50)+chr$(75):"(4) well" print chr$(27)+"-"+chr$(51)+chr$(75):"(5) extremely" print chr$(27)+chr$(40) respS = "" dummyS =• inkeyS while respS < "1" or respS > "5" respS = inkeyS wend 135 1610 1620 1630 1640 1641 1650 1660 1670 1680 1690 1700 1710 1720 1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000 2010 2020 2030 2040 2050 2060 2061 2062 2063 2064 2065 print chr$(27)+chr$(40) for i% - 1 to 14: type%(I%) = 3: next i% for i% = 15 to 28: type%(i%) - 2: next i% for i% - 29 to 42: type%(i%) - 1: next i% for i% - 43 to 56: type%(i%) - 4: next i% et prmptSU) - "mean" iet prmpt$(2) -et prmpt$(3) -et prmpt$(4) -et prmptS(5) -et prmpt$(6) -let prmptS(7) -et prmpt$(8) -et prmpt$(9) = et prmptS(lO) •= et prmptS(ll) •» et prmpt$(12) -let prmpt$(13) = et prmpt$(14) -et prmpt$(15) -let prmpt$(16) -et prmpt$(17) = et prmpt$(18) -et prmptS(19) -let prmpt$(20) -et prmpt$(21) •= et prmpt$(22) -et prmpt$(23) = et prmpt$(24) -et prmpt$(25) = et prmpt$(26) = [et prmpt$(27) = et prmpt$(28) = et prmptS(29) » et prmpt$(30) = let prmpt$(31) = et prmpt$(32) -et prmptS(33) = et prmpt$(34) -et prmpt$(35) -et prmpt$(36) = et prmpt$(37) = et prmpt$(38) = et prmpt$(39) = et prmpt$(40) = et prmpt$(41) = :et prmpt$(42) = et prmptS(43) = et prmpt$(44) = et prmptS(45) = et prmpt$(46) = et prmptS(47) = messy " shallow" "naive" "immoral" "vain" •homely" "stubborn" "dishonest" "immature" "conceited" "repulsive" "greedy" "sneaky" "depressed" "defeated" "despondent" "dismal" "desolate" "sluggish" "lifeless" "dejected" "dreary" "sorrowful" . "mournful" "downcast" "blue" "sad" "anxious" "jumpy" "jittery" "worried" "restless" "enraged" "afraid" "apprehensive" "fearful" "panicky" "tense" "excited" "hyperactive" "nervous" "gifted" "thriving" "attentive" "assertive" "loved" 136 2066 let prmpt$(48) = "spirited" 2067 let prmpt$(49) = "industrious" 2068 let prmpt$(50) - "happy" 2069 let prmpt$(51) = "confident" 2070 let prmpt${52) •» "rested" 2071 let prmpt$(53) = "productive" 2072 let prmpt$(54) = "cheerful" 2073 let prmpt$(55) = "secure" 2074 let prmpt$(56) - "dynamic" 2080 let ignorS(l) = "dull" 2090 let ignorS(2) - "talkative" 2100 let ignor$<3) = "outgoing" 2110 let ignorS(4) - "grim" 2140 print chr$(27)+chr$(90) 2150 print "Good. Now you are ready to start. You will see 60 words, one at a time." 2160 print"You are to respond just as you have to the sample word. Remember you are" 2170 print"to respond as quickly and as accurately as you can. 2180 print 2190 print" After the list, the experimenter will come in, and will ask you to recall" 2200 print"as many of the 60 words as you can." 2210 print 2220 print" If you have any questions, please ask the experimenter now. Otherwise," 2230 print"press RETURN to begin the test." 2240 print 2250 print 2260 print 2270 print 2280 print 2290 print 2300 print 2310 print 2320 print 2330 input "press RETURN to start the test.",dummy$ 2340 rem -- first the two words that don't count 2350 for cur% - 1 to 2 2360 print chr$(27)+chr$(40) 2370 print chr$(27)+chr$(90) 2380 for i% - 1 to seconds3%: next 2390 print chrS(27)+"="+chr$(42)+chrS(60): " READY" 2400 for i%= 1 to seconds3%: next 2410 print chr$(27)+chr$(90) 2420 for i%=l to halfsecond%: next 2430 print chr$(27)+"="+chr$(42)+chrS(60):ignor$(cur%) 2440 print chrS(27)+chr$(41) 2450 print chr$(27)+"="+char$(47)+chr$(75):"(l) yes" 2460 print chr$(27)+"="+char$(48)+chr$(75):"(2) no" 2470 print chr$(27)+chr$(40) 2480 respS - "" 2490 dummyS - inkeyS 2500 count% - 0 137 2510 while respS <> "1" and respS <> "2" 2520 if count% < 32700 then count% = count% + 1 2530 respS = inkeyS 2540 wend 2550 ir(cur%) - count% / insec 2560 print chr$(27)+"="+chr$(44)+chr$(52):"how much does this describe you?" 2570 print chr$(27)+chr(41) 2580 print chrS(27)+"="+chrS(47)+chr$(75):"(l) not at all" 2590 print chr$(27)+"="+chr$(48)+chr$(75):"(2) slightly" 2600 print chr$(27)+"="+chr$(49)+chr$(75):"(3) somewhat" 2610 print chrS(27)+"="+chrS(50)+chrS(75):"(4) well" 2620 print chr$(27)+"="+chrS(51)+chr$(75):"(5) extremely" 2630 print chr$(27)+chr$(40) 2640 respS ="" 2660 dummyS = inkeyS 2670 while respS < "1" or respS > "5" 2680 respS - inkeyS 2690 wend 2700 next cur% 2710 rem 2720 rem -- and now for the 56 that do matter 2730 rem 2740 etl - 0 2750 etavt = 0 2760 for cur% = 1 to words% 2770 print chrS(27)+chr$(40) 2780 print chrS(27)+chr$(90) 2790 for i% = 1 to seconds3%: next 2800 print chr$(27)+"="+chr$(42)+chrS(60): "ready" 2810 for i% •» 1 to seconds3%: next 2820 print chr$(27)+chr$(90) 2830 for i% " 1 to halfsecond%: next 2840 print chr$(27)+"="+chr$(42)+chr$(60):prmpt(order%(cur%)) 2850 print chrS(27)+chr$(41) 2860 print chr$(27)+"="+chrS(47)+chrS(75):"(l) yes" 2870 print chr$(27)+"="+chr$(48)+chr$(75):"(2) no" 2880 print chr$(27)+chr$(40) 2890 respS ="" 2900 dummyS = inkeyS 2910 count% - 0 2920 while respS <> "1" and respS <> "2" 2930 if count% < 32700 then count% - count% + 1 2940 respS - inkeyS 2950 wend 2960 reactime(order%(cur%)) = count% / insec 2970 etavt = etavt + reactime(order%(cur%)) 2980 etract(type%(order%(cur%)))-etract(order%(cur%)))+reactime(order%(cur%)) 2990 etans$(order%(cur%))=resp$ 3000 if resp$="l" then etl-etl+1: etones(type$(order%(cur%))) = etones (type%(order%(cur%))) + 1 3010 print chr$(27)+"="+chr$(44)+chr$(52):"How much does this describe you?" 3020 print chr$(27)+chr$(41) 133 3030 print chr$(27)+"="+chr$(47)+chr$<75):"U) not at all" 3040 print chr$(27)+"='*+chr$(48)+chr$(75):"(2) slightly" 3050 print chr$(27)+"="+chr$(49)+chr$(75):"(3) somewhat" 3060 print chr$(27)+"="+chr$(50)+chr$(75):"(4) well" 3070 print chr$(27)+"-"+chr$(51)+chrS(75):"(5) extremely" 3080 print chrS(27)+chr$(40) 3090 resP$="" 3110 dummyS = inkeyS 3120 count% - 0 3130 while respS < "1" or respS > "5" 3140 if count% <32700 then count% = count% + 1 3150 respS - inkeyS 3160 wend 3170 r2(order%(cur%)) = count% / insec 3180 etrateS(order%(cur%)) = respS 3190 etrating%(type%(order%(cur%)))-etrating%(type%(order%(cur%)))+cint(val(resp$)) 3200 next cur% 3210 etl - etl * 100 /words% 3220 et2 - 100 - etl 3230 etavt - etavt /words% 3240 for cur% = 1 to 4 3250 etreact(cur%) - etreact(cur%) / 14! 3260 next 3270 rem 3280 rem — and finally for the last two than don't count 3290 rem 3300 for cur% - 3 to 4 3310 print chr$(27)+ chr$(40) 3320 print chrS(27)+chr$(90) 3330 for i% = 1 to seconds3%: next 3340 print chr$(27)+"-"+chr$(42)+chr$(60): " ready" 3350 for i% •» 1 to seconds3%: next 3360 print chr$(27)+chr$(90) 3370 for i%— 1 to halfsecond%: next 3380 print chr$(27)+"="+chrS(42)+chr$(60): ignor$(cur%) 3390 print chr$(27)+chrS(41) 3400 print chr$(27)+"-"+chr$(47)+chr$(75):"(l) yes" 3410 print chr$(27)+"="+chr$(48)+chrS(75):"(2) no" 3420 print chr$(27)+chr$(40) 3430 respS 3440 dummyS=inkey$ 3450 count%=0 3460 while respS <> "1" and respS <> "2" 3470 if count% < 32700 then count% = count% + 1 3480 respS = inkeyS 3490 wend 3500 if(cur%) - count% / insec 3510 print chr$(27)+"="+chr$(44)+chr$(52):"How much does this describe you?" 3520 print chr$(27)+chrS(41) 3530 print chr$(27)+"="+chr$(47)+chr$(75):"(l) not at all" 3540 print chr$(27)+"="+chr$(48)+chr$(75):"(2) slightly" 3550 print chr$(27)+"="+chrS(49)+chrS(75):"(3) somewhat" 3560 print chr$(27)+"=-"+chr$(50)+chr$(75):"(4) well" 3570 print chrS(27)+"="+chr$(51)+chr$(75):"(5) extremely" 3580 print chr$(27)+chr$(40) 3590 respS -"" 3610 dummyS - inkeyS 3620 while respS < "1" or respS > "5" 3630 respS - inkeyS 3640 wend 3650 wend 3650 next cur% 3660 print chr$(27)+chr$(90) 3670 print chr$(27)+"="+chr$(42)+chr$(50):"one moment please" 3680 print 3690 print 3700 open "0",l,file$ 3710 for i% - 1 to 4 3720 write#l,ir(I%) 3730 next i% 3740 for i% = 1 to words% 3750 writetfl, order%(i%),reactime(i%), r2(i%),etrate$(i%),etans$(i%) 3760 next i% 3770 for i% - 1 to 4 3780 write«l,etract(I%),etreact(I%),etrating%a%)Ietones(I%) 3790 next i% 3800 writettl.etl.etavt 3810 close n 3820 print chr$(27)+chr$(90) 3830 print chrS(27)+chr$(41):" You have now finished this test." 3840 print 3850 print" Please get the experimenter." 3860 end Appendix G 95% confidence intervals for BDI, DAS, STAI-state, and STAI-trait scores for Study 2 141 95% confidence intervals for BDI, DAS, STAI-state.and STAI-trait scores for Study 2. Questionnaire score Groupwise comparison 95% confidence interval Beck Depression Inventory MO vs MI 5.36 10.75 » MO vs ND 14.24 20.12 * MI vs ND 6.57 11.68 * Dysfunctional Attitude Scale MO vs MI 2.50 36.87 * MO VS ND 46.88 84.39 « MI VS ND 29.64 62.26 * STAI-state MO vs MI -4.89 13.52 MO vs ND 4.16 24.26 * MI vs ND 1.16 18.63 s STAI-trait MO vs MI -0.82 14.09 MO vs ND 13.12 29.39 • MI vs ND 7.54 21.70 * £ < .05 MO = moderately depressed group MI - mildly depressed group ND = nondepressed group Appendix H ANOVA summary table for the endorsement rate index for Study 2 143 ANOVA summary table for the endorsement rate index, as a function of subject group and content condition for Study 2. Source of variation df MS Between subjects Groups 2 .45 8.31*** Subjects w groups 85 .06 Within subjects Content condition 3 2.44 64.87*** Group x content 6 1.18 31.22*** Content x subjects w groups 255 .04 *** £ < 001 Appendix I ANOVA summary table for the descriptiveness ratings index for Study 2 145 ANOVA summary table for the descriptiveness ratings index, as a function of subject group and content condition for Study 2. Source of variation df MS Between subjects Groups 2 3.26 10.63*" Subjects w. groups 85 .31 Within subjects Content condition 3 18.84 87.39*** Group x content 6 6.00 27.85*** Content x subjects w. groups 255 .22 *** E < 001 Appendix J ANOVA summary table for the reaction time index for Study 2 147 ANOVA summary table for the reaction time index, as a function of subject group and content condition for Study 2. Source of variation df MS Between subjects Groups 2 2.93 2.91 Subjects w. groups 85 1.01 Within subjects Content condition 3 .54 5.78 Group x content 6 .30 3.23** Content x subjects w. groups 255 .09 " E < 01 *** £ < -001 Appendix K ANOVA summary table for the recall index for Study 2 149 ANOVA summary table for the recaii index, as a function of subject group and content condition for Study 2. Source of variation df MS Between subjects Groups 2 .01 .34 Subjects w groups 85 .04 Within subjects Content condition 3 .02 .87 Group x condition 6 .03 1.52 Cond.subjects w. groups 255 .02 Appendix L ANOVA summary table for BDI scores for Study 3 151 ANOVA summary table for BDI scores, as a function of subject group and time for Study 3. Source of variation df MS F Between subjects Groups 2 2670.99 84.63*** Subjects w. groups 85 31.56 Within subjects Time 1 183.29 21.32*** Group x time 2 220.28 25.63*** Time x subjects w. groups 85 8.60 ** £ < 001 Appendix M ANOVA summary table for DAS scores for Study 3 153 ANOVA summary table for DAS scores, as a function of subject group and time for Study 3. Source of variation df MS Between subjects Groups Subjects w. groups 2 85 48052.15 642.65 74.77" Within subjects Time 1 Group x time 2 Time x subjects w. groups 85 1.37 830.84 144.05 .01 5.77" " £ < .01 • * * 2 < 001 Appendix N ANOVA summary table for STAI-state scores for Study 3 155 ANOVA summary table for STAI-state scores, as a function of subject group and time for Study 3. Source of variation df MS Between subjects Groups Subjects w. groups 2 85 1851.21 121.51 15.24*** Within Subjects Time 1 Group x time 2 Time x subject w. groups 85 135.64 174.16 78.85 1.72 2.21 *** £ < .001 Appendix 0 ANOVA summary table for STAI-trait scores for Study 3 157 ANOVA summary table for STAI-trait scores, as a function of subject group and time for Study 3. Source of variation df MS Between subjects Groups 2 5011.70 29.99*** Subjects w. groups 85 167.12 Within subjects Time 1 1114.39 9.56** Group x time 2 116.84 1.00 Time x subjects w groups 85 116.51 *« 2 < -01 *** £ < 001 Appendix P ANOVA summary table for the endorsement rate index for Study 3 - Time 1 159 ANOVA summary table for the endorsement rate index for Study 3 - Time 1 data. Source of variation df MS Between subjects Groups 2 .37 6.50' Subjects w. groups 85 .06 Within subjects Content condition 3 3.15 68.75 Group x content 6 .83 18.11*** Content x subject w. groups 255 .05 ** £ < .01 *** £ < .001 Appendix Q ANOVA summary table for the descriptiveness ratings index for Study 3 - Time 1 161 ANOVA summary table for descriptiveness ratings for Study 3 - Time 1 data. Source of variation df MS Between subjects Groups 2 3.14 10.14*** Subjects w. groups 85 .31 Within subjects Content condition 3 22.58 88.47*** Group x content 6 4.32 16.93*** Content x subjects w. groups 255 .26 * * * 2 < 001 Appendix R ANOVA summary table for the reaction times index for Study 3 - Time 1 ANOVA summary table for reaction time index for Study 3 - Time 1 data. Source of variation df MS F Between subjects Groups 2 2.72 2.68 Subjects w. groups 85 1.01 Within subjects Content condition 3 .54 5.70** Group x content 6 .22 2.31* Content x subjects w. groups 255 .09 * £ < .05 ** £ < .01 Appendix S ANOVA summary table for the recall index for Study 3 - Time 1 165 ANOVA summary table for the recall index for Study 3 - Time 1 data Source of variation df MS F Between subjects Groups 2 .02 .40 Subjects w. groups 85 .04 Within subjects Content condition 3 .02 .93 Group x content 6 .01 .37 Content x subject w. groups 255 .02 Appendix T MANOVA summary table for Study 3 MANOVA summary table for Study 3. 167 Effects Wilks' lambda F-ratio Between subjects Groups .597 F(12,160) - 3.91 *** Within subjects Content condition .344 F(18,708) - 18.01 Group x content .563 F(36,H00) - 4.28 * " Time .662 F(6,80) - 6.81 " * Group x time .869 F(12,160) - .97 Content x time .865 F(18,708) = 2.07 " Group x content x time .816 F(36,1100) - 1.45 * • £ < .05 •* £ < .01 *** £ < .001 Appendix U ANOVA summary table for the endorsement rate index for Study 3 ANOVA summary table for the endorsement rate index for Study 3. Source of variation df MS Between subjects Groups 2 .79 8.18" Subjects w. groups 85 .10 Within subjects Content condition 3 7.33 92.41*** Group x content 6 1.34 16.94*** content x subjects w. groups 255 .08 Time 1 1.62 1.62 Group x time 2 .03 1.54 Time x subjects w. groups 85 .02 Content x time 3 .12 6.31*** Group x content x time 6 .07 3.52** Content x time x subj w groups 255 .02 * £ < .05 " 2 < 01 * * * 2 < 001 Appendix V ANOVA summary table for the descriptiveness ratings index for Study 3 ANOVA summary table for the descriptiveness index for Study 3. Source of variation df MS Between subjects Groups 2 6.57 11.24*** Subjects w. groups 85 .59 Within subjects Content condition 3 47.40 118.89*** Group x content 6 8.65 21.68*** content x subjects w. groups 255 .40 Time 1 1.00 6.48* Group x time 2 .49 3.17* Time x subjects w. groups 85 .16 Content x time 3 .50 4.56** Group x content x time 6 .33 3.05** Content x time x subj.w group 255 .11 * £ < .05 ** £ < .01 *** £ < .001 Appendix W ANOVA summary table for the reaction time index for Study 3 ANOVA summary table for the reaction time index for Study 3. Source of variation df MS Between subjects Groups 2 5.64 2.97 Subjects w. groups 85 1.90 Within subjects Content condition 3 1.24 8.84'** Group x content 6 .30 2.12" content, x subjects w. groups 255 .14 Time 1 6.00 17.64"* Group x time 2 .31 .97 Time x subjects w. groups 85 .34 Content x time 3 3.06 .71 Group x content x time 6 .02 .21 Content x time x subj w groups 255 .09 *** £ < .001 Appendix X ANOVA summary table for the recall index for Study 3 ANOVA summary table for recall index for Study 3. Source of variation df MS F Between subjects Group 2 .03 .41 Subjects w. groups 85 .06 Within subjects Content 3 .01 .49 Group x content 6 .01 .54 Content x subjects w. groups 255 .02 Time 1 .28 20.81**" Group x time 2 .00 .06 Time x subjects w. groups 85 .01 Content x time 3 .01 .85 Group x content x time 6 .01 .54 Content x time x subj w groups 255 .01 *** £ < .001 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0097257/manifest

Comment

Related Items