JOURNAL OF PERSONALITY ASSESSMENT, 1990, 54(3é&), 546-563, Copyright © 1990, Lawrence Erlbaum Associates, Ine. Measuring Generalized Expectancies for Negative Mood Regulation: Initial Scale Development and Implications Salvatore J. Catanzaro Illinois State University Jack Mearns University of Connecticut ‘A scale measuring generalized expectancies for negative mood regulation (NMR) vas developed. The construct was defined as the expectancy that some behavior ot Cognition will alleviate a negative mood state. Data from five samples of college tndergraduates (N = 1,630) were reported. Internal consistency, discriminant Validity from social desirability, and temporal stability were demonstrated for @ iter scale derived from an initial pool of 50 items. Further analyses revealed (a) modest correlations of the 30-item scale with internal-external control, (b) that high scorers on the NMR scale reported few sympcoms of depression, and (c) that the NMR scale predicted a different pattern of emotions than the Beck Depression, Inventory did. Discussion focused on possible roles for expectancies for negative mood regulation in the coping process and directions for future research on the ‘mood regulation process. Research has expanded the study of behavioral self-regulation (Mischel, 1973) to affective self-regulation. Rippere (1976, 1977, 1979) found that many people try to cope with depression, behaviorally and cognitively, in commonsensical ways. For example, subjects reported attempting to see a friend or analyzing the reason for their depression. ‘Doerfler and Richards (1981) studied selfinitiated attempts to cope with depression. They interviewed adult women who reported either having beer seoressful at coping with a previous depression or not. Two differences emerged between the successful and nonsuccessful copers. First, che successful copers pnore often reported making a change in their social environment (¢.g., moving). Second, successful copers believed the coping statements they made to themNEGATIVE MOOD REGULATION 547 selves (e.g., ‘Tl feel better soon”). Although these data were based on retrospec- tive self-reports from a small sample, they are suggestive of the utility of studying individual differences in regulating negative moods. In a similar vein, Burns, Shaw, and Croker (1987) reported that depressed women had lower expectan. cies that their attempts to cope with sadness would be helpful than nondepressed women did. These findings suggest that a generalized expectancy for problem solving otter, Chance, & Phares, 1972) is an important dimension of affective self-regulation. This expectancy is a subjectively held belief that behaviors in the individual's repertoire will successfully solve a problem as it is encountered in a variety of situations. Franko, Powers, Zuroff, and Moskowitz (1985) defined a generalized expectancy for affective self-regulation as “an expectancy that some overt behavior or cogni- tion will alleviate a negative state or induce a positive one" (p. 210). Through interviews, they found that children who were asked to describe situations in which they felt sad and angry also generated numerous solutions for these problems. This ability was taken as evidence of these children’s generalized expectancies for affective self-regulation. Following the definition of a generalized expectancy for affective self regulation of Franko et al. (1985), the questionnaire described in this article was designed to measure generalized expectancies for the self-regulation of negative moods. SOCIAL LEARNING THEORY AND GENERALIZED EXPECTANCIES FOR PROBLEM SOLVING In social learning theory (Rotter, 1954, 1982; Rotter et al., 1972), generalized expectancies for problem solving combine with situational expectancies and reinforcement values in the prediction of behavior. The influence of generalized expectancy varies as a function of the novelty of the situation: The more novel a situation, the more a generalized expectancy contributes to behavior (Rotter, 1975). Generalized expectancies for problem solving can also determine be. havior when the generalized belief defines or categorizes situations in a manner that implies or precludes certain behavioral choices. For example, people with an external locus of control (Rotter, 1966) may behave helplessly because they believe the world is unresponsive to their efforts to change it Zuroff, 1980). Another category of expectancy has been elaborated by Kirsch (1985). He defined response expectancy. as the expectancy of the occurrence of a nonvolitional (e.g., autonomic, emotional, and hypnotic) response and summa. "zed evidence indicating that these expectancies are self-confirming. Therefore, the expectancy that the successful execution of a given behavior will change « mood state is itself likely to influence that mood state. Thus, a generalized548 | CATANZARO AND MEARNS expectancy for alleviating negative moods may be mote precisely defined as a response expectancy that generalizes across situations in which negative mood states may be experienced. Assessing generalized expectancies for affective self-regulation should, therefore, predict successful attempts to cope with nega tive mood states. “These attempts at negative mood regulation are hypothesized to have short- term emotional consequences and long-term consequences for adjustment and health, Ifa person has a low expectancy for alleviating negative moods, attempts to alter such a mood will be feeble at best, and the negative mood state will continue or became worse. Failing to regulate these-negative moods will serve to confirm expectations about the inability to do so. In contrast, positive expect- ancies should generate coping responses. These coping behaviors thay or may ‘not have intrinsic mood-altering properties; even if they do-not, the belief that they will may lead to mood enhancement (Kirsch, 1985), thereby initiating « benign cycle. “The results of the interview studies of Rippere (1977) and Franko et al. (1985) were tused t0 develop the Generalized Expectancy for Negative Mood Regulation (NMR) Scale, the self-report questionnaire described in this article, This ques- tionnaire wos expected to afford an economical alternative ro interviews for data collection, thereby facilitating research on the construct. A S:point Likere'scale format was used. Initially, 50 items were wrieten. Items consisted of expectancy statements about coping behaviors, all with the stem, “When I'm upset, [believe that. ...” Five types of items were included. General items referred to the possibility that negative moods could or could not be alleviated, rather than to Specific sttatgies of alleviation. Cognitive iems sanaples thoughts and cognitive strategies that might influence negative moods. Similarly, behavioral—alone and behavioral--social items sampled actions that would influence negative moods that one might perform alone or with other people, respectively. Finally, behavioral unspecified items sampled actions with which one might affectione’s mood either along or with others. The instructions are found in the Appendix. In this article, we addressed thece issues. First, we sought to. establish the ingernal consistency of a set of items) selected from jan. initial peel, thelr discriminant validity from social desirabilicy (Crowne Su Marlowe, 1964), and the temporal seability of the retained items. Second, we soughe to demonstrate the discriminant validity of the new scale from locus of control Rarter, 1966). Finally, we began! co explore the-relations amongithe NMR scale, messes of deprestion, and othet variables relevant to: mood ind aifjustment: SAMPLES AND PROCEDURES Five nonoverlapping samples of college students provided the data’ reported here. Descriptions of each, the testing sprocedures they ‘experienced, and the measures they completed in addition to the NMR follow.NEGATIVE MOOD REGULATION 549 Sample 1 Subjects were 508 female and 281 male (N = 789) introductory psychology students at the University of Connecticut participating in a’ general testing procedure. Early in the semester, all introductory psychology students partici- Pated in a mass test, completing several questionnaires in a class hour. This sample completed a 50-item version of the NMR. Overlapping but different subsamples completed the Social Desirability Scale (SDS; Crowne & Marlowe, 1964), and Rotter’s (1966) Internal-External Locus of Control (I-E) Scale also. Questionnaires were interspersed among other inventories in a randomly determined order. The SDS is a 33-item crue-false inventory that measures the tendency of subjects to portray themselves in an acceptable light by admitting to behaviors that are socially desirable but statistically improbable or denying behaviors that are common but socially unacceptable. The I-E scale is a 23-item forced-choice inventory (plus 6 filler items) that requires subjects to endorse statements reflecting either the belief that reinforcements they receive are the result of their own behavior (internal control) or are caused by fate, chance, or powerful others (external control). Sample 2 A sample of 80 female and 33 male students at the University of Connecticut completed the NMR twice over a 3- to 4-week interval. The first assessment was made during group testing sessions similar to those in Sample 1. The second assessment was also a group session for which subjects were recruited via telephone. The minimum time span between the two sessions was 24 days; the maximum was 30 days. Most subjects experienced at least a 4-week interval between sessions. Subjects received course credit for participating in the second session. Sample 3 Subjects were 140 female and 65 male students at Ilinois State University who had volunteered to participate in a two-session study for extra course credit. The two sessions were no fewer than 6 weeks apart and no more than 8 weeks apart. Sample 4 ‘The NMR was administered to a sample of 272 female and 174 male students at the University of Connecticut in a procedure similar to that used with Samples 1 and 2. Of these students, 37 men and 125 women were also administered the SDS and the short form of the Beck Depression Inventory (BDI-SF; Beck, Rial,550 CATANZARO AND MEARNS & Rickels, 1974). The Beck Depression Inventory (BDI; Beck, Rush, Shaw, & Emery, 1979) is a 2Litem questionnaire that requires subjects to select state- ments descriptive of them during the past week from sets of four, which indicate increasingly severe depressive symptopms (c.g, sleep disturbance and loss of pleasure). The BDI-SE consists of 13 items selected from the more familiar BD! And wasused with this sample because of time constraints imposed by the testing procedure. Beck et al. (1974) reported correlations of the BDI-SF with the regular BDI ranging from .89 to .97. Sample 5 Subjects were 77 members of social fraternities at the University of Connecticut.’ They were recruited via their fraternities, and payment for participation was made to the organizations. These subjects completed the NMR, the regular version of the BDI, the Rotter Incomplete Sentences Blank (SB; Rotter & Rafferty, 1950), and Mosher and White's (1982) revision of zard’s (1972) Differential Emotions Scale (DES) in two different sessions. “The ISB is a semiprojective instrument which requires that subjects write completions of 40 sentence stems (e-g., "A mother. ..) $0 as to reveal their true feelings. These are scored on a 7-point scale for level of conflict, with high scores indicating high levels of conflict. A randomly selected set of 20 protocols was scored independently;? correlating these scores with those of the first author evealed good interrater reliability, x18) = 82, p < .OOL, and at test for Correlated samples revealed no systematic bias, (18) = — 1.27, ns. Thescores of the firse auchor were used in all analyses. The validity of the ISB as a broad treasure of maladjustment in college students has been well established (Lah, 1989; Watson, 1978). ‘The DES measures the discrete emotions of interest, enjoyment, surprise, dlistress, anger, fear, guilt, contempt, and shame. Each emotion was represented by 6 adjectives; subjects indicated how intensely, on a 1 to 5 scale, they feteach emotion “tight now,” with higher scores indicative of greater emotional inten sity, The validity of the DES as a measure of patterns of emotional states has generally been supported (¢, lard, 1972; Mosher & White, 1962). "The first session was a group testing” at which subjects completed the ISB, the ‘Feary mambers were recruited to afford use of peer noination procedure which included = number of etjutment cated tat measures (€, happy v8. unhappy) and control vatiablés (© aa offends). The pattern of relations among the peer nominations indicted that subjects did earner celina between the erste and that nominations may have been influenced by halo effects SFoopulary, Therefore, no resule from thismeasure are presented here. An aril adress these and ocher measurement issues is in preparation. “We are grateful to Ronald Anderson for scoring the 20 ISB protocols Tee Soca geoupe the Bret session took pace a ther eatery ous; the fourth group was tested ‘ata mecting room on campus. No group differences were found on any ofthe measuresNEGATIVE MOOD REGULATION 551. NMR, and the BDI. The first author, one male assistant, and one female assistant monitored the session and ensured that subjects followed instructions. The second session consisted of individual testing. This session usually took place within 2 weeks of the group session. The majority of individual testing sessions took place within 1 week of the group testing. Here, subjects completed the BDI and the DES in the presence of the first author, who was blind to their ‘scores on measures from the first session. SCALE DEVELOPMENT AND PSYCHOMETRIC PROPERTIES Item Selection These analyses proceeded with the initial goal of establishing the general psy- chometric properties of the NMR separately for men and women. Items were retained if they met the four following criteria: (a) a mean score that was within the range of 1.5 and 4.5 on the Likert scale; (b) a standard deviation near 1.0, indicating adequate dispersion of scores without too great a range; (c) a corrected item-total correlation between .30 and .60, indicating a balance between con- sistency with the rest of the scale and unique contribution to the total scale score; and (d) a correlation with social desirability lower than the corrected item-total correlation. Further, items were chosen to minimize sex differences in the items statistics, thereby creating a single scale with utility for both men and women. Application of the item retention criteria led to the inclusion of 30 items in a single scale that appeared useful for both men and women. The average corrected item-total correlations were .40 for men and .42 for women (both p < 001), Of the retained items, half were scored in the positive direction; for the remainder, the scoring was reversed. The 30 retained NMR items can be found in Table 1. Descriptive statistics and alpha internal consistency coefficients for the 30-item scale obtained from all samples can be found in Table 2. Note that mean Scores are generally consistent in this population, that there are no statistically significant sex differences in mean scores, and that the alpha coefficients are uniformly high. As might be expected, given that the retained items were selected partially because of their low correlation with social desirability, the NMR correlated av a low level with the SDS. This correlation was significant for the women in Samples 1 who completed both measures, 1(414) = .17, p < .05, but not the men, 1(248) = .09, p > .05. However, the correlations for men and women are not significantly different from each other, p > .05, Intercorrelations of subscales based on item content are presented in Table 3, using Sample I data. Reflecting the high internal consistency of the entire scale, the subscales correlated from a moderate to strong degree. These correlations552 CATANZARO AND MEARNS: TABLE 1 NMR Scale Items 1. Tecan usually find a way to cheer myself up. 2. Lean do something to fel better. ‘Wallowing init is all Tcan do. (a) Tl feel okay if chink about more pleasant times. ‘Being with other people will be a drag. (n) Tan feel better by treating myself to something Tike. Til feel betver when I understand why I fel bad. won't be able to get myzelf to do anything about i. (n) {won't feel much better by trying to find some good in the situation. (x). 10. Ie wont be long before Ican calm myself down. (o) 11. Ie will be hard to find someone who really understands. (n) 12. Telling myself twill pas will help me calm down. 13. Doing something nice for someone else will cheet me up. 14, Tend up feeling really depressed. (n) 15. Planning how Fl eal wih chings will help. 16. Lean forget about whats upsetting me pretty easily. 17, Catching up with my work will help me calm down. 18. ‘The advice friends give tae won't help me feel better. (a) 19. wont be able to enjoy the things Tusually enjoy. (n) 20, Vest find a way to relax. 21, Trying to work the problem out in my head will only make it seem worse. (n) 22. Seeing a movie won't help me fel bere. (n) 2B. Going out to dinner with friends will help. 24. Fl be upset for a long tine. (n) 25, wont be able to put i 6ut of my mind. (a) 26. Lcan feel better by doing something creative. 27. stare to feel really down about myself (n) 38. Thinking that things will eventually be better won't help my feel any better (n) 39. ean find some humor in the situation and fecl beter. 50. Tei with @ group of people; Tl feel “alone in a crowel.”(n) Nae All ners have the same stem: “When Tm upset, Tbelieve that. ‘pln Scale ranging fom song dsareement (1) tong agreement (5). Negative items are denowed ‘by (a); scoring is reversed prior to computation of statistics. vena ? Irema were rated ot suggest that although different subscales may prove useful in the prediction of some criteria (e.g., using cognitive vs. behavioral mood regulation strategies), the 30 items measure a broad, unitary construct." ‘Temporal Stability One final question of initial scale development regards the temporal stability (Geat-retest reliability) of NMR scores. Because itis conceived as a measure of a TA principal components factor analysis of Sample 1 dat yieked a soltion conden with cis inxerettion, Beonse these resus are redundant with those ofthe internal consistency analyses, they are not reported here. More information is available from the authors.NEGATIVE MOOD REGULATION 553 TABLE 2 Descriptive Statistics for the NMR Scale in Five Samples Sample * ™ ‘SD Range « 1 Men 281 99.68 143 aT 8 Women 508 99.14 14.3 58-141 87 2 Time 1 Men 3 98.55 11.82 2-133 90 Women 2 102.26 15.95 61-132 87 Time 2 Men 3 101.2 1884 os 3 Women 8 105.00 ts21 51-150 92 3 Time 1 Men 6 104.43 15.08 70-132 88 Women 40 104.96 M410 O18 90 Time 2 Men 65 106.83 1527 2135 87 Wornen 40 105.60 1434 8-136 88 4 Mea 14 103.66 1720 S48 30 Women m 102.99 16.69 51-13 89 5 Men 7 107.40 1498 ort 86 Note. High score = high expectancy TABLE 3 Intercorrelations of NMR Subscales in Sample 1 General Cognitive Behavioral Ment General 7 Cognitive 8 - Behavioral 52 53 = Women General - Cognitive 1 - Behavioral 4 3 - Note, Allcorrelations are significant, p < .001. General items are 1,2,3,6, 10,14, 19;20,04,and 37, Cosnitive items are 4,7, 9, 12, 15, 16,21, 25,28, nd 29. Behavioral items are 5 6,11, 13,17, 18, 22, 23, 26, and 30. ‘n= 281." = 508, set of generalized expectancies, we expect that NMR scores will prove to be stable over time, in the absence of experiences (like psychotherapy) that may provide a basis for changing these belief. The stability obtained from Sample 2 for NMR scores over a 3- to 4-week period for women was 1(78) = .74; for ‘men, 731) = .76. For men in Sample 3, who were retested at a 6- to Bweek interval, 1163) = .67; for women, (138) = .78, all ps < .001. Note thar NMR scores tended to increase at Time 2 in both samples (see Table554 CATANZARO AND MEARNS 2), Repeated measures analyses of variance (ANOVAS) revealed that for Sample 2, this effect was significant, F(1, 111) = 4.18, p < .05; for Sample 3, the effect approached significance, F(1, 203) = 3.74, p < .06. No interactions with sex vere found, Because the mean increases are small relative to the range of scores, it is unlikely that the statistical significance of these results indicates much practical significance. Ic is an open empirical question if these differences reflect feactivity to repeated testing or an actual increase in confidence among these ‘young adults, predominantly freshman and sophomore college students, in their ‘mood regulation abilities. RELATIONSHIP WITH INTERNAL-EXTERNAL CONTROL We expected that NMR items would exhibit discriminant validity from locus of control (Rotter, 1966) because locus of control pertains to external reinforcing events, whereas affective self-regulation pertains to one’s own emotional states. Still, some overlap between the constructs does exist; thus, significant, but moderate negative correlations between measures of the constructs were pre- dicted, because higher I-E scores are indicative of endorsing more external items. ‘Data were available from 358 subjects (135 men and 223 women) from Sample 1. NMR scores were significantly correlated with I-E scores in the predicted direction for women, 1221) = —.35, p < 001, but not for men, 1133) = —.09, p > .05. These correlations are significantly different, t = 2.49, p < .05. “Although the original hypothesis did not take sex differences into account, this finding may reflect a previously observed tendency for the I-E scale to Comelate with some measures of adjustment for women but not men (Phares, 1976; Warehime & Foulds, 1971). As such, the results are consistent with our conception of expectancies for negative mood regulation as one aspect of the ‘adjustment process. Further, the low magnitudes of these correlations are evidence of the discriminant validity of the NMR from a measure of locus of control. RELATIONSHIPS WITH MEASURES OF DEPRESSION, ADJUSTMENT, AND AFFECT Because depression is associated with the duration as well as the occurrence of negative moods, generalized expectancies for negative mood regulation should be an important predictor of dysphoria and depression. Our purpose here was twofold: to establish the relationship of the NMR to self-reported depressionNEGATIVE MOOD REGULATION 555 and to demonstrate the discriminant validity of the NMR from measures of depression or dysphoria. The correlation matrices of the NMR scale with the BDI-SF and the SDS for subjects in Sample 4 who completed all three measures are presented in Table 4. Correlations of the NMR with the SDS followed the same pattern as was found for Sample 1. The correlations with the BDI-SF showed that subjects who reported a strong belief that they can alter negative moods are less depressed than those who do not hold this belief as strongly. This relation was particularly robust for men. However, for women, there appeared to be some shared response style variance due to social desirability among the scales, which might account for the high correlations. ‘We sought to rule out the hypothesis that, for women, the apparent relation between the NMR and the BDI-SF was, in fact, due to a shared tendency for both scales to elicit socially desirable responses. Therefore, we tested the contribution of the NMR to variance in the BDI-SF with social desirability statistically controlled in a hierarchical multiple regression. Results indicated that the NMR and BDI-SF shared variance independent of that shared with social desirability, F(1, 122) = 16.56, p < .001. The increment in R? was .11. The correlation of the NMR with a measure of depression is a first step in establishing its validity as a measure of an important determinant of negative mood regulation, but it leaves open the question of discriminant validity. Demonstrating that the NMR has a different pattern of relationships than does the BDI with broader adjustment measures and vatious emotional states would be evidence that the NMR is not merely a proxy measure of depression, Thus, analyses of data from Sample 5 were designed to replicate and extend those from Sample 4 in the following ways: First, we correlated the NMR with ewo separate ‘administrations of the regular 21-item version of the BDI; second, we correlated TABLE 4 Correlations of NMR Scores With Social Desirability and Depression Sample 4 NMR. BDL-SF SDS Ment NMR - BDL-SF 566 S SDs 20 al 5 Women" ‘NMR - BDI-SE 398 5 SDs 23 ~296 = Note. NMR = Generalized Expectancy for Negative Mood Regulation Scale; BD-SF = Beck Depression Inventory, Short Form; SDS = Marlowe Crowne Social Desirability Scale. t= 37. = 125, *b < 01. *p <0556 | CATANZARO AND MEARNS both the NMR and BDI with the ISB; and third, we examined the relationships among the NMR, the two BDI assessments, and measures of emotional states taken at the second session. ‘We expected the NMR: (a) to predict depression scores at Time 2 and (b) to be related to negative emotions. In addition, we expected that the NMR and BDI would display different patterns of relations with affect and adjustment meas- tures. We based this reasoning on two things: first, on evidence that depression is emotionally complex and includes the experience of negative emotions other than sadness, such as shame and guilt (zard, 1972; Watson & Clark, 1984; ‘Watson, Clark, & Tellegen, 1988). Indeed, the BDI includes items that specifi- cally assess guilt, irritability, and self accusation (Beck, Steer, & Garbin, 1988). On the other hand, the NMR items, although they refer to the genetal feeling of being “upset,” emphasize feelings of sadness more than other unpleasant emo- tions, We thoughtit unlikely, therefore, that the NMR would predict emotions such as guile, shame, disgust, or contempt. We expected that the NMR would predict sadness measured at Time 2 and that the BDI would be related to saciness ‘he well as a number of additional negative emotions, including guilt and shame, at Time 2, ‘Second, because the BDI elicits reports of specific symptoms of psychopathol- ogy, high BDI scores are indicative of poor adjustment. In contrast, NMR items Go not elicit symptom reports; thus, we expected that the BDI, but not the NMR, would be correlated with the ISB. ‘Conrelations among the NMR, ISB, and the two BDI administrations are presented in Table 5. Replicating and extending the results obtained with Sample 4, the NMR correlated significantly with both BDI scores. However, the NMR alko correlated with a broad adjustment measure, the ISB. By using T2 (Williams, 1959) to compare the correlation between the NMR and the ISB to the correlation between the BDI given at the same session and the ISB, we Showed that chese were significantly different, T;(73) = —4.74, p < .05. This TABLES Correlations of NMR Scores With Adjustment and Two Assessments of Depression ‘Sample 5 NMR. 1SB BDI-L ‘BDI-2 NMR = 1B - BDI-1 sie - BDL2 2 386 138 S BOR rn ed SB Noa N= 7 men NMR = Generalized Expectancy for Negative Mood Regulation Sealey ISB = Roter Incomplete Sentences Blank; BDI-1 = Beck Depression Inventory fist administrations BDL-2 = Beck Depression Inventory, second administration. p< 05. "4p < 001,NEGATIVE MOOD REGULATION 557 finding is consistent with our view that the NMR scale, although related to depression and other adjustment outcomes, is not a measure of adjustment. The correlations of the NMR and each BDI score with DES scores are presented in Table 6. As hypothesized, the BDI was strongly related to a number of negative emotions; the NMR was only related to sadness, In summary, the results of these analyses indicate that although the NMR is correlated with depression, it clearly has a different pattern of relations with related variables than the BDI does. These data also suggest that the NMR scale is not a proxy measure of negative affect (Watson et al., 1988) and provide evidence for the discriminant validity of the NMR from measures of depression, GENERAL DISCUSSION These studies were intended to examine the utility of a preliminary pool of theoretically derived items as a measure of generalized expectancies for negative mood regulation. The 30-item NMR scale has acceptable internal consistency, temporal stability, and discriminant validity from social desirability and locus of control. Generalized expectancies for negative mood regulation and depression are inversely related, but measures of these two constructs display different Patterns of relations with a broad adjustment measure as well as measures of affective states, suggesting that they are distinct constructs. More recent data lend further support to the NMR's discriminant validity from depression. Kirsch, Mearns, and Catanzaro (in press) found that stressful life events, a powerful predictor of depressive and physical symptoms, was not TABLE 6 Conelations of NMR Scores and Depression With Self-Reported Emotions—Sample 5 reach aaa va it wie a eo ee Emotion NMR BDI-1 BDI-2 Interest, 19 05 or ‘Surprise -02 o 20 dey 15 1 =05 Sadness ~28" cee ‘s8eee ‘Anger =.01 05 20 Disguse -.06 18 370 Fear ~.10 308 foe Guile 10 agree 55t Contempe ary 28% ad Shame 2 30% ee tt _N 5, 17 men. NMR = Generalized Expectancy for Negative Mood Regulation Sle EDEL = Beck Deprenon Inventory, fst administration; BDL-2 = Beck Depeson Inventory second administration, Emotions are from the Differential Emotions Scale, which was administered with the second administration of the BDL "p< .5 p< .O1. 4? < 001,558 CATANZARO AND MEARNS: correlated with the NMR at all. Mearns (1989) found that the NMR adds to the prediction of changes in depression over time even when controlling for initial levels of depression. 'A number of theoretical, empirical, and practical issues remain. Before con- sidering the implications of our results for further research, we wish to clarify the relation between negative mood regulation expectancies and three related con- structs: negative affectivity (Watson & Clark, 1984), self-control (Rosenbaum, 1980), and anxiety sensitivity (Reiss, Peterson, Gursky, & McNally, 1986): ‘Negative affectiviry (NA) is defined as the disposition to experience unpleasant mood states. Expectancies for negative mood regulation concern one’s ability to terminate unpleasant mood states, rather than the tendency to experience them in the first place. Although we expect a modest empirical relation between measures of the two constructs, they lend themselves to different predictions. In future studies, we expect the NMR to predict attempts to terminate negative moods and the duration of negative moods following upsetting events that are likely to induce them in most people; in contrast, measures of NA should predict the onset and intensity of negative moods, but might not be related co attempts to terminate the negative moods or the efficacy of those attempts. Although our data do not address these issues, the finding that NMR scores predict sadness cores but not other aspects of NA is an initial indication that NMR scores are not trait NA scores. "The SelfControl Schedule (SCS) was developed by Rosenbaum (1980), to measure the behavioral construct of selfcontrol. Both the SCS and the NMR spring from the notions: (a) that individuals can use a variety of cognitive and Lehavioral strategies to assert control over negative moods, and (b) that individuals are likely to vary in the degree to which they attempt to do so. However, in contrast to our focus on the assessment of the belief that one can effectively terminate an unpleasant mood, the SCS focuses on the frequency ‘vith which the person relies on various self-control strategies by means of inquiring about the strategies! descriptiveness of the respondent. Despite few SCS items that indizectly assess expectancies similar to those assessed by the NMR scale (e-g., “Often by changing my way of thinking I am able to chanige!my feelinye about almost everything’), the SCS appears to give a picture of the individual’ style and level of coping behavior, rather chan expectancies for tnood regulation. In this respect, the SCS may be more similar to measures of oping behavior (cf. Moos, Cronkite, Billings, & Finney, 1983) than to the NMR scale. Furthermore, itis likely chat the NMR will not predict a number of different criteria that the SCS does but will predicr some criteria that the SCS Goes not. For example, SCS. scores have been related. to. procrastination (Milgram, Sroloff, & Rosenbaum, 1988); however, expectancies for negative rnood regulation chould predict sych criteria (ie. criteria affected by planning, goal serting, and selfreinforcement) only when they are a funcson of the aeration of a negative mood state. The NMR should prove more useful than theNEGATIVE MOOD REGULATION 559 SCS in predicting the attempts to alleviate the adverse emotional effects of negative life events and the impact of affect-mediated consequences of these events. Anxiety sensitivity is defined as the fear of fear: the belief that the experience of anxiety or fear will cause further anxiety, illness, ot embarrassment. Reiss et al. (1986) developed the Anxiety Sensitivity Index (ASI) as a measure of these expectancies and demonstrated their role in the experience of anxiety. Although these expectancies are clearly related to generalized expectancies for negative mood regulation, an important distinction can be drawn. The ASI does not assess the person's expectancy that anxious states can be effectively terminated likewise, the NMR does not assess the expectancy that additional negative outcomes will follow negative moods. The ASI is most likely to prove useful as predictor of the intensity and duration of anxious states; again, we emphasize that the NMR scale is intended to predict the number of attempts to terminate negative moods and the success of these mood regulation attempts. Because Problems arising from the experience of anxious states are probably a function of both sensitivity and mood regulation expectancies, the NMR and the ASI will probably provide joint prediction of a number of criteria. Implications for Research Research on the construct we are measuring should contribute to an under- standing of stress and coping. A specific hypothesis consistent with the cognitive-phenomenological framework of Lazarus and Folkman (1984) is that the NMR scale measures a variable influencing a person’s secondary appraisal of one’s coping resources when emotionfocused coping is possible or required. Although we have focused on emotional outcomes, itis likely that generalized expectancies for negative mood regulation will also be related to somatic Outcomes of stress. ecause emotion-focused coping strategies can affect, for better or worse, patterns of physiological disturbance as well as the behavioral choices made to adapt to physiological disturbance (Lazarus & Folkman, 1984, Pp. 214-219), beliefs concerning one’s ability to alleviate negative emotional states can play an important role in this process. Consistent with these hypotheses, Kirsch etal. (in press) found that not only do NMR scores predict coping behavior, they have direct and indirect (¢hrough coping attempts) effects on self-reported coping outcomes of depression and physical symptoms, That is, the belief that one can regulate negative moods is in itself associated with more adaptive outcomes, in addition to its effects via coping behavior. Mearns (1989) found that the NMR and initial levels of depression interacted to predict changes in depression over time as subjects were recovering from the end of a romantic relationship. That is, prediction of emotional recovery from a significant loss was enhanced by inclusion of the NMR.560 | CATANZARO AND MEARNS The NMR is related to self-report measures of depression and sadness. These might be termed broad outcomes ofthe mood regulation process. Research is also needed to examine the relation between expectancies for negative mood regu- lation and the aspects of the ongoing process of mood regulation itself. In particular, there may be “controlled” cognitive activities in which successful ‘mood regulators engage that counteract apparently “automatic” cognitive effects of dysphoric mood states (cf. Blaney, 1986; Bower, 1981; Hasher & Zacks, 1979). Engaging in such activities should be at least partially a function of the expectancy of their success. High and low NMR scorers may differ in the initial intensity of negative moods, the speed with which they recover from these, or both. It should be noted that, as with any broad measure of generalized expectancy, we expect the NMR to predict such criteria in conjunction with situational parameters. There are probably situations in which even the most confident mood regulator would be hard pressed to be successful and indeed may not expect or, possibly, want to be successful. One example of a situation that might preclude mood. ‘regulation artempts despite strong generalized expectancies for cheir success might be an experiment in which subjects are exposed to a strong and ‘obvious negative qood induction procedure and are exhorted to continue to feel the negative moed (as is often done in research investigating effects of mood on cognitive processes). Thus, it is more appropriate to use nonobvious or ambiguous mood Faduction procedures, such as films and tape-recorded stories with positive and negative elements, to test hypotheses about the role of generalized expectancies in the mood regulation process. Examples of naturally occurring siguations in which med regulation efforts might not immediately be appropriate are likely to be found when individuals are grieving deaths andiother losses or reacting to teaumatie stress. High NMR scorers are not hypothesized to be immune to the slings and arrows of lifes. misfortunes; they are expected to eventually take cognitive and behavioral arms against such troubles and recover more quickly than low scorers, however. Cognitive theories of personality, emotion, and therapeutic change are consistent with our view: that what people think will happen when they experience negative emotions has important implications for their experience of those states and the development ‘and treatment of a number of psychiatric and physical disorders. Developing a more thorough understanding of processes vslved in newitive mood regulation, individual differences in this process, and their treatment inplications appear to be important and fruitful directions for farcher research. ‘ACKNOWLEDGMENTS Portions of this article were presented at the annual convention of the American Psychological Association, August 1987.NEGATIVE MOOD REGULATION 561 We thank Kevin Corcoran, Laurie Pearlman, Elizabeth Portman, Janis Phelps, J. Conrad Schwarz, and especially Julian B. Rotter and Irving Kirsch, for their helpful comments. We also thank Robin Coltey, Gregory Greenwood, Gene Herrmann, Jill Mattia, Leslie McGovern, Colleen Quinlan, and Michael Silverstein for their assistance with various phases of data collection. REFERENCES Beck, A. T., Rial, W. Y., & Rickels, K. (1974). Shore form of depression inventory: Cros validation. Peychological Reports, 34, 1184-1186, Beck, A. T., Rush, A. J. Shaw, B. F.,& Emery, G. (1979). Cognitive therapy of depression. New York: Guilford. Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Peychometric properties of the BDI: 25 years later. Clinical Pychology Review, 8, 77-100. Blaney, P. H. (1986). Affect and memory: A review. Prychologial Bulletin, 99, 229-246. Bower, G. H, (1981). Mood and memory. American Psychologist, 36, 129-148. Burns, D. D., Shaw, B.F., & Croker, W. (1987). Thinking sles and coping strategies of depression women: An empirical investigation, Behaviour Research and Therapy, 25, 223-225. Crowne, D. P., & Marlowe, D. (1964). The approval motive. New York: Wiley. Doerfler, L. A., & Richards, C. 5. (1981). SelFiniiated attempts to cope with depression. Cognitive ‘Therapy and Research, 5, 367-371. Franko, D. L., Powers, T. A., Zuroff, D. C., & Moskowite, D. S. (1985). Children and affect: Strategies for self-regulation and sex differences in sadness. American Journal of Orthopsychiatry, 55, 210-219, Hasher, Ly & Zacks, R. T. (1979). Automatic and effortul processes in memory. Joumal of Experimental Peychology: General, 108, 356-388, lard, C. E. (1972), Patterns of emotions. New York: Academic. Kirsch, 1. (1985). Response expectancy as a determinant of experience and behavior. American Pychologist, 40, 1189-1202, Kirsch, L., Mearns, J., & Catanzaro, S.J. in press). Mood regulation expectancies as determinants ‘of depression in college students. Journal of Counseling Psychology. Lah, ML. (1985). New validity, normative, and scoring data for the Rotter Incomplete Sentences Blank. Journal of Personality Assessment, 53, 607-620. Lazarus, R. S., & Folkman, S. B. (1964), Stress, appraisal, and coping. New York: Springer. Mearns, J. (1989). Expectancy for negative mood regulation and mood recovery following the end of a ‘romantic relationship. Unpublished doctoral dissertation, University of Connecticut, Storrs Milgram, N. A., Srolof, B, & Rosenbaum, M. (1988). The procrastination of everyday lf, Journal of Research in Personality, 2, 197-212. Mischel, W. (1973). Toward a cognitive social learning reconceptualization of personality. Poycho- logical Review, 80, 252-283, Moos, R. H., Cronkite, R. C., Billings, A. G,,& Finney, J. W. (1983). Health and Daily Living Form ‘manual. (Available from Social Ecology Laboratory, Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, CAA 94305) Mosher, D. Ly & White, B. B. (1982). On differentiating shame and shyness. Motivation and Emotion, 5, 61-74. Phares, E. J. (1976). Locus of contol in personality. Mortistown, NJ: General Learning Press. Reiss, S.J, Peterson, R. A., Gursky, D. M., & McNally, R. J. (1986). Anxiety sensitivity, anxiety frequency, and the prediction of fearfulness. Behaviour Research and Therapy, 24, 1-8562 CATANZARO AND MEARNS Rippere, V. (1976). Antidepressive behaviour: A preliminary report. Behaviour Research and Therapy, 14, 289-299. Rippere, V. (1977).“What’s the thing to do when you're feling depressed?”— A pilot study. Behaviour Research and Therapy, 15, 185-191. Rippere, V. (1979). Scaling the helpfulness of antidepressive activities, Behaviour Research and ‘Therapy, 17, 439-449. Rosenbaum, ML. (1980). A schedule for assessing self-control behaviors: Preliminary findings. Behavior Therapy, 11, 109-121. Roeter, J. B. (1954) Socal learning and clinical prychology. Englewood Cis, NJ: Prentice Hall Rotter,], B. (1966), Generalized expectancies for internal versus external control of reinforcement. aychalogical Monographs, 80(1, Whole No. 605). Rotter, B. (1975). Some problems and misconceptions related to the construct of internal versus ‘external contcol of reinforcement. Journal of Consulting and Clinical Paychology, #3, 56-67. Rotter, J.B. (1982). The development ana application of socal learning theory: Selected papers. New York Praeger Rotter, J.B Chance, J. E., & Phares, E. J. (Eds). (1972). Applications of a social leaning theory of ‘personality. New York: Holt, Rinehart & Winston, Rotter, J. B., & Rafferty, J. (1950). The Rete incomplete sentences blank: College form. New Yorks aychological Corporation. Worthime, R. G., & Foulds, M. F. (1971). Perceived locus of control and personal adjustment. “Journal of Consulting and Clinical Peychology, 37, 250-252. ‘Wateon, D. A» & Clark, L. A. (1984). Negative affectivity: The disposition to experience negative ‘emotional sates. Prchological Bulletin, 98, 219-235. Watson, D. A., Clark, L.A., & Tellegen, A. (1988). Development and validation of brief scales of positive and negative alfece: The PANAS Scales. Journal of Personality and Socal Pcholog, 13, 901-914. ‘Watson, RL. (1978). The eentence completion method. In B. B. Wolman (Ed), Clinial diagnosis of ‘mental disorders: A handbook (pp. 255-279). New York: Plenum. Williams, EJ (1955). The comparison of regression variables. Jounal ofthe Royal Statistical Societ, Series B, 21, 396-399, Zurofl,D.C. (1980). Learned helplessness in humans: An analysis of learning processes and the roles ss individual and situational differences. Journal of Personality and Social Pycholog, 39, 130-146, Salvatore Catanzaro Ilinois State University DeGarmo Hall 435, Normal, IL 61761 Received March 1, 1989 Revised May 30, 1989 APPENDIX Instructions for the NMR Scale a The Attitudes Toward Feelings Scale This is a questionnaire to find out what people believe they can do about upsetting emotions or feelings. Please answer the statements by giving as true aNEGATIVE MOOD REGULATION 563 picture of your own beliefs as possible. Of course, there are no right or wrong answers. Remember, the questionnaire is about what you believe you can do, ‘not about what you actually or usually do. Be sure to read each item. carefully and show your beliefs by marking the appropriate number on. your IBM answer sheet. If you strongly disagree with an item, fill in the space numbered 1. Mark the ‘space numbered 2 if you mildly disagree with the item. That is, mark the space numbered 2 if you think the item is more generally untrue than true according to your beliefs. Fill in the space numbered 3 if you feel the item is about equally true as untrue. Fill in the space numbered 4 if you mildly agree with the item. That is, mark number 4 if you think the item is more true than untrue. If you strongly agree with an item fill in the space numbered 5. 1. Strongly disagree 2. Mildly disagree 3. Agree and disagree equally 4. Mildly agree 5. Strongly agree Please be sure to fill in the spaces completely and to erase completely any marks to be changed. Make no extra marks on either the answer sheet or the questionnaire. ence a ne nner nTTETTEENESEEEEIESEY Instructions for other reporting formats are available from S. J. Catanzaro (the first author).