Association between daily coping and end-of-day mood.

Copyright 1995 by the American Psychological Association, Inc. 0278-6133/95/S3.00 Health Psychology 1995, Vol. 14, No. 4,341-349 Association Between Daily Coping and End-of-Day Mood Arthur A. Stone, Eileen Kennedy-Moore, and John M. Neale State University of New York at Stony Brook Research indicates that self-reports of coping with stressful occurrences are associated with psychological and health outcomes. However, measures of coping may be biased by retrospective distortion as they assess coping over relatively long periods. In this study, a sample of 79 men completed a coping assessment daily for several weeks about the day's most "bothersome" problem. Repeated daily measurement of coping allowed analysis of within-subject effects of coping efforts. Same-day mood reported by the men (targets) and reports of the men's mood by their spouses (observers) were outcome variables. Within-subject analyses indicated that catharsis and social supports were associated with increased negative affect, whereas use of acceptance was associated with less negative affect. Use of distraction, acceptance, and relaxation were associated with increased positive affect. These findings held for both target- and observer-reported mood. Key words: coping behavior, mood, positive affect, negative affect, daily coping assessment, self-reported/observer-reported mood list is a sample of the large number of conditions and outcomes that have been the focus of coping research. Research on coping has blossomed during the last decade with advances in both theory (e.g., Lazarus & Folkman's, 1984, transactional stress and coping model) and methods for measuring coping activities with self-report, situation-specific inventories (such as the Ways of Coping Inventory, Folkman & Lazarus, 1980). The transactional model is one that, in general terms, is adhered to by many stress and health researchers. Coping is an integral part of the model: It is a potential modifier of the effect of stressful environments and threatening appraisals on subsequent physiological and health outcomes. For this reason, studying coping is important for health researchers. Dozens of studies have been conducted that have explored associations between coping and health outcomes. For example, there are studies on coping with asthma (Deenan & Klip, 1993), temporomandibular joint osteoarthrosis (Jaspers, Heuvel, Stegenga, & de Bont, 1993), post-coronary-bypass surgery (Agren, Ryden, Johnsson, & Nilsson-Ehle, 1993), rheumatoid arthritis (Manne & Zautra, 1990), cancer (Ettinger & Heiney, 1993), irritable bowel disease (Kinash, Fischer, Lukie, & Carr, 1993), premenstrual syndrome (Pirie & Halliday Smith, 1992), the threat of AIDS (Leserman, Perkins, & Evans, 1992), coronary heart disease (Denollet & de Potter, 1992), chronic diseases (Wikblad & Montin, 1992), genital herpes (Keller, Jadack, & Mims, 1991), chronic fatigue syndrome (Blakely et al., 1991), diabetes mellitus (Kvam & Lyons, 1991), epilepsy (Snyder, 1990), organ transplants (Voepel-Lewis, Starr, Ketefian, & White, 1990), and on the physiological correlates of coping (Ursin & Olff, 1993). This Assessing Coping on a Daily Basis In this study, we argue for examining coping in a more intensive manner than has typically been done. Previous studies have usually had participants report on their coping efforts for a period of from 1 week to 1 month. Recognizing the relatively specific detail requested by most coping questionnaires (e.g., how often participants were thinking particular thoughts), there is likely to be considerable distortion of reporting in this retrospective procedure (Stone & Shiftman, 1992, 1994). Bias will depend upon the degree of problem resolution at the time of questionnaire completion as well as upon the simple inability to accurately remember particular thoughts and actions over relatively long periods (Stone, Greenberg, Kennedy-Moore, & Newman, 1991; Stone & Shiftman, 1994). Recently, Ptacek, Smith, Espe, and Rafferty (1994) have shown that there is even substantial degradation of recall when coping is assessed several days after a stressful event, as opposed to when coping is assessed concurrently, thus confirming concerns about retrospective distortion in coping assessment. Therefore, the need for more accurate data concerning the effects of coping requires a fine-grained assessment. In this report we present the results of a longitudinal study of coping with daily problems and how coping efforts are related to mood recorded at the end of the day. Although still retrospective over the period of the day, these daily reports of coping should be much less prone to reporting bias than are assessments of coping over the previous week or month. We measured coping with an instrument developed specifically for daily administration (Stone & Neale, 1984). Instead of presenting a series of many specific questions, this questionnaire first asks participants to indicate the most bothersome problem of the day and then whether they handled it with any thoughts or Arthur A. Stone, Department of Psychiatry, State University of New York at Stony Brook; Eileen Kennedy-Moore and John M. Neale, Department of Psychology, State University of New York at Stony Brook. Correspondence concerning this article should be addressed to Arthur A. Stone, Department of Psychiatry, State University of New York at Stony Brook, Stony Brook, New York 11794-8790. Electronic mail may be sent to [email protected]. 341 342 A. STONE, E. KENNEDY-MOORE, AND J. NEALE actions that fit into the following eight broad conceptual categories of coping (Stone & Neale, 1984; Schwartz & Stone, 1993): distraction, situation redefinition, direct action, catharsis, acceptance, seeking social support, relaxation, and religion. We chose these coping categories because for each of them there is at least some evidence that they can moderate the effects of life stress (see Stone & Neale, 1984). However, in the context of a study of daily coping, predictions about the effects of each coping type are more difficult to make. For some of the coping categories being assessed here (situation redefinition, catharsis, relaxation, and distraction), there is experimental evidence that suggests that they can alter negative affect (NA) in the short term (e.g., Benson, 1975; Josephs & Steele, 1990; Pennebaker & Beall, 1986); but for several of the other coping modes, the only data available concerning their effects come from the study of longer term outcomes. Indicators of Coping There are many kinds of outcomes that coping researchers have considered, such as short-term well-being effects, health outcomes, and achievement of particular goals. We have chosen to explore same-day mood because it may be thought of as a proximal indicator of processes linked to distal health outcomes. More proximal measures of outcome have a distinct advantage over more distal measures of outcome: The former should have stronger associations with coping because there are fewer opportunities for intervening factors to modify the effects of coping. On the other hand, the effects of coping may be different for short- and long-term outcomes: Achieving desirable short-term outcomes may impair the achievement of successful long-term outcomes. For example, ignoring a lump in one's breast may reduce immediate NA but in the long run may allow for a life-threatening disease to progress. However, the use of cathartic coping might produce a same-day increase in NA, yet also be linked to subsequent, favorable mood and health outcomes (e.g., Pennebaker & Beall, 1986). Unfortunately, very little is known about the effects of coping efforts on outcomes measured at different time intervals from the coping efforts. In terms of the role of mood in health, a strong case can be made for the effect of mood on physiological processes that could influence disease. Conceptual models describing the potential pathways between psychosocial factors (e.g., stressors) and health outcomes place mood in a central role. In their general model, which, we believe, reflects current thinking about psychobiological mechanisms, Herbert and Cohen (in press) have mood mediating the effects of environmental demands and of the appraisal process on both biological and behavioral responses, which in turn affects risk of health disorders. There is also empirical evidence that negative moods are related to corticosteroid and catecholamine levels, immunological measures, and autonomic nervous system functioning (Elliot & Eisdorfer, 1982), all of which have the potential for influencing susceptibility to illness. In this article, we consider the possible differential effects of coping on positive affect (PA) and NA. Most studies of coping have observed its effects on reducing negative outcomes but have not examined its effect on the orthogonal variable of PA. Yet research is showing that changes in PA are important in their relation to health outcomes (Diener & Emmons, 1984). We pose the following question: Are some forms of coping associated with the decreased NA, increased PA, or both, assessed at the end of the day? Because participants were studied for several weeks, we could conduct ipsative analyses of the association between coping and mood that address the question of whether coping is associated with mood across an individual's reporting. Interestingly, there is no required statistical correspondence between nomothetic analyses, which examine between-subject relationships, and ipsative analyses as they address very different questions (Michela, 1990). In other words, the fact that a strong, nomothetic, positive correlation between direct action coping and mood is observed in a sample of participants does not mean that the day-to-day, within-subject association will also be positive; it is, in fact, possible for the variables to be negatively correlated. A final feature of this article is our analysis of both self- and other-reported mood. One problem in the coping area, even with daily reports, is that the outcome of the coping efforts are known to the participant at the time the assessments are completed and may therefore bias the coping reports. In this study, husbands were the target participants and reported their own mood and coping. Wives' reports of their husbands' moods were used to lessen the possible biasing effect just described. Method Participants Married couples were solicited for participation in this study by means of letters of introduction mailed to households within 10 miles (16 km) of our university. It is difficult to calculate the acceptance rate of participants into the study because we do not know what percentage of households receiving letters were actually eligible for the study by virtue of being occupied by a married couple. Given the arduous nature of the study and the effect that it probably had on participants' interest, the sample is not likely to be representative of the general population. Of the 150 participants who initially began the daily recording procedures, 71 (47%) were eliminated from the analyses presented here because they did not meet the criterion of having completed at least 82 days of data (noncompleters), the minimum we felt necessary for the analyses. The final sample of 79 men (completers) reported for an average of 109 days and missed 6.3% of days they could have reported. The sample had an average age of 43.2 years (SD = 10.1 years)1, an average family income of $30,965 (SD = $10,532; 1980 dollars), and an average of 13.7 years (SD = 2.2 years) of education. The demographic characteristics of the completers and noncompleters were compared to determine if there was a systematic bias in participants selected for the analyses. Of the five comparisons that were made (age of husband and wife, number of years married, number of children, number of years of education of husband and wife, and household income), only one was significant at the .05 level: Completers were married an average of 3.5 years longer than noncompleters (16.6 vs. 13.2). 1 One participant refused to report his age, so the sample size for this mean is 78. ASSOCIATION BETWEEN DAILY COPING AND END-OF-DAY MOOD Materials 343 Table 1 Description and Labels of the Eight Coping Categories A coping assessment was part of the daily questionnaire booklet, which also included a mood assessment, a daily event checklist and appraisal section, and questions about physical symptomatology. There were two versions of the booklet, one for husbands (targets) and one for wives (observers). Targets reported about their own mood, and observers also reported about targets' mood, providing another view of mood based more on observable facets of affect and less on its experiential component (see Stone, 1981, for a more thorough discussion of this procedure). Information about coping efforts was answered by both targets and observers about themselves. Because this article focuses on the relationship between coping and mood, only targets' coping efforts are considered, inasmuch as observers did not report about their own mood. The mood measure was the revised 12-item version of the Nowlis Mood Adjective Checklist (MACL; Nowlis, 1965). Because of low factor loadings, one item was dropped, leaving the Negative Engagement factor with 5 items and the Positive Engagement factor with 6 items. NA includes original Nowlis mood adjectives of "angry," "clutched up," "concentrating," "skeptical," and "sad," while PA includes "playful," "elated," "energetic," "kindly," "self-centered," and "leisurely" (Stone, 1981). Each MACL item was rated on a 3-point scale, with 1 = definitely applies, 2 = slightly applies, and 3 = does not apply. To make the mood measures comparable, the MACL items were reversed by subtracting them from 4, so that larger numbers meant higher scores on both mood scales. These reverse-scored items were summarized into positive and negative mood scales (see Stone, 1981). The assessment of coping was the final section of the questionnaire. The development of the instrument was described in detail in a previous article (Stone & Neale, 1984). Participants were asked to write a brief description of "the most bothersome event or issue of the day." The instructions stated that these problems could be something that happened that day, something that happened in the past, or something that was anticipated to happen in the future. Overall, only 3% of the reported problems happened in the past and fewer than 5% were anticipated. Eight appraisal questions followed the narrative description. Control over the problem's occurrence was answered with a 3-point scale (1 = no control, 2 = some control, and 3 = quite a lot I complete control); undesirability was answered with a 6-point bipolar scale (1 = extremely desirable, 2 = moderately desirable, 3 = slightly desirable, 4 = slightly undesirable, 5 = moderately undesirable, and 6 = extremely undesirable); the degree to which it changed or stabilized their lifestyle was also answered with a 6-point bipolar scale (1 = extremely changing, 2 = moderately changing, 3 = slightly changing, 4 = slightly stabilizing, 5 = moderately stabilizing, and 6 = extremely stabilizing); degree of anticipation was answered with a 4-point bipolar scale (1 = completely unexpected, 2 = somewhat unexpected, 3 = somewhat anticipated, and 4 = completely anticipated); meaningfulness of the problem was answered with a 3-point scale (1 = extremely meaningful, 2 = moderately meaningful, and 3 = slightly meaningful); whether the problem was a single event or more chronic was answered with a 2-point scale (1 = single event and 2 = long lasting); and whether the situation had ever happened before was answered with a yes-no scale. Participants also rated the problem's "stressfulness" using a 0- to 100-point scale. The final section of the instrument asked participants what they did "to try to feel better or handle the problem" that was just described. Eight statements followed that described different broad categories of coping. The statements (included in the questionnaire) and our descriptive labels (not printed in the questionnaire) are found in Table 1. Using a yes-no scale, participants indicated whether they used any actions or thoughts falling into each category to handle the problem. When yes was chosen, participants were requested to describe the Descriptive labels Category 1. Distraction 2. Situation redefinition 3. Direct action 4. Catharsis 5. Acceptance 6. Seeking social support 7. Relaxation 8. Religion Diverted attention away from the problem by thinking about other things or engaging in some activity Tried to see the problem in a different light that made it seem more bearable Thought about solutions to the problem, gathered information about it, or actually did something to try to solve it Expressed emotions in response to the problem to reduce tension, anxiety, or frustration Accepted that the problem had occurred, but that nothing could be done about it Sought or found emotional support from loved ones, friends, or professionals Did something with the explicit intention of relaxing Sought or found spiritual comfort and support particular action or thought on several blank lines provided directly below the statement. We have previously shown (Stone & Neale, 1984) that participants' narrative descriptions are appropriate for the category selected. An additional, ninth, category labeled Other was provided to allow the recording of any other actions or thoughts used to handle the problem that did not fit into the eight categories. Procedure Participants were trained at home to use the coping materials by reading over the descriptions of the coping categories with a research assistant. Additional examples of coping behaviors for each category were presented to ensure that participants understood the meaning of the coping descriptions. Participants completed a coping assessment for the previous day in the presence of the research assistant and then discussed their responses. The assistant collected demographic information, including sex, age, education, and family income. Participants were given a supply of questionnaire booklets. They completed a booklet at the end of each day and mailed the prestamped booklet to us on the following day. Results Analytic Method An analysis that proceeded directly to the question of coping and mood could be misleading because of potential confounding by the characteristics of the problem. Perhaps the most important characteristics are the subjective impressions or appraisals an individual has about the problem; these have been demonstrated to affect the type and frequency of coping (Schwartz & Stone, 1993; Stone & Neale, 1984). For instance, compared with the effects of less stressful problems, more undesirable or stressful problems are likely to both evoke more coping efforts and result in higher levels of negative mood. If undesirability is not controlled in an analysis of coping and mood, incorrect conclusions might be drawn, such as more 344 A. STONE, E. KENNEDY-MOORE, AND J. NEALE coping being associated with worse mood. Although this might actually be true, the reason for the association is that highly undesirable problems are associated with more coping and more negative mood. Thus, appraisal of problem desirability is covaried out in all analyses. The specific type of statistical technique used here is a form of least squares regression that has demonstrated utility in the analysis of multilevel data. Two levels exist in the daily data: between-subjects and within-days associations. To remove the potentially confounding effects of differential reporting levels by participants, all between-subjects variance is removed from both independent and dependent variables prior to exploration of within-subject associations (see Bolger, DeLongis, Kessler, & Schilling, 1989; Repetti, 1989; Schwartz & Stone, 1993). In essence, this is accomplished by creating a set of dummy-coded variables, one for each participant except the last, and entering these variables into the regression before any other predictors are entered.2 All between-subject variance attributable to persons is thereby removed, yet the original within-subject variability is retained. Furthermore, because it is likely that these data are autocorrelated (eig., mood on one day is related to mood on the prior day), some of the regression models that we present included a term for removing the effects of the prior day's level of the outcome variable. Finally, it is possible that the effect of coping on mood is not constant for all participants. Therefore, we also examine the interaction between participants and coping in some analyses; a significant interaction term means that the effect of the coping varies among participants. In order to understand the relationships among the uses of the eight coping modes, we computed correlations, in addition to the regressions, for each participant among all of the coping modes. These coefficients were then averaged, and the mean correlations are presented. However, for statistical testing, coefficients were Fisher-Z transformed in order to normalize their distributions prior to taking their average. These means were then tested for deviation from zero with a t test (J. Cohen & Cohen, 1983). Appraisal and Coping Although there were eight questions tapping problem appraisal included in the assessment, we focus on the concept most strongly associated with affect: perceived undesirability (Schwartz & Stone, 1993). Although most problems were rated on the undesirable pole of the bipolar desirable /undesirable response key, a small number were not and were dropped from the analyses. We contrast the probability of reporting a type of coping for the three levels of undesirability (slightly, moderately, extremely), with individual-difference effects already removed. Table 2 presents the proportion of days targets used each coping mode broken out by level of undesirability. Frequency of use of distraction, acceptance, relaxation, and religion was not related to the level of problem undesirability. In contrast, frequency of use of direct action, catharsis, and seeking social supports were monotonically related to increasing undesirability, F(2, 6929) = 8.63, 36.23, and 6.41, p < .001, respectively, while the frequency of use of situation redefinition was Table 2 Proportion of Days Coping was Used for Low, Medium, and Highly Undesirable Problems Undesirability Coping mode Low Medium Distraction Situation redefinition Direct action Catharsis Acceptance Seeking social support Relaxation Religion .23 .22*** .48*** .35*** .28 .16*** .14 .03 .23 .20*** .53*** .46*** .28 .19*** .14 .03 High .23 .17*** .54*** .55*** .28 .24*** .15 .04 ***p < .001. inversely related to undesirability, F(2,6929) = 2.18,p < .001. Because of these associations between problem appraisal and coping usage, level of undesirability is statistically controlled in subsequent analyses. Intercorrelations Among Coping Modes Before computing the regressions between appraisals and coping predicting mood, we examined the relationships among the daily usage of the eight coping modes. Although the aggregated correlations shown in Table 3 are not very large, they do suggest that some degree of multicollinearity will be present in simultaneous regressions of mood on all eight coping modes. Note that it is the average correlations (with each participant contributing, if possible, one correlation to the average) that are presented in Table 3; tests for significance are based on the standard error of the mean of the Z-transformed correlations (with the degrees of freedom based on the number of participants, not the total number of days). Appraisal and Coping Predicting Target-Reported Mood Two sets of regressions were computed in which NA and PA were separately predicted from undesirability and the eight coping modes (with between-subjects effects removed first). In all of these regressions there was, as expected, a strong effect of undesirability on mood, for NA, F(3, 7554) = 102.56, p < .0001, and for PA, F(3, 7549) = 46.54, p < .0001, such that increases in undesirability were associated with more NA and less PA. Because the coping modes are correlated, two strategies were used to understand the effects of coping on mood. First, to assess the effect of each coping mode without removing effects shared with other coping modes, a set of regressions was computed in which undesirability and a single coping mode predicted mood (target- and spouse-rated). The second set of regressions was simultaneous with regard to the coping modes. These regressions reveal the unique effects of each coping mode with the remaining coping modes statisti2 The statistical software SAS has a procedure called GLM that automatically generates the matrix of dummy-coded variables. The GLM option is "Absorb," and in addition to being convenient, it is a much more efficient way to compute the regression. 345 ASSOCIATION BETWEEN DAILY COPING AND END-OF-DAY MOOD Table 3 Ipsatized Correlations Among Coping Modes Coping mode 1. 2. 3. 4. 5. 6. 7. 8. Distraction Situation redefinition Direct action Catharsis Acceptance Seeking social support Relaxation Religion 4 5 6 7 8 -.04" -.05" -.06" .00 -.01 .09" -.03 -.31" -.02 .03 .01 .06 .00 -.06" — .01 — 3 1 2 — .01 — -.20" -.17" — — .01 -.03 .01 .01 .06 — .09" .02 .01 .12" .03 — Note. The Ns for these aggregated correlations range from 44 to 78 because some participants had no variability on a coping mode and, hence, were eliminated from the aggregated mean. **p < .01. negative coefficients represent the case where there was less mood when the coping was used. The separate regressions revealed that several of the coping modes significantly predicted mood. In general, when effects of coping are evident for both NA and PA, the effects are in the opposite direction, as indicated by the sign of the regression coefficients. Distraction and acceptance were related to less NA and more PA. Direct action and social supports were associated with more NA and less PA. There were also several cases in which a coping mode was related only to either NA or cally removed (partialed). Because shared variance does not play a role in the evaluation of statistical significance in the simultaneous modeling, our expectation was that there would be fewer significant effects of coping in the simultaneous case compared with the separate regressions. Table 4 presents the results of the separate regressions and Table 5 the simultaneous regressions. The betas in the tables represent the mean differences when the coping mode was used versus when it was not used. Positive coefficients indicate that more negative or positive mood was associated with use of the coping, whereas Table 4 Separate Regressions of Undesirability and Eight Coping Modes Predicting Negative and Positive Affect Controlling for Individual Differences Negative affect8 Coping mode Distraction Self-view Spouse view Situation redefinition Self-view Spouse view Direct action Self-view Spouse view Catharsis Self -view Spouse view Acceptance Self-view Spouse view Seeking social supports Self-view Spouse view Relaxation Self-view Spouse view Religion Self-view Spouse view Positive affectb 0 t Interaction -0.032 -0.020 -3.08" -1.95 * *** 0.067 0.050 6.55"* 4.82*** *** ns 0.004 -0.005 0.47 -0.50 ns *** 0.001 0.007 0.11 0.71 ns ns P t Interaction0 P 0.036 0.021 4.11*" 2.48* *** ** -0.030 -0.026 -3.46*** -3.10** ns ns 0.085 0.068 8.10*** 6.52*" ns *#* -0.006 -0.002 -0.56 -0.20 ns ns -0.056 -0.038 -5.97*** -4.04*** ns 0.026 0.031 0.087 0.094 6.16*** 6.77"* *** ns -0.030 -0.015 »* *## 0.069 0.054 -0.020 -0.022 0.057 0.065 -1.54 -1.76 1.99* 2.28* # * -0.023 -0.004 2.76** 3.35*** -2.14* -0.37 5.43*** 4.28"* -0.84 -0.15 ns » ** ns *** ** ns ns Note. The regression also includes undesirability of the problem and individual differences in level of report. "For self-view, N = 6,781; for spouse view, N = 6,478. Tor self-view, N = 6,778; for spouse view, N = 6,467. The interaction column indicates the significance of the interaction between the coping and subject identification number. "p < .05. "/> < .01. ***p < .001. ns 346 A. STONE, E. KENNEDY-MOORE, AND J. NEALE Table 5 Subject X Coping Interactions Simultaneous Regression Model of Undesirability and Eight Coping Modes Predicting Negative and Positive Affect Controlling for Individual Differences Negative affect" Coping mode Distraction Self-view Spouse view Situation redefinition Self-view Spouse view Direct action Self-view Spouse view Catharsis Self -view Spouse view Acceptance Self-view Spouse view Seeking social supports Self-view Spouse view Relaxation Self-view Spouse view Religion Self-view Spouse view P -.02 -.01 .01 .00 t Positive affectb P -2.11* -1.35 .06 .04 5.67*** 4.16*** 0.88 -0.30 .00 .01 0.08 0.68 -.01 -.01 -1.21 -0.81 -0.22 0.06 .02 .01 2.38* 1.23 .08 7.57*** 5.90*** .00 .06 -.04 -.03 -4.27*** -2.93** .02 .03 .07 5.34*** 6.16*** -.03 -.01 .09 .00 .06 -.02 -.02 -1.56 -1.90 .05 .04 1.62 1.82 -.03 -.01 .05 t To assess the possibility that the effect of coping was not the same for all participants, a set of dummy variables formed by the cross-product of participant identification number and coping was entered into the regression after the main effects were entered. This was done individually for each of the coping variables predicting NA and PA for both target and observer. Significant interactions were observed (see Table 4) for some, but not all, of the coping variables. For example, the relationship between catharsis and NA (self-view) is constant across participants, as indicated by the nonsignificant interaction term. Although there is not a clear pattern of significance, there are some trends in the pattern of interactions. First, for distraction and for relaxation there is strong support for Negative Affect 1.97* 2.75** n 1n VJ . 1 \J * Cothorsis -2.23* -0.52 4.82*** 3.69*** -1.02 -0.36 Note. The regression also includes undesirability of the problem and individual differences in level of report. a For self-view, N = 6,781; for spouse view, N = 6,478. bFor self-view, N = 6,778; for spouse view, N = 6,467. *p < .05. **p < .01. ***/> < .001. - 0.06 - * Soc Sup - 0.04 - x 1- # Dir Act •* 3f o 0.00 -0.02 § -0.04 in - rx] rvj -• U in (n — Distraction 1 * A LJ Relax • K n nfi O Positive Affect z Q. O O Lu o * 0.06 Relax - 0 str action LJ LJ Appraisal and Coping Predicting Observer-Reported Mood Table 4 also includes the simultaneous regression results for targets' appraisal and coping that predicted observer-reported mood. Overall, these results closely replicate the relationships observed with target-reported mood. Every strong (.001 level) relationship between coping and target-reported mood was duplicated by the observer-reported mood finding. However, there were cases in which a relationship between coping and target-reported mood that was significant at the .01 or .05 level was not corroborated by observer-reported mood; for example, distraction predicted a decrease in target-reported, but not observer-reported, NA. There were also two instances in which a stronger association was found for observer data than for target-report data: with religion and NA and with acceptance and PA. LJ IT O Religion Redef P^J 0.02 LJ H- (A PA. Catharsis and religion were associated with increased NA; relaxation had a positive association with PA. These effects are plotted according to the strength of their relationships in Figure 1. The pattern for the simultaneous regressions was remarkably similar to that of the individual regressions with one exception: The association between direct action and both NA and PA, which was rather strong in the separate regressions, was almost entirely eliminated. Also, the weak association between religion and NA was weakened even further. 0.08 0.04 - 0.02 0.00 Accept L J ^ DO 1 II B (_ I 6$d Redef DO Catharsis 1— -0.02 Dir Act ^2 Religion 1— Soc a: o LJ Sup n nd. Figure 1. Magnitude of standardized regression coefficients for regressions in which target-reported negative affect and positive affect were regressed on problem stressfulness and each coping mode. Redef = situation redefinition; Dir Act = direct action; Soc Sup = social support; Relax = relaxation. *p < .01. ASSOCIATION BETWEEN DAILY COPING AND END-OF-DAY MOOD individual differences in the association between coping and mood. Second, there is relatively little evidence for such differences with situation redefinition and religion. Third, generally speaking, there is more support for individual differences in NA than in PA. Fourth, the interactions accounted for about 1% to 2% additional variance, which is not a huge amount. Fifth, none of the main effects reported above were affected in a significant manner by the inclusion of the interaction terms. Simultaneous Relationships Controlling for Previous Day's Mood There is a possibility that the relationships observed between coping and mood are affected by the autocorrelation that is extremely likely to be present for moods. To test this hypothesis, we recomputed the pooled regressions by entering into the equation a variable representing the previous day's mood (separately for NA and PA). The sample sizes for these analyses are reduced relative to the previous analyses because the creation of a lagged variable eliminates cases when the lag cannot be computed (e.g., the first day for each participant, when there is a missing day). As expected, there was a strong autocorrelation in mood: target NA,f(l, 5671) = 326.61,p < .0001; target PA, F(l, 5671) = 48.45, p < .0001; observer NA, F(l, 5306) = 39.17,p < .0001; and observer PA, F(l, 5298) = 410.83, p < .0001. Generally speaking, these analyses replicated the stronger results of the previous simultaneous regressions. All of the effects in that regression that were significant at the .01 level or were more significant remained highly significant in these analyses. In contrast, all of the findings in which the significance was only at the .05 level were reduced to marginal significance levels (p < .10) in these analyses. In two cases, results that were previously marginally significant became significant at the .05 level (signs of the effects remained the same): observer NA with distraction and observer NA with relaxation. Discussion The purpose of this study was to determine if men's reports about coping with daily problems were related to same-day mood. A secondary aim was to explore the relationship between appraisals of daily problems, coping, and daily mood. To simplify analyses, we limited our exploration of appraisals to undesirability. The undesirability of the daily problem was directly related to NA and, somewhat less strongly, inversely related to PA, regardless of the coping efforts reported. This pattern confirmed our concern about reporting coping efficacy outcomes without taking this appraisal factor into account. It is also of interest that reported undesirability of the problem was related to end-of-day mood reports at all, considering that mood is presumably determined by multiple factors including, but not limited to, the experience of other desirable and undesirable daily events and health. Certainly, the problem that was reported in the coping section of the questionnaire is part and parcel of the experience of daily events, but is only one aspect of that experience. Several coping modes were significantly associated with level 347 of problem undesirability. Less frequent use of situation redefinition was linked to more undesirable problems, whereas increases in use of catharsis, seeking social supports, direct action, and relaxation were associated with more undesirable problems. This pattern of coping response is not as simple as that which might have been expected—for example, that more severe problems would predict increased coping for all eight coping modes. A factor that may complicate the interpretation of these relationships, and one that is not considered in the analyses presented, is that of problem content. As we have already shown with data from this article (Schwartz & Stone, 1993), as problem undesirability increases, people alter their coping differently depending on whether the problem is work related, spouse related, or injury or illness related. However, such differences should not affect the coping-mood analyses that are the focus of this report. As in a previous report on a small sample of this data set, there are significant associations between some of the coping modes; we briefly discuss these relationships, inasmuch as they are similar to those previously reported (Stone & Neale, 1984). The mean coefficient with the greatest magnitude was that between acceptance and direct action (inverse), with an average of —.31. Although substantial, this is not a very large overall average correlation and supports keeping the coping scales separate, rather than combining them. Generally, the directionality of the correlations is consistent with expectations. When direct action is used, it is less likely that distraction, situation redefinition, acceptance, and relaxation are reported. On the other hand, use of relaxation is positively related to use of distraction and religion, all of which are emotion-focused types of coping. Also notable in these analyses are the relationships between religion and seeking social support (direct), relaxation and direct action (inverse), direct action and acceptance (inverse), and direct action and distraction (inverse). Although none of these associations is particularly novel, they do support the face validity of the daily coping assessment. It is somewhat surprising that use of catharsis was inversely related to distraction, situation redefinition, direct action, acceptance, and relaxation and was unrelated to the remaining coping modes. This suggests that catharsis is not used with other coping modes and that it does not fit neatly into the traditional emotion- or problem-focused dichotomy of coping responses. Finally, the relationships observed in this study were often consistent with those observed in the previous study, although there are some exceptions. But inasmuch as this study examined only men and the earlier analyses were of both men and women, differences are not entirely unexpected, both because sex differences have been previously observed in coping (Porter & Stone, 1993) and because this study has considerably greater statistical power than the previous one. Once problem undesirability and individual differences were controlled, the analyses indicated that individual coping efforts were differentially related to mood and that these relationships were not the same for NA and PA. We focus on the strongest findings—those in which both target-reported and observer-reported affect were related to coping efforts—and we discuss the findings for positive and negative mood scales separately. These findings were also observed when the autocorrelation among day-to-day moods was taken 348 A. STONE, E. KENNEDY-MOORE, AND J. NEALE into account. Catharsis and seeking social supports were associated with more NA, whereas acceptance was associated with less NA. The different directions of the associations are important, because all of the coping modes might have been negatively associated with NA, perhaps suggestive of a reporting bias. Furthermore, it is notable that some coping efforts were associated with increased NA; for PA, distraction, relaxation, and acceptance were associated with more PA. Again, all of the coping efforts might have been directly associated with PA, yet only some actually were. Comparing results across PA and NA shows that acceptance was inversely related to NA and directly related to PA; a similar, although less robust, pattern was observed for distraction. Some of our findings regarding the association between coping and mood are consistent with experimental, laboratory studies. For example, Pennebaker and Beall (1986) found that writing about a past traumatic event, presumably a cathartic experience, produced increases in negative mood. Many experiments have shown that relaxation reduces negative mood (Benson, 1975) and that distraction can have a similar effect (e.g., Josephs & Steele, 1990). Our data indicate that seeking social supports is associated with more NA, a finding inconsistent with the generally held view of the salutary effect of social supports (S. Cohen & Wills, 1985). However, other research is consistent with ours in suggesting that support may be associated with worse mood (Coyne & Downey, 1991). We were also surprised that direct action, often regarded as a highly effective means of coping, and situation redefinition had weak relationships to mood (for direct action only a weak univariate relationship emerged that was eliminated in the simultaneous regression). Likewise, others have found religiosity to be an effective means of coping (e.g., Park, Cohen, & Herb, in press), yet our religion scale showed no association with mood. Interpreting these results involves speculating about the nature of the coping process and the definition of coping outcome measures, and it leads to questions about the success of our efforts to understand coping. First, why were some coping modes associated with seemingly negative outcomes, such as more NA? Part of the answer may lie in the use of immediate (same-day) mood as an outcome variable. Coping with problems is best conceptualized as an ongoing process, sometimes spanning several days; exploring daily cross sections of the process may lead to counterintuitive results (Stone & Kennedy-Moore, 1991; Stone, Kennedy-Moore, Newman, Greenberg, & Neale, 1992). Coping immediately with a problem may not have instantaneous beneficial effects, but rather may increase distress before leading, perhaps, to a later resolution of the problem. For example, seeking social support with a therapist could lead to immediate NA as the problem was hashed out, yet despite this apparently negative outcome, it could in the long run also be the most effective manner of dealing with the situation. Distraction or relaxation may result in decreased NA and increased PA in the short run, but it is not clear what effects putting off dealing with the problem could have in the long run. Similarly, in Pennebaker and Beall's (1986) study, although catharsis produced immediate increases in NA, the long-term result was favorable—fewer visits to the student health center. A second question concerns the possibility that participants' moods might influence their reports of coping, and vice versa. Although we believe it is possible, we have four alternative thoughts about this issue. First, mood was the first task in the questionnaire and was not adjacent to the coping assessment. It is therefore unlikely that the reporting of the daily problems (which might elicit affect concerning the problem) influenced mood reports. The opposite causal pattern—mood induced by answering the mood adjectives influencing the coping reports— seems much less likely because the coping assessment (except catharsis) is not affect oriented. Second, the analyses of the observers' reports of the targets' moods could not be affected by the targets' coping reports as this information was not available to observers. Observer-reported mood had the same relationship to coping as did target-rated mood, suggesting that simple reporting biases operating within a single individual (the target) cannot explain the results. Third, it should be noted that although many of the associations between coping and mood were statistically significant, they accounted only for a small percentage of mood's variance. Although mood is a more proximal indicator of daily coping, it is also caused by many variables including other events of the day and physiological factors. These other influences were not part of the modeling procedure and thus appear as error variance in these analyses, reducing the variance accounted for by coping. Fourth, we cannot make a strong claim that differential coping is causing the observed increases or decreases in PA and NA. It is possible, for example, that mood changes cause different types of coping to be adopted. To more fully address the causality issue would require either a fine-grained analysis of life experiences, mood, and coping over the day (e.g., using ecological momentary assessment [EMA; Stone & Shiftman, 1994], in which individuals are queried many times throughout the day about their immediate experiences) or an analysis of lagged effects, such as predicting mood on Day 2 from coping on Day 1. Unfortunately, EMA data on coping are not yet available and, in this instance, the lagged analysis would not be informative because even severe daily events do not generally yield lagged associations with mood (Stone, Neale, & Shiffman, 1993). Our coping assessment—and every other coping assessment we know of—misses the ongoing, transactional nature of the process by either condensing it over many days or by not viewing days as sequences in which occurrences on one day influence processes on subsequent days. Clearly, a consideration of time course is essential for proper understanding of the coping process and its effects. This is not to say that there are obvious solutions to the problem, although future studies might consider following particular problems from their onset through their resolution. References Agren, B., Ryden, O., Johnsson, P., & Nilsson-Ehle, P. (1993). Rehabilitation after coronary bypass surgery: Coping strategies ASSOCIATION BETWEEN DAILY COPING AND END-OF-DAY MOOD predict metabolic improvement and return to work. Scandinavian Journal of Rehabilitation Medicine, 25, 83-95. Benson, H. (1975). The relaxation response. New York: Morrow. Blakely, A. A., Howard, R. C, Sosich, R. M., Murdoch, J. C., Menkes, D. B., & Spears, G. F. (1991). Psychiatric symptoms, personality and ways of coping in chronic fatigue syndrome. Psychological Medicine, 21, 347-362. Bolger, N., DeLongis, A., Kessler, R. C., & Schilling, E. A. (1989). Effects of daily stress on negative mood. Journal of Personality and Social Psychology, 57, 808-818. Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed). Hillsdale, NJ: Erlbaum. Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98, 310-357. Coyne, J., & Downey, G. (1991). Social factors and psychopathology: Stress, social support, and coping processes. Annual Review of Psychology, 42, 401^*25. Deenen, T. A., & Klip, E. C. (1993). Coping with asthma. Respiratory Medicine, 87, 67-70. Denollet, J., & de Potter, B. (1992). Coping subtypes for men with coronary heart disease: Relationship to well-being, stress and Type- A behaviour. Psychological Medicine, 22, 667-684. Diener, E., & Emmons, R. A. (1984). The independence of positive and negative affect. Journal of Personality and Social Psychology, 47, 1105-1117. Elliot, G. R., & Eisdorfer, C. (1982). Stress and human health: Analysis and implications of research. New York: Springer. Ettinger, R. S., & Heiney, S. P. (1993). Cancer in adolescents and young adults. Psychosocial concerns, coping strategies, and interventions. Cancer, 71, 3276-3280. Folkman, S., & Lazarus, R. S. (1980). An analysis of coping in a middle-aged community sample. Journal of Health and Social Behavior, 21, 219-239. Herbert, T. B., & Cohen, S. (in press). Measurement issues in research on psychosocial stress. In H. B. Kaplan (Ed.), Perspectives on psychosocial stress. New York: Academic Press. Jaspers, J. P., Heuvel, F., Stegenga, B., & de Bont, L. G. (1993). Strategies for coping with pain and psychological distress associated with temporomandibular joint osteoarthrosis and internal derangement. Clinical Journal of Pain, 9, 94-103. Josephs, R. A., & Steele, C. M. (1990). The two faces of alcohol myopia: Attentional mediation of psychological stress. Journal of Abnormal Psychology, 99, 115-126. Keller, M. L., Jadack, R. A., & Mims, L. F. (1991). Perceived stressors and coping responses in persons with recurrent genital herpes. Research in Nursing and Health, 14, 421-430. Kinash, R. G., Fischer, D. G., Lukie, B. E., & Carr, T. L. (1993). Coping patterns and related characteristics in patients with IBD. Rehabilitation Nursing, 18, 12-19. Kvam, S. H., & Lyons, J. S. (1991). Assessment of coping strategies, social support, and general health status in individuals with diabetes mellitus. Psychological Reports, 68, 623-632. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal and coping. New York: Springer. Leserman, J., Perkins, D. O., & Evans, D. L. (1992). Coping with the threat of AIDS: The role of social support. American Journal of Psychiatry, 149, 1514-1520. Manne, S. L., & Zautra, A. J. (1990). Couples coping with chronic illness: Women with rheumatoid arthritis and their healthy husbands. Journal of Behavioral Medicine, 13, 327-342. Michela, J. L. (1990). Within-person correlational design and analysis. In C. Hendrick & M. S. Clark (Eds.), Research methods in personality and social psychology (pp. 279-311). Newbury Park, CA: Sage. 349 Nowlis, V. (1965). Research with the Mood Adjective Checklist. In S. S. Tomkins & C. E. Izard (Eds.), Affect, cognition, and personality (pp. 352-389). New York: Springer. Park, C., Cohen, L. H., & Herb, L. (in press). Intrinsic religiousness and religious coping as life stress moderators for Catholics versus Protestants. Journal of Personality and Social Psychology. Pennebaker, J., & Beall, S. (1986). Confronting a traumatic event: Toward an understanding of inhibition and disease. Journal of Abnormal Psychology, 95, 274-281. Pirie, M., & Halliday Smith, L. (1992). Coping with PMS. A women's health center has success with a lifelike skills model. Canadian Nurse, 88, 24-25. Porter, L. S., & Stone, A. A. (1993). Sex differences in coping: Debunking the myth. Manuscript submitted for publication. Ptacek, J. T., Smith, R. E., Espe, K., & Rafferty, B. (1994). Limited correspondence between daily coping reports and retrospective coping recall. Psychological Assessment, 6, 41-49. Repetti, R. (1989). Effects of daily workload on subsequent behavior during marital interaction: The roles of social withdrawal and spouse support. Journal of Personality and Social Psychology, 57, 651-659. Schwartz, J. E., & Stone, A. A. (1993). Coping with daily work problems: Contributions of problem content, appraisals, and person factors. Work and Stress, 7, 47-62. Snyder, M. (1990). Stressors, coping mechanisms, and perceived health in persons with epilepsy. International Disability Studies, 12, 100-103. Stone, A. A. (1981). The association between perception of daily experiences and self- and spouse-related mood. Journal of Research in Personality, 15, 510-522. Stone, A. A., Greenberg, M. A., Kennedy-Moore, E., & Newman, M. G. (1991). Self-report, situation-specific coping questionnaires: What are they measuring? Journal of Personality and Social Psychology, 61, 648-658. Stone, A. A., & Kennedy-Moore, E. (1991). Assessing situational coping: Conceptual and methodological considerations. In H. Friedman (Ed.), Hostility, coping, and health (pp. 203-214). Washington, DC: American Psychological Association. Stone, A. A., Kennedy-Moore, E., Newman, M. G., Greenberg, M., & Neale, J. M. (1992). Conceptual and methodological issues in current coping assessments. In B. Carpenter (Ed.), Personal coping: Theory, research, and application (pp. 15-29). New York: Praeger. Stone, A. A., & Neale, J. (1984). A new measure of daily coping: Development and preliminary results. Journal of Personality and Social Psychology, 46, 892-906. Stone, A. A., Neale, J. M., & Shiftman, S. (1993). How mood relates to stress and coping: A daily perspective. Annals of Behavioral Medicine, 15, 8-16. Stone, A. A., & Shiftman, S. (1992). Reflections on the intensive measurement of stress, coping, and mood, with an emphasis on daily measures. Psychology and Health, 7, 115-129. Stone, A. A., & Shiftman, S. (1994). Ecological Momentary Assessment (EMA) in behavioral medicine. Annals of Behavioral Medicine, 16, 199-202. Ursin, H., & Olff, M. (1993). Psychobiology of coping and defence strategies. Newopsychobiology, 28, 66-71. Voepel-Lewis, T., Starr, A., Ketefian, S., & White, M. J. (1990). Stress, coping and quality of life in family members of kidney transplant recipients. Anna Journal, 17, 427-431. Wikblad, K. F., & Montin, K. R. (1992). Coping with a chronic disease. Diabetes Educator, 18, 316-320.