Coping As a Mediator of Emotion

Coping as a Mediator of Emotion

Susan Folkman and Richard S. Lazarus University of California, Berkeley

There is widespread conviction among health care professionals that coping affects emotion. Yet theory and research have traditionally emphasized the effects of emotion on coping. The present research addresses this imbalance bv evaluating the extent to which coping mediated emotions during stressful encounters in two Caucasian, community-residing samples. Subjects' recently experienced stressful encounters, the ways they coped with the demands of those encounters, and the emotions they experienced during two stages of those encounters were assessed repeatedly. The extent to which eight forms of coping mediated each of four sets of emotions was evaluated with a series of hierarchical regression analyses (of residuals). Coping was associated with changes in all four sets of emotions, w ith some forms of coping associated with increases in positive emotions and other forms associated with increases in negative emotions.

There is a long-standing and widely held cons'iction among Child & Waterhouse, 1952; Sarason, Mandler, & Craighill, researchers and practitioners in the fields of mental health and 1952). behavioral medicine that the ways people cope with the de- The understanding of the relationship between emotion and mands of a stressful event make a difference in how they feel coping that derives from these two theoretical models is incom- emotionally. Yet despite this conviction, there is little under- plete for two important reasons. First, the complexity of emo- standing about the ways coping processes actually affect the tion and coping processes is underestimated. Emotion, for ex- emotion response. ample, tends to be treated as unidimensional drive or arousal. Historically, coping has been viewed primarily as a response However, emotions depend on cognitive appraisals of the sig- to emotion. Within the animal model of stress, for example, nificance of the person-environment relationship for the indi- coping is defined as learned behaviors that contribute to sur- vidual's well-being and the available options for coping (cf. Laz- vival in the face of life-threatening dangers (e.g.. Miller. 1980: arus, 1982; Lazarus, Averill, & Opton. 1970: Lazarus & Folk- Ursin. 1980). These behaviors are initiated by fear, which moti- man, 1984; Lazarus, Kanner. & Folkman, 1980). Any one vates the behavioral response of avoidance or escape, and by stressful event, even an ordinary daily encounter, usually has anger, which motivates confrontation or attack. Within the ego- more than one implication for well-being and more than one psychology model, coping includes cognitive processes, such as option for coping (e.g., Folkman, Lazarus, Dunkel-Schetter, denial, repression, suppression, and intellectualization, as well DeLongis, & Gruen, 1986). As a result, people are likely to ex- as problem-solving behaviors, that are invoked to reduce or perience multiple and often conflicting emotions, as has been manage anxiety and other distressing emotion states (e.g.. Men- seen in young children who feel both happy and sad about ninger. 1963; Vaillant. 1977). something that has transpired (Harris, 1985; Terwogt, Schene, Much of the research on the relationship between emotion & Harris. 1985) and in students preparing for exams who feel and coping in humans has focused on the ways in which emo- both threat and challenge emotions (Folkman & Lazarus, tion—in the form of anxiety—can interfere with cognitive 1985). And in these models coping tends to be treated as either functioning and hence coping (e.g., rCrohne & Laux. 1982: approach-avoidance behavior or defensive processes. However, Schwarzer, 1984: Spielberger, 1966, 1972; van der Ploeg, people use not just approach-avoidance behavior or defensive Schwarzer. & Spielberger. 1984). Two mechanisms of interfer- processes to cope with the complex demands and constraints of ence have been emphasized, a motivational one in which atten- a given stressful encounter, but a wide range of cognitive and tion is redirected from a task at hand to a more pressing emer- behavioral strategies that have both problem-solving and gency (e.g., Easterbrook, l959;Schonpflug, 1983), and a cogni- emotion-regulating functions (e.g., Felton, Revenson, & Hin- tive one in which anxiety-related thoughts that are irrelevant to richsen, 1984; Folkman & Lazarus, 1980, 1985; Folkman, Laz- performance impede functioning (e.g., Alpert & Haber, 1960; arus, Dunkel-Schetter, DeLongis, & Gruen, 1986; Folkman, Lazarus, Pimley. & Novacek, 1987; McCrae, 1982, 1984; Vital- iano, Russo, Carr, Maiuro, & Becker, 1985). Second, these theoretical models emphasize a unidirectional This research was supported by grants from the MacArthur Founda- causal pattern in which emotion affects coping both by motivat- tion and the National Institute on Drug Abuse (DA02976). ing it and impeding it. However, the relationship between emo- Correspondence concerning this article should be addressed to Susan Folkman. who is now at the University of CaliforniaSchool of Medicine, tion and coping in stressful encounters is bidirectional, with Division of General Internal Medicine, A405. 400 Parnassus Avenue, each affecting the other. The behavioral flow begins with a trans- San Francisco. California 94143-0320. action that is appraised as harmful, beneficial, threatening, or

466 COPING AS A MEDIATOR OF EMOTION 467

Person-Environment Encounter based on data from two samples, one in which the subjects were community-residing married couples with young children living at home and a second in which the subjects were retired, Appraisal community-residing adults over the age of 65. Although the

Priaary Secondary samples are demographically similar, some procedures and measures were modified for use with the older sample to ad- I > EMOTION <—I dress questions concerning aging that are dealt with elsewhere Quality and Intensity (cf. Folkman, Bernstein, & Lazarus, 1987; Huffine, Folkman, & Lazarus, 1987). Despite these differences, the overlap in the I two studies was sufficient to allow us to draw upon both for the COPING present analysis.

Problem-focused Emotion-focused Three major questions were addressed: (a) Does coping in 1 fact mediate the emotion response in stressful encounters? (b) Altered Person-Environment Changge in AttentioAttention or If so, are the effects limited to certain types of emotions, or are Relationship Meaning they evident across all emotions, including positive and negative ones? (c) To what extent is there specificity in the association between diverse forms of emotion-focused coping and problem- Reappraisal focused coping and changes in the emotion response? For exam- EMOTION ple, are all forms of coping associated with the reduction of dis- Quality and Intensity tress, or are some forms associated with increases in distress?

Method Hew Person-Environment Encounter Samples Figure I. Coping as a mediator of emotion. The people selected for the two studies were primarily Protestant or Catholic, had at least an eighth-grade education, had an above-marginal family income ($18,000 for a family of four in 1980 in the younger sample, and $10,000 for retired individuals or couples in 1982 in the challenging. The appraisal process generates emotion. The ap- older sample), and were not bedridden. Qualified subjects were identi- praisal and its attendant emotions influence coping processes, fied through random-digit dialing. Prospective subjects received a letter which in turn change the person-environment relationship. explaining the study and then a telephone call from a project interviewer who answered questions and requested a home interview. The altered person-environment relationship is reappraised, Younger sample. The younger sample consisted of 85 married cou- and the reappraisal leads to a change in emotion quality and ples living in Contra Costa County with at least one child at home. The intensity. Viewed in this way, coping is a mediator of the emo- female portion of the sample was restricted to women between the ages tion response. The process is summarized in Figure 1. of 35 and 45; their husbands, whose ages were not a criterion for eligibil- Mediator variables are often confused with moderator vari- ity, were between the ages of 26 and 54 and were also studied. Forty-six ables. Moderators are antecedent conditions such as gender, so- percent of the qualified couples who received letters agreed to be in the cioeconomic status, or personality traits that interact with other study. The mean age of the women was 39.6 years, and the mean age of conditions in producing an outcome. An example is a goal hier- the men was 41.4 years. The mean number of years of education was archy that a person brings to an encounter. This hierarchy inter- 15.5, and the median family income was $45,000. Eighty-four percent acts with relevant environmental variables to produce an ap- of the men and 57% of the women were employed for pay. People who refused to be in the study were compared on age, religion, and educa- praisal and its attendant emotion response. A mediating vari- tion; they differed significantly from those who participated in the study able, on the other hand, is generated in the encounter, and it only in years of education (M = 14.3). Ten couples dropped out of the changes the relationship between the antecedent and the out- study, an attrition rate of 11.8%. These couples were excluded from the come variable. As a mediator, coping arises during the en- analysis, yielding a final sample of 75 couples, or 150 men and women. counter and transforms the original appraisal and its attendant Subjects who dropped out did not differ from those who remained in emotion in some way. The difference between moderator and the study with respect to age, religion, or education (information on mediator variables is conceptually and methodologically im- income was not available for those who dropped out). Interviews were portant and often misunderstood (cf. Baron & Kenny, 1986; conducted in two 6-month waves from September 1981 through August Frese, 1986; Stone, 1985;Zedeck, 1971). 1982. Older sample. The older sample consisted of 161 people living in We examined the extent to which eight different forms of the San Francisco East Bay Area. Forty-four percent of the people who coping mediated four types of emotions in stressful events of received letters agreed to be in the study. The mean age of the women day-to-day living. The research was part of a larger program- was 68.9 years, and the mean age of the men was 68.3 years. The mean matic series of studies of stress and coping processes (cf. number of years of education was 14.7, and the median family income DeLongis, Folkman, & Lazarus, in press; Folkman & Lazarus, was $22,500. People who refused to be in the study were compared on 1980, 1985, 1986; Folkman, Lazarus, Dunkel-Schetter, age, gender, religion, and education. The only significant difference was DeLongis, & Gruen, 1986; Folkman, Lazarus, Gruen, & in education: Those who refused to be in the study had a mean of 13.2 DeLongis, 1986; Folkman et al, 1987). The present report is years. Twenty people dropped out of the study, an attrition rate of 12%, 468 SUSAN FOLKMAN AND RICHARD S. LAZARUS leaving a final sample of 141 people. Forty-seven percent of the final times be used to regulate emotion, as when a person decides that the best sample were male, and 53^ were female. Sixty-one percent were mar- way to reduce anxiety is to tackle the task that is causing the anxiety; and ried, I CTr divorced. 26°c widowed, and 3°t< never married. Those who an emotion-focused strategy, such as taking a tranquilizer, can have a dropped out of the study did not differ from those who remained in the problem-focused function when it is used to reduce anxiety that is im- study in age. gender, religion, or education (information on income was peding task-related work. Thus, our conceptual preference is to assume not available for those who dropped out). Interviews were conducted in that any act or thought can have more than one coping function depend- two 6-month waves from September 1983 through August 1984. ing on the psychological context in which it occurs (cf. Lazarus & Folk- man. 1984). The 50 items comprising these scales were used to assess coping in Procedures the younger sample. As noted above, interviews with the older subjects were shorter than those with the younger subjects, which meant that we Subjects were interviewed in their homes once a month for 6 months. had less time for the assessment of coping. Therefore, a shortened, 31- In the younger sample, husbands and wives were interviewed separately item version was used to assess coping in the older sample. The items by different interviewers on the same day and. if possible, at the same for Ihe short version were selected on the basis of their factor loadings lime. In the older sample each subject was interviewed alone. In both from the full-length Ways of Coping Questionnaire that was used with studies each subject had the same interviewer throughout. Interviews the younger sample, with the exception of the Positive Reappraisal scale. lasted about 1V: to 2 hr for the younger subjects and 1 to 1Vi hr for The items for the short version of the Positive Reappraisal scale empha- the older subjects. We limited the length of the interviews for the older sized the use of religion as a coping mechanism, which we thought was subjects because we did not want to tire them. The data reported here important to investigate in the older sample. The items on the short were gathered during the second through sixth interviews in the younger version were "found new faith": "rediscovered what is important in sample and in the second through fifth interviews in the older sample. life": and "I prayed." The long version of this scale also includes items such as "changed or grew as a person in a good way" and "I came out Measures of the experience better than when I went in" that suggest a less spiritual form of reappraisal. The alphas for the eight scales based on the short- A structured protocol was used to reconstruct a recently experienced ened version ranged from .47 to .74.' stressful or emotional encounter. In the younger sample subjects were In both samples, subjects rated on a 4-point Likert scale the extent to asked to report the most stressful event they had experienced during the which they used each strategy during the stressful encounter. The Ways week preceding the interview. The directions were modified in the older of Coping Questionnaire is usually self-administered during an inter- sample to include the month preceding the interview because the older view in which a specific stressful encounter is reconstructed. This was subjects sometimes did not have any stressful event to report from the the procedure used with the younger subjects. The procedure was modi- preceding week. The protocol was administered monthly five times with fied for the older subjects. Following the reconstruction of the stressful the younger sample and four times with the older sample. The present encounter, a sort board was placed in front of the subject and he or she study drew only on the protocol questions about the subject's emotions was handed a pack of 3 X 5 inch cards, each of which contained one and coping during the encounter. item and its corresponding number from the Ways of Coping Question- Coping. Coping was assessed with the revised Ways of Coping Ques- naire. The subject put the card in the slot that was labeled with the tionnaire, which describes a broad range of cognitive and behavioral appropriate point on the Likert scale. After the task was completed, the strategies thai people use to manage internal and/or external demands subject read the numbers on the cards in each slot to the interviewer, in a stressful encounter. A factor analysis, which w« described in a prior who recorded the responses on the questionnaire. This procedure was report (Folkman, Lazarus, Dunkel-Schetter, DeLongis, & Gruen, used on the premise that it would place fewer demands on the older 1986). suggested eight coping scales: (a) Confrontive Coping (e.g., subjects. Scales were scored by summing ratings. "stood my ground and fought for what I wanted"; "tried to get the per- Emotion. Emotion was assessed by asking subjects to indicate the son responsible to change his or her mind"); (b) Distancing (e.g., "went extent to which they experienced each of a number ofemotions, which on as if nothing had happened"; "didn't let it get to me—refused to were selected from the literature on emotion and on the basis of pilot think about it too much"); (c) Self-Control (e.g.. "I tried to keep my interviews. A similar list ofemotions has been used in much of our feelings to myself": "kept others from knowing how bad things were"); research. The younger subjects were asked to recall their emotions at (d) Seeking Social Support (e.g.. "talked to someone who could do the beginning of the encounter, during the encounter, and, if the en- something concrete about the problem"; "accepted sympathy and un- counter was concluded, at the end. The older subjects were asked to derstanding from someone"); (e) Accepting Responsibility (e.g.. "criti- recall their emotions w'hen the encounter was most stressful and, if the cized or lectured myself"; "realized I brought the problem on myself"); encounter was concluded, at its end. (In both studies, encounters that (f) Escape-Avoidance (e.g., "wished that the situation would go away or subjects said were still going on were not included in the analysis.) A somehow be over with": "tried to make myself feel better by eating, factor analysis of the emotion data from the younger sample suggested drinking, smoking, using drugs or medications, etc."): (g) Planful four scales: Worried/Fearful (3 items, a = .81); Disgusted/Angry (4 Problem-Solving (e.g.. "I knew what had to be done, so I doubled my items, a = .87); Confident (3 items, a = .82); and Pleased/Happy (6 efforts to make things work": "I made a plan of action and followed it"); items, a = .80). (For a full description of the factor analysis, see Folk- and (h) Positive Reappraisal (e.g.. "changed or grew as a person in a man & Lazarus. 1986.) Several items, including "apprehensive." good way": "found new faith"). "hurt," "demeaned," and "enthusiastic," were added to the list for the Two of the scales (Confrontive Coping and Planful Problem-Solving) older sample. A factor analysis of the data from the older sample showed probably have primarily problem-focused functions. Four of the scales (Distancing, Self-Comrol, Accepting Responsibility, and Positive Reap- praisal) probably have primarily emotion-focused functions, and one 1 The coping scales that were used with the younger sample were re- scale (Seeking Social Support, which assesses informational and emo- scored according to the short versions in order to compare the short tional support) can undoubtedly serve both functions. It is important and long versions. The correlations between the short and long versions to note, however, that our designation of a scale as problem-focused or ranged from .88 to .98. The alphas for the shortened scales ranged from emotion-focused is provisional. Problem-focused coping can some- .53 to .76. COPING AS A MEDIATOR OF EMOTION 469 complete convergence with the data from the younger sample for the associated with the Disgusted/Angry scale and negatively asso- items that were common to both samples. The additional items in the ciated with the Confident scale and the Pleased/Happy scale. older sample loaded on the existing factors and did not change their Accepting responsibility was associated only with the Confident meaning. The four scales in the older sample were: Worried/Fearful (5 scale (in a negative direction). Self-control, seeking social sup- items, a = .74); Disgusted/Angry (5 items, a = .78); Confident (4 items, port, and escape-avoidance did not contribute significantly to a = .66); and Pleased/Happy (5 items, a = .83). The content of these scales is theoretically consistent with four kinds of cognitive appraisal the variance in any of the emotion scales. (Folkman & Lazarus, 1985): threat (Worried/Fearful); harm (Dis- gusted/Angry); challenge (Confident); and benefit (Pleased/Happy). Older Group

The intercorrelations among the residual scores of the coping Results scales for the older group are shown in Table 3. The four regres- The extent to which coping mediated emotion during stress- sion analyses for the older group are shown in Table 4. ful encounters was evaluated with hierarchical regression analy- After partialing out within-person means and controlling for ses of residual scores. Because multiple encounters were re- emotions at the height of the stressful encounters, coping con- ported by subjects, residual scores were used to control for de- tributed significantly to the variance in all four emotion scales. pendency in the data and for individual difference variables, Planful problem-solving contributed significantly to the vari- including gender, that were extraneous to the question ad- ance in three emotion scales: The association was positive for dressed here. This procedure allowed us to treat each encounter the Confident scale and negative for the Disgusted/Angry and as a separate observation. To residualize the regression analyses, Pleased/Happy scales. Seeking social support was positively as- each person's within-person mean on each of the variables in a sociated with the Confident scale and negatively associated with given analysis was partialed out by entering it on the first step the Pleased/Happy scale. Accepting responsibility and positive of the regression analysis. reappraisal were associated (in a positive direction) only with For each of the two samples there were four regression analy- the Worried/Fearful scale; escape-avoidance was associated (in ses. The dependent variable in each analysis was one of the four a positive direction) only with the Confident scale; and distanc- concluding emotion scales. In the younger sample, the first step ing was associated (in a negative direction) only with the of each regression consisted of within-person means on the Pleased/Happy scale. Self-control and positive reappraisal did emotions at the beginning of the encounter, the eight coping not contribute significantly to the variance in any of the four scales, and the emotions at the conclusion of the encounter; emotion scales. scores on the emotion scale at the beginning of the encounter were entered on the second step; and scores on each of the eight Alternative Explanations coping scales were entered on the third step. The same proce- The extent to which the same group of coping strategies dure was followed for the older sample, the only difference being (planful problem-solving, distancing, confrontive coping, and that emotions at the height of the encounter were entered in- positive reappraisal in the younger group and planful problem- stead of emotions at the beginning of the encounter. The data solving in the older group) seemed to account for changes in from the younger sample and the older sample were analyzed 2 emotions prompted additional analyses. Individual difference separately because of the differences in assessment procedures. variables were to a large extent controlled by residualizing the regressions, which makes it unlikely that the consistency in the Younger Group findings was due to person factors. However, this consistency could have been the result of lack of independence among the The intercorrelations among the residual scores of the coping emotion scales. scales for the younger group are shown in Table 1. The four The intercorrelations among the emotion scales are shown in regression analyses for the younger group are shown in Ta- Table 5. The correlations of interest include those among the ble!. Disgusted/Angry, Confident, and Pleased/Happy scales. The After partialing out within-person means and controlling for emotions at the beginning of the encounter, coping explained a significant amount of variance in the Disgusted/Angry scale, 2 The data from the younger sample were also analyzed using the the Confident scale, and the Pleased/Happy scale. Coping did short versions of the eight coping scales to evaluate differences due to not make a significant contribution to the variance in the Wor- scoring procedures. Coping contributed significant variance in the same ried/Fearful scale. Four coping scales, Planful Problem-Solving, three regression analyses using the long and short versions. Across the Positive Reappraisal, Distancing, and Confrontive Coping, con- three regression analyses, 14 beta coefficients were significant using the tributed a significant amount of variance to all three concluding long version. Of those 14, 8 were also significant, and in the same direc- emotion scales. Planful problem-solving and positive reap- tion, using the short version. Of the remaining 6 beta coefficients, 5 were in the same direction as those using the long version, and the sixth beta praisal were negatively associated with the Disgusted/Angry coefficient, though in the opposite direction from its long-version coun- scale and positively associated with the Confident scale and the terpart, indicated virtually no contribution (ft - .01). The only system- Pleased/Happy scale. The pattern of association for the Con- atic difference between the two versions pertained to the Positive Reap- frontive Coping and Distancing scales was the opposite of the praisal scale. It was significant in all three regression analyses in its long pattern for the Planful Problem-Solving and Positive Reap- form, whereas in its short form it did not contribute significantly in any praisal scales; confrontive coping and distancing were positively of the analyses. 470 SUSAN FOLKMAN AND RICHARD S. LAZARUS

Table 1 Intercorrelations Among the Coping Scales in ihe Younger Group

Coping scale

1. Confrontive Coping 2. Distancing -.06 3. Self-Control .35 .33 4. Seeking Social Support .26 -.08 .28 5. Accepting Responsibility .25 .23 .26 .06 .22 .35 .22 6. Escape-Avoidance .26 .36 — 7. Planful Problem-Solving .26 .14 .38 .30 .08 .13 — 8. Positive Reappraisal .23 .10 .43 .30 .19 .26 .39 — A/ 7.13 6.02 6.76 9.38 4.29 5.75 9.21 4.17 SD 2.91 2.03 2.39 3.53 1.64 2.05 2.96 1.84 at .70 .61 .70 .76 .66 .72 .68 .79

Note- Number of observations was 444.

Worried/Fearful scale is of less interest than the other three Although the correlations among the scales of interest range scales because coping did not contribute significantly to its vari- from -. 14 to -.57, which suggests that some of the consistency ance in the younger sample, and in the older sample the coping in the coping patterns across emotion scales might be attribut- pattern that was associated with change in the Worried/Fearful able to lack of independence among them, it is important to scale was unlike the pattern associated with the other three note that even in the case of the correlation of greatest magni- scales. tude, only 32°o of the variance is shared.

Table 2 Regression Analysis: Coping as Mediator of Emotions in Younger Group

Emotion scale Adjusted R- R-change

Worried/Fearful (N = 443) Step 1: Within-person means .56 .55 Step 2: Emotions at beginning of encounter .58 .57 .02* .18" Step 3: Coping scales .60 .58 .01

Disgusted/Angry (,V = 444) Step 1: Within-person means .44 .43 Step 2: Emotions at beginning of encounter .49 .48 .04*** .14*" Step 3: Coping scales: .58 .58 .09*** Distancing .17*** Planful Problem-Solving -.20*** Confrontive Coping .27*** Positive Reappraisal -.24*" Confident (N = 444) Step I: Within-person means .58 .57 Step 2: Emotions at beginning of encounter .63 .62 .04*' .24*' Step 3: Coping scales: .68 .66 .05* Accepting Responsibility Confrontive Coping -.08* Distancing .11** Planful Problem-Solving .14*** Positive Reappraisal .19***

Pleased/happ> (N = 442) Step 1: Within-person means .46 .45 Step 2: Emotions at beginning of encounter .52 .51 .06" .24*** Step 3: Coping scales: .61 .59 Confrontive Coping -.22*** Distancing -.11 Planful Problem-Solving .19*** Positive Reappraisal .25***

* Standardized beta coefficient. *p<.05.**p<.01.***/x.OOI. COPING AS A MEDIATOR OF EMOTION 471

Table 3 Jnlercorrelations Among the Coping Scales in ihe Older Group

Coping scale

1. Confrontive Coping 2. Distancing -.06 3. Self-Control .16 .25 4. Seeking Social Support .32 -.16 .12 5. Accepting Responsibility .16 .18 .21 .10 6. Escape-Avoidance .09 .12 .04 — .31 .19 — -.06 .19 .28 .24 .03 7. Planful Problem-Solving .20 — 8. Positive Reappraisal .10 .06 .23 .27 .25 .31 .05 — A/ 2.30 2.87 3.09 3.86 1.15 2.16 5.00 1.78 SD 2.79 2.91 2.70 3.90 1.76 2.30 3.65 2.42 « .64 .68 .51 .72 .49 .47 .74 .63

. A shortened version of the coping scales was used with the older group. Numbers of observations was 299.

Another possibility was that the changes in emotion were due ity. And planful problem-solving and positive reappraisal were to situational characteristics that were common across encoun- used more in encounters with favorable outcomes than they ters. rather than to coping. For example, in previous research were in encounters with unfavorable outcomes. Thus, it is pos- the coping strategies that were used in an encounter and the sible that the observed changes in emotion could have been de- outcome of the encounter were to a certain extent dependent termined by the changeability of the reported encounters, and on whether or not the encounter was appraised as amenable to coping might have had very little independent effect, change (Folkman & Lazarus, 1980:Folkman, Lazarus, Dunkel- We explored this possibility with a series of regression analy- Schetter, DeLongis, & Gruen, 1986). Planful problem-solving, ses in which the effects of changeability were statistically con- confrontive coping, and positive reappraisal were used more fre- trolled. As a part of the reconstruction of each stressful en- quently in encounters that were appraised as low in changeabil- counter we had asked subjects to indicate on a Likert scale the

Table 4 Regression Analysis: Coping as Mediator oj Emotions in Older Group

Emotion scale R2 Adjusted R: R- change (3°

Worried/Fearful (N = 299) Step 1: Within-person means .62 .61 Step 2: Emotions at height of encounter .66 .65 .04*** .27*** Step 3: Coping scales: .69 .67 .03"* Accepting Responsibility .15** Positive Reappraisal .21 Disgusted/ A ngry( N = 299) Step 1 : Within-person means .61 .59 Step 2: Emotions at height of encounter .71 .70 1 1*** 42*** Step 3: Coping scales: .74 .72 .02** Planful Problem-Solving -.18*** Confident (W= 299) Step 1 : Within-person means .65 .64 Step 2: Emotions at height of encounter .67 .66 .02*** .21*** Step 3: Coping scales: .70 .68 .03**

Seeking Social Support 1 1* Planful Problem-Solving J4** Escape-Avoidance 1 1* Pleased/Happy (N= 299) Step 1 : Within-person means .61 .60 Step 2: Emotions at height of encounter .62 .61 .01*** .12*** Step 3: Coping scales: .67 .65 .05** Seeking Social Support .13* Planful Problem-Solving 24** Distancing -.17***

" Standardized beta coefficient. *p<.05.**p<.01.***/><.001. 472 SUSAN FOLKMAN AND RICHARD S. LAZARUS

Table 5 jects, who were asked to reconstruct an event from the previous Inlercorrelaiions Among Emotions Scales week. The only way this possibility can be addressed with assur- ance is to monitor the processes in stages during stressful en- Emotion scale counters, somewhat as we tried todo with a college examination

Vbunger group (Folkman & Lazarus, 1985). However, serious practical diffi- culties tend to militate against such strategies in most research Emotions at beginning of contexts, especially naturalistic ones. encounter: With respect to the role of situational characteristics, the ap- 1. Worried/Fearful — 2. Disgusted/Angry .24 — praised changeability of encounters did not account for the ob- 3. Confident -.42 -.31 — served emotion effects. However, it is possible that other situa- 4. Pleased/Happy -.10 -.41 -.43 — tional characteristics that we were unable to evaluate in the Emotions at conclusion present analysis, such as whether or not the encounter was a of encounter: 1. Worried/Fearful — major life event or a daily hassle, may have been responsible for 2. Disgusted/Angry .38 — the changes in emotion. 3. Confident -.26 -.36 — Keeping in mind the limitations imposed by the retrospective 4. Pleased/Happy -.24 -.57 -.51 — nature of the data and our inability to rule out the effects of

Older group unassessed situational characteristics, the findings of the present study are nevertheless consistent with a mediational inter- Emotions at height of pretation: Coping was associated with changes in all four types encounter: 1. Worried/Fearful — of emotions—disgust and anger, pleasure and happiness, con- 2. Disgusted/Angry -.10 — fidence, and to a lesser extent, worry and fear. Our discussion 3. Confident -.01 -.14 — focuses on two aspects of the findings: age-group differences in 4. Pleased/Happy -.06 -.24 .36 — the patterns of findings and possible mechanisms through Emotions at conclusion which coping affected emotion. of encounter: 1. Worried/Fearful — 2. Disgusted/Angry .17 — Age-Group Differences 3. Confident -.08 _ 72 4. Pleased/Happy -.16 -.50 .45 — There were notable differences in the effects of coping on emotion in the two age groups. For example, positive reappraisal was associated with a decrease in distress (in disgust/ extent to which they could change or do something about the anger) and an increase in positive feeling (in pleasure/happiness encounter they reported. We residualized this variable and en- and confidence) in the younger group and with a worsened emo- tered it in the regression equation before the coping variables tion state (more worry/fear) in the older group. Confrontive were entered. Only the significant coping variables were in- coping was associated with increased distress in the above-men- cluded in this analysis. Changeability did not make a significant tioned emotions in the younger group, but it showed no associa- contribution to any of the scales that assessed emotion at the tion with the emotion state in the older group. And seeking so- conclusion of encounters, nor did it affect the relationships be- cial support was associated with increased positive emotions in tween coping and emotion. The coping scales that were associ- the older group, but it showed no association in the younger ated with significant changes in emotion in the preceding analy- group. sis continued to be significant when we controlled for situa- The age-group differences in the pattern of findings could be tional characteristics related to appraised changeability. due to methodology. In the younger group the first emotions that subjects were asked to report were those for the beginning Discussion of their stressful encounters, whereas in the older group the first emotions that subjects were asked to report were those they ex- The findings in both samples are consistent with the view that perienced when the encounter was at the height of stressfulness. coping mediates emotions in stressful encounters. However, we which could have been at any time during the stressful en- cannot make definitive causal statements because (a) the data counter. Differences in the scale lengths and methods of admin- were retrospective, and self-reports may have been affected by istration could also account for the age-group differences. This memory bias, and (b) the effects of situational characteristics possibility is especially relevant in the case of the Positive Reap- that were not assessed could not be ruled out. praisal scale. The long version of the scale, which was used with To a certain extent, bias due to individual differences in re- the younger sample, contained both nonreligious and religious sponse styles was controlled statistically through residualized strategies for reappraising an encounter. As noted earlier, the regression analysis. However, we must allow for the possibility short version that was used with the older sample emphasized that situational memory bias, independent of response style, in- religious strategies. These methodological inconsistencies fluenced the reports of emotions and coping within each en- should be addressed in future studies. counter. This possibility could have been more likely among the The age-group differences in the associations between coping older subjects, who were asked to reconstruct the most stressful and emotion could also be explained by differences in the types event from the previous month, than among the younger sub- of stressful encounters the subjects reported, which we have de- COPING AS A MEDIATOR OF EMOTION 473 scribed elsewhere (Folkman et al.. 1987). For example, the sub- ing was associated with favorable encounter outcomes (Folk- jects in the younger group reported more encounters having to man, Lazarus. Dunkel-Schetter, DeLongis, & Gruen, 1986). In do with their jobs than did the subjects in the older group, who such cases, the effects on the emotion response are indirect. were retired, and the subjects in the older group reported more Most likely, planful problem-solving can work in both ways in encounters having to do with health than did the subjects in the any given encounter. younger group (Folkman et al., 1987). Further, although both Confrowive coping was consistently associated with worsened groups reported similar proportions of family encounters, the emotion states in the younger group. Perhaps expressing anger family encounters of the younger group tended to involve a and hostility does not always provide the relief that is suggested spouse or child, whereas other family members, such as siblings, in the aphorism "Getting it off your chest makes you feel bet- played a more important role in the family encounters of the ter." In fact, the findings suggest that expressing anger and hos- older group (Folkman etal., 1987). A given coping mechanism's tility may make one feel worse. Support for this idea comes from association with emotion might be influenced by these differ- a previous analysis (Folkman & Lazarus, 1986) in which we ences in context. Confrontive coping, for example, might be found that people who were high in depressive symptoms used more likely to exacerbate the emotion state in encounters with more confrontive coping than did people who were low in de- close family members than in encounters with more peripheral pressive symptoms. Further, we found that this form of coping family members and friends, because encounters with close was associated with unfavorable encounter outcomes (Folk- family members are likely to have more enduring and/or sig- man, Lazarus, Dunkel-Schetter, DeLongis, & Gruen, 1986). nificant consequences. Thus, it is possible that confrontive coping as we assessed it is a Another possibility is that age differences in the associations maladaptive form of problem-focused coping in that it fails not between coping and emotion are due to developmental changes only to improve person-environment relationships but even to in coping efficacy. For example, older people may be more tem- provide relief from distress emotions. perate in their interpersonal confrontations than younger peo- Theoretically, positive reappraisal and distancing should help ple are, meaning that the confrontive coping of older people is reduce distress. Positive reappraisal had the expected effect in likely to have less of a negative effect on the emotion state than the younger group, where it was associated with improved emo- that of the younger people. The possibility that older people are tion states in three analyses. However, in the older group it was more skilled than younger people in interpersonal coping could associated with an increase in distress (worry/fear) rather than also explain why social support seeking was helpful for the older a decrease. And contrary to expectations, in the three analyses people but had no effect for the younger people. The whole ques- where distancing had a significant effect, including two in the tion of age differences in stress and coping processes deserves younger group and one in the older group, it consistently con- close attention (cf. McCrae, 1982). tributed to a worsened emotion state. One explanation for this is that the reduction of distress achieved cognitively through Possible Mechanisms of Coping Effects on Emotion distancing and positive reappraisal may be difficult to sustain. For example, with respect to distancing, unexpected cues from Four types of coping were strongly associated with changes in the environment may redintegrate the significance of an en- emotion: planful problem-solving, positive reappraisal, con- counter, as when a woman's efforts not to think of the pending frontive coping, and distancing. Although the design of this results of a breast biopsy fail because she unexpectedly hears a study precludes definitive causal statements about the relation- report about cancer on her car radio or in passing sees a friend ship between coping and emotion, the findings suggest possible who had a bout with cancer the previous year. Efforts at distanc- mechanisms of affect that are consistent with previous research ing can also be undermined by cues from within the person and clinical observation. in the form of fleeting cognitions or images, or what Horowitz Planful problem-solving contributed significantly to the ex- (1976) calls intrusive thoughts. To the extent that distancing plained variance in the seven regression analyses that showed a cannot be sustained, not only might it fail to diminish distress, significant mediating effect for coping. The pattern of associa- but when it diverts attention from needed problem-solving it tion was consistent: Planful problem-solving was associated could even lead to increases in distress, as when an individual with an improved emotion state, that is, it was associated with uses distancing to avoid thinking about (and therefore seeking less negative emotion and more positive emotion. One explana- medical attention for) a symptom (e.g., Katz, Weiner, Gal- tion for this association is that people can begin to feel better lagher, & Hellman, 1970). We are impressed with how difficult when they turn to the problem that is causing distress. For ex- it is to achieve distancing when one wants to, for example, in ample, persons who experience distress when receiving notice situations in which one must await a decision in which one has of imminent layoff are likely to feel better when they begin to a strong stake. make plans for finding new work. In such cases planful prob- Finally, although it is tempting to argue from the above rea- lem-solving can have a direct effect on emotions even though soning that planful problem-solving is an inherently adaptive the adaptational problem may remain unresolved. Another ex- form of coping and confrontive coping and distancing are inher- planation is that planful problem-solving, when effective, can ently maladaptive, it is important not to lose sight of the princi- result in an improved person-environment relationship, which ple that the adaptive value of a coping process often depends should in turn lead to a more favorable cognitive appraisal and on the context (cf. Folkman & Lazarus, 1980; Lazarus, 1981; hence a more positive emotion response. This explanation is Lazarus & Folkman, 1984; Vaillam, 1977). Thus, planful supported by a previous study in which planful problem-solv- problem-solving can be maladaptive if people persist in it in 474 SUSAN FOLKMAN AND RICHARD S. LAZARUS circumstances where nothing can be done to alter the outcome stress and coping processes in older adults. Psychology and .Aging. 2. (cf. Collins, Baum, & Singer. 1983), and confrontive coping can 366-374. be adaptive when the situation calls for getting another person Folkman, S.. & Lazarus. R. S. (1980). An analysis of coping in a middle- aged community sample. Journal of Health and Social Behavior, 21. to act. We note that the maladaptive appearance of confrontive 219-239. coping in this research may be due to the type of hostile and Folkman, S., & Lazarus. R. S. (1985). Hit changes it must be a process: aggressive strategies that we sampled. Interpersonal strategies Study of emotion and coping during three stages ofa college examina- that have a less aggressive tone should also be evaluated. And tion. Journal of Personality and Social Psychology, 48. 150-170. we should be wary of labeling distancing as inherently maladap- Folkman, S., & Lazarus, R. S. (1986). Stress processes and depressive tive. As noted above, its failure to improve the emotion state in symptomatology. Journal of Abnormal Psychology, 95. 107-113. the present research may have been due to unexpected intru- Folkman, S.. Lazarus. R. S., Dunkel-Schetter, C. DeLongis, A., & sions from the environment that undid the distancing effect or Gruen, R. (1986). The dynamics ofa stressful encounter: Cognitive to its being used in contexts that required action. appraisal, coping, and encounter outcomes. Journal of Personality and Social Psychology. 50. 992-1003. Folkman, S., Lazarus, R. S., Gruen, R., & DeLongis, A. (1986). Ap- Conclusions praisal, coping, health status, and psychological symptoms. Journal of Personality and Social Psychology, 50, 571-579. Three points that we regard as important are made by this Folkman. S., Lazarus, R. S., Pimley, S., & Novacek, J. (1987). Age study. First, coping processes that are generated during the heat differences in stress and coping processes. Psychology and Aging. 2, of a stressful encounter are associated with changes in a wide 171-184. range of ongoing emotions. Theoretical formulations on the re- Frese, M. (1986). Coping as a moderator and mediator between stress lations between emotions and coping that emphasize the effects at work and psychosomatic complaints. In M. Appley & R. Trumbull of emotion on coping without also considering the effects of cop- (Eds.). Dynamics of stress (pp. 183-206). New York: Plenum. ing on emotion are therefore incomplete. Second, both pro- Harris. P. L. (1985). What children know about the situations that problem-focused and emotion-focused forms of coping were asso- voke emotion. In M. Lewis & C. Saarni (Eds.), The socialization of ciated with changes in emotions. Thus, conceptualizations of a/ect(pp. 161-185). New York: Plenum. coping that are limited to unidimensional approach-avoidance Horowitz, M.0976). Stress response syndromes. New York: Jason Ar- behaviors or defensive ego processes provide incomplete under- onson. Huffine, C. Folkman, S., & Lazarus, R. S. (1987). Psychoaclive drugs standing of the complex relationship between coping and emo- and alcohol, and stress and coping processes in older adults. Manu- tion. Third, some forms of coping, such as planful problem- script submitted for publication. solving, may have a salubrious effect on the emotion response, Katz. J. J., Weiner, H., Gallagher, T G., & Hellman, L. (1970). Stress, whereas other forms, such as confrontive coping and distancing, distress, and ego defenses. Archives of General Psychiatry. 23. 131- may make things worse, at least in some populations and in 142. some contexts. The extent to which these effects are context Krohne, H. W., & Laux. L. (Eds.). (1982). Achievement, stress, and anx- specific, developmental in origin, and/or inherent in the coping iety. Washington, DC: Hemisphere. strategy is not clear as yet. A full understanding of these differ- Lazarus, R. S. (1981). The costs and benefits of denial. In B. S. Dohren- ential effects is especially important if we are to make progress wend& B. P. Dohrenvjend (Eds.), Stressful life events and theircon- in developing effective coping-related interventions. le.\ts (pp. 131 -156). (Monographs in PsychosocialEpidemiology. No. 2. B. Z. Locke & A. E. Slaby, Eds.). New York: PRODIST (a division of Neale Watson Academic Publications). References Lazarus, R. S. (1982). Thoughts on the relations between emotion and Alpert, R.. & Haber, R. N. (1960). Anxiety in academic achievement cognition. American Psychologist, 37. 1019-1024. situations. Journal of Abnormal and Social Psychology. 61, 207-215. Lazarus, R. S., Averill. J. R., & Opton, E. M., Jr. (1970). Toward a cogni- Baron. R. M., & Kenny, D. A. (1986). The moderator-mediator vari- tive theory of emotions. In M. Arnold (Ed.), Feelings and emotions able distinction in social psychological research: Conceptual, strate- (pp. 207-232). New York: Academic Press. gic, and statistical considerations. Journal of Personality and Social Lazarus, R. S., & Folkman, S. (1984). Coping and adaptation. In Psychology, 51. 1173-1182. W. D. Gentry (Ed.). The handbook of behavioral medicine (pp. 282- Child, I. L. & Waterhouse, I. K. (1952). Frustration and the quality of 325). New York: Guilford. performance: I. A critique of the Barker, Dembo, & Lewin experi- Lazarus, R. S., Kanner, A. D., & Folkman, S. (1980). Emotions: A cog- ment. Psychological Review. 59. 351-362. nitive-phenomenological analysis. In R. Plutchik & H. Kellerman Collins, D. L., Baum, A., & Singer, J. E. (1983). Coping with chronic (Eds.), Theories of emotion (pp. 189-217). New York: Academic stress at Three Mile Island: Psychological and biochemical evidence. Press. Healih Psychology. 2, 149-166. McCrae, R. R. (1982). Age differences in the use of coping mechanisms. DeLongis, A., Folkman, S., & Lazarus. R. S. (in press). Hassles, health, Journal of Gerontology, 37. 454-460. and mood: A prospective study with repeated daily assessments. Jour- McCrae, R. R. (1984). Situational determinants of coping responses: nal of Personality and Social Psychology. Loss, threat, and challenge. Journal oj Personality and Social Psychol- Easterbrook. J. A. (1959). The effect of emotion on cue utilization and ogy. 46.9\<)-<)2%. the organization of behavior. Psychological Review, 66. 183-201. Menninger, K. (1963). The vital balance: The life process in menial Felton, B. J., Revenson, T. A., & Hinrichsen. G. A. (1984). Coping and health and illness. New York: Viking. adjustment in chronically ill adults. Social Science ana1 Medicine, 18, Miller, N. E. (1980). A perspective on the effects of stress and coping on 889-898. disease and health. In S. Levine & H. Ursin (Eds.), Coping and health Folkman, S., Bernstein, L., & Lazarus, R. S. (1987). Drug misuse and (NATO Conference Series HI: Human factors). New York: Plenum. COPING AS A MEDIATOR OF EMOTION 475

Sarason. S. B., Mandler, G., & Craighill, P. C. (1952). The effect of the meetings of the International Society for Research on Emotion, differential instructions on anxiety and learning. Journal of Abnor- Cambridge, MA. mal and Social Psychology. 47. 561-565. Ursin, H. (1980). Personality, activation and somatic health. In S. Schonpilug, W. (1983). Coping efficiency and situational demands. In Levine & H. Ursin (Eds.), Coping and health (NATO Conference Se- G. R. J. Hockey (Ed.), Stress and faiigite in human performance (pp. ries III: Human factors). New York: Plenum. 299-330). New York: Wiley. Vaillani, G. E. (1977). Adaptation to life. Boston: Little, Brown. Sehwarzer, R. (Ed.). (1984). The sell in anxiety, stress and depression van der Ploeg, H. M., Schwarzer, R., & Spielberger, C D. (1984). Ad- Amsterdam: Elsevier. vances in lest anxiety research (Vol. 3). Hillsdale, NJ: Erlbaum. Spielberger, C. D. (Ed.). (1966). Anxiety and behavior New York: Aca- Vitaliano. P. P., Russo, J.. Carr. J. E.. Maiuro, R. D., & Becker, J. (1985). demic Press. The Ways ofCoping Checklist: Revisionand psychometric properties. Spielberger, C. D. (Ed.). (1972). Anxiety: Current trends in theory and Mtillivariate Behavioral Research, 20, 3-26. research (Vols. 1 & 21. New York: Academic Press. Zedeck, S. (1971). Problem with the use ofmoderator" variables. Psy- Stone, A. A. (1985). Assessment oCcopingefficiency: A comment. Jour- chological Bulletin, 76.295-310. nal of Behavioral Medicine, S. 115-117. Received January 18, 1987 Terwogt, M. M, Schene, J., & Harris, P. L. (1985, June 23-26). Self- Revision received July 20, 1987 control of emotional reactions hy young children. Paper presented at Accepted July 24, 1987 •

STATIMINT OP OWNtMHIf. MANAGIMINT AMD CIRCULATION XMMW*> ItUIC Hill 1* TITtl Of HJ»UCAfto« !• ftJIklCATtON ««O 1. OATIO"ILIN« Journal of Personality and Social Psychol ogy | 2J 8J1 J9 A 0 10/1/87 3. *HIQUIMCV 0* IHUI 1A. »««u»u.MO. O« IUU» U PueulNIO 11.nu AflNUALUMCfllT'On $80-«emberM Monthly " S160-Indv/$320-Inat 4 COM»LITI MAILING AOOMM 0* «NO«M df *ICI Of PwfuCATlOM limit dr.. Cfffrr. *•• «• JJPM Cmi •*•! ^umt/ 1400 N. Uhle Street, Arlington, VA 22201

i COM»klTI MAILING 100*IUO* rMI HIAOOWAffTIRf Of OINIMAL tuWNIM QMICIS OT THl PutLIIMIJI //*•( »Bnnr/ 1200 17th Street, B.W. , Washington, DC 20036

•UIUIHIN ,.%.-. *M CMVMM MMMf AHHmt American Psychological Aasoclatlon, 1200 17th Street, H.W. , Washington, DC 22036 !OiTOMf>wM«M<:*«*M>.MMnf M*W Steven J. Sherman, Dept. ot rsycnology, inaiana university Bloonlngton, IN 47405/Harry T. Eels, Dept. of Psychology, University of Rochester, Rochester, NY 14627/Irwln G. Sarason, Dept. of Psychology, University of Uashlngtot

Susan Knapp, 1400 ». Uhle Street, Arlington, VA 22201

H:S™r:7~^^

*ULLNAM« COM»LtT| UAILIMQ AOOHIU American PsvcholoKical Association Waahinnton. DC 20036

* KNOWN lOHQHOLDiai. MOnrOMIIf. AND OTMIM »ICU«1TY HOLMIU OVWINO 0« K)LD1 W 1 f I*CINT 0» UOMI 0* TOTAL • MOUNT Of IONOI MOMTOAQH ON QTMIR I|CU«ITI II /If »,- M — . » H.KI

'U..LHAXI £OM»Lm UAILIM4 AODNIM Wfinp

• FC" CSMmriON IT nON**0'JH(kn> til 11 3MWM/J

111 \Jt -*I NOT CMANOIO OUHIMQ i— i -*1 CMANOtO OU"INO ilf^f*- »-•«»' -p-r ..-.. »*wn4v »/ 3 '-ICIOiNG H MONTH! |_| »WICIO

. 'OTALHO CO»HI

1 -J*i bBHf*r>«n ».,« ««.•* >^MnMU 3,016 2,516

C TOT*L»AiO««>*.»«|QuHfTIOemCUt*TlO* S*» tf I9tl ** 10*11 S.433 4.533 D '"II OilTdHUTiOtt |V MAIk. CAMUR ON OTMIff MIAN* ^"CLII. COMrUMIMTAHV. .NO OTHIP. *H|| COOif I 279 | 211 1 TOTAL Ol*T»ituTIO»lf*H. •/£*•• Ol 5.712 4.744 ' COfltSNOT DUTOlluTID 1,288 2,695 I. n 1 1 u rn liom KrM * (.rri .. 2 TOTA^fc,./,.^^,.^...^^.^,^^.^^ 7,000 i 7,439

" . V ll-ONATudl AND TlTll OP IOlTO*.Wl.i|N|H. MINIM MANAd" -.".«.- 1 c^ttty thtt tht itatWTMfttt mad* by \J r ) / // / - / m« tbovt ift corrtct >nd comptttt ^ №r~&* tfUt/tlSWb (&f*. &%• />'