Subjective Past, Present, and Future Fit: Effects on ... - Semantic Scholar

3 downloads 4939 Views 1MB Size Report
Motivational poor fit had its greatest effect on strain when poor fit dealt with anticipated or current fit. These findings are discussed in terms of hypotheses, which ...
Copyright 1985 by tlie American Psychological Association, Inc. 0022-3514/85/J00.75

Journal of Penonality and Social Psychology 1985, Vol. 48, No. 1, 180-197

Subjective Past, Present, and Future Fit: Effects on Anxiety, Depression, and Other Indicators of Weil-Being Robert D. Caplan University of Michigan

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Rama C. Tripathi and R. K. Naidu University of Allahabad, Allahabad, India Hypotheses were tested regarding the effects of subjective current, retrospected, and anticipated person-environment fit on affective and somatic strain and wellbeing. Two-hundred seven university students preparing for annual academic examinations at an Indian university completed self-administered questionnaires measuring parameters of the hypotheses. A distinction was made between cognitive (e.g., meeting demands for intelligence, good memory) and motivational (e.g., being able to muster the effort) fit. Cognitive poor fit had a greater effect on predicting high levels of strain across all time frames but did best in the past and current frames. Motivational poor fit had its greatest effect on strain when poor fit dealt with anticipated or current fit. These findings are discussed in terms of hypotheses, which suggest that the more subjectively uncontrollable a dimension of fit, the more salient it is in the past time frame, whereas the more subjectively controllable the dimension, the more salient it is in the future time frame.

H. G. Wells' Time Machine (1968) allowed its inventor to move back and forth through time, free from the contemporary restraints of Victorian England. Although those less fortunate do not have access to such a wonderful machine, it is still the case that our subjective experiences carry us beyond the present to the past and to the anticipated future. These time perspectives are not some special "separate reality" (Castenada, 1971). They are a common, everyday influence on current actions, future plans, and immediate feelings of well-being. This study investigates this proposition. Although some hypotheses have been offered about different time perspectives as

This study was made possible by funding from the U.S. Educational Foundation of India Fulbright Scholarships to all three authors, the University Grants Commission of the Government of India, University of Michigan Sponsored Computing Funds, NIMH Grant #MH34586-02, and funding by Hoffmann-La Roche to R. Caplan. We thank Antonia Abbey, John R. P. French, Jr., Willard L. Rodgers, and two anonymous reviewers for their comments and suggestions regarding an earlier version of the text. Requests for reprints should be sent to Robert Caplan, Institute for Social Research, University of Michigan, 426 Thompson Street, Ann Arbor, Michigan, 48106.

antecedents of affect and well-being (e.g., Lazarus, 1966; Campbell, Converse, & Rodgers, 1976), little is actually known in an integrated manner about the conditions under which time perspectives gain and lose salience and affect well-being. Indeed, the psychological past, present, and future have received little attention despite the long-standing hypothesis that the behavior of the person is governed by all three time perspectives (Lewin, 1951). This study tests hypotheses about the relative contribution of these time perspectives to positive and negative affect and to somatic complaints. There are two major propositions: (a) that all time perspectives are not equally salient but each may make unique contributions to well-being or ill-being, and (b) that the components of stressful events that affect emotion and somatic symptoms may be different for past, present, and anticipated future events. Model and Hypotheses The hypotheses are derived from a more elaborate theory of person-environment misfit and time frames described elsewhere (Caplan, 1983). The hypotheses, that follow, examine ISO

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

PAST, PRESENT, AND FUTURE FIT

the effects of misfit on well-being. The elements of the hypotheses are defined as follows: Stress is defined as the lack-of-fit between commensurate characteristics of the person (P) and of the environment (E). One may examine the fit of abilities and needs of the person with commensurate demands on those abilities and supplies to meet those needs. This conceptualization is based on the theory of French, Rodgers, and Cobb (1974).1 Alternatively, one can view such fit as the level of attained versus desired performance, where the desired level refers to a standard set by self or others. Subjective fit refers to perceptions by the person, whereas objective fit refers to fit assessed independently of the subjective biases of the person. This study focuses on subjective fit. Nevertheless, this study includes a small amount of validation data on the link between subjective fit and an objective measure of performance. Three forms of subjective fit are used in the hypotheses. Retrospected fit (FR), refers to the person's recollections of fit at some past time, (t — n). Current subjective fit (Fc), refers to the person's reports about the present. The third form of subjective fit is anticipated fit (FA), and it refers to some particular point in the future (t + n). The hypotheses apply to both perceived fit and Fs. Fs is fit that is computed by asking the person for separate ratings of the subjective person and the subjective environment (Ps and Es, respectively). Perceived fit is obtained by asking the person directly to provide a single rating of goodness of fit. In this study the tests of the hypotheses use perceived fit rather than Fs. Using a measure of directly perceived fit eliminates the opportunity to test the contribution made by P and E components in the assessment of stress and the prediction of strain. Such tests, however, have been conducted elsewhere. They have shown that PE fit consistently explains additional, albeit small, amounts of unique variance in strain and well-being beyond that explained by either E or P alone (e.g., Harrison, 1978; French, Caplan, & Harrison, 1982; Kulka, Klingel, & Mann, 1980). Although the main purpose of this study is not to test PE fit theory, per se, the concept of perceived fit is retained

181

conceptually and operationally because it reminds us that the stressors in this study are the result of some interplay between characteristics of the person and of the environment.2 The remaining elements in the hypotheses are broadly referred to as strains. They are deviations from normal states of the person and can include negative affect or loss of positive affect, somatic symptoms, and illness states. Main

Effects

Before stating hypotheses about the relative effects on strain of past, present, and future misfit, it is necessary to consider how these three time perspectives may, themselves, be interrelated. Hypothesis 1: Perceptions of Jit from different time frames influence one another. This hypothesis derives, in part, from level of aspiration theory (Lewin, Dembo, Festinger, & Sears, 1944) and from more recent derivations of that theory, specifically the theory of learned helplessness (e.g., Seligman, 1975). Current (Fc) and past (FR) perceived fit should influence anticipated fit (FA) to the extent that Fc and FR are a basis for expectancies regarding preparedness to handle the future. Current fit, Fc, should be related to retrospections about past fit, FR, to the extent that the past performance influences current attempts to cope with environmental challenges. The prediction that follows is that F R) Fc, and FA will be positively correlated. A more specific hypothesis follows. Hypothesis la: The more adjacent the time frames for subjecti\e fit, the greater their mutual influence. Following Lewin, this hypothesis involves the assumption that the psychological field is, in part, organized by

1 A full examination of person-environment fit theory is not possible here. For theoretical and empirical references, see French et al., 1974; Harrison, 1976, 1978; Kulka, 1979; French et al, 1982; and Caplan, 1983. 2 In this study E (the environment, annual university exams), is a relatively invariant stressor across all the respondents so that the main variant is P in the PE fit model. Although the study does provide an adequate opportunity to test the effects of time perspectives, a test of the value of PE fit theory itself would require variance in P and E.

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

182

R. CAPLAN, R. TRIPATHI, AND R. NAIDU

time so that adjacent fields influence one another more because fewer psychological time boundaries need to be permeated. The prediction that follows from this topological conception is that current subjective fit will be more strongly correlated with retrospected fit and anticipated fit than retrospected fit and anticipated fit will be with each other. Adjacent past and present subjective time frames should also be interrelated to the extent that current objective fit (F0) is most influenced by more recent than by less recent, past F0. This proposition is not testable in this study, however, because measures of F0 were not included. Hypothesis 2: Subjective misfit increases strain. This hypothesis has received support in a number of studies covering a wide variety of respondent groups: in 22 different occupations (Harrison, 1976, 1978; Caplan, Cobb, French, Harrison & Pinneau, 1980; French et al., 1982), in work settings in several countries (Tannenbaum & Kuleck, 1978), in the community of Tecumseh, Michigan (House, 1972), in samples of high school students (Kulka, Klingel, & Mann, 1980; Kulka, Mann, & Klingel, 1980), among elderly persons (Kahana, Liang, & Felton, 1980), and among Navy personnel (French, Doehrman, Davis-Sacks, & Vinokur, 1982). The examination of retrospected and anticipated as well as current fit is a new element in this research, and it leads to the following general research quetion. Research Question 1: What are the relative, independent contributions of subjective retrospected, current, and anticipated fit to strain? Research comparing anticipation of events to confrontation with events suggests that anticipated misfit will have the most powerful effect on strain (see Lazarus, 1966; Kasl & Cobb, 1979), partly because anticipation involves the additional stressor, ambiguity. A similar prediction is suggested by Gurr (1970) and by Morrison (1971). They suggest that "anticipated relative deprivation" will have a greater effect on strain than the gap between past expectations and current outcomes. Such a hypothesis contains the assumption that future well-being and goal attainment are generally more important to people than current well-being and attainment. Past misfit, like anticipated misfit, may also affect current well-being. Effects of past

misfit on strain may appear after a long period of time. For example, past misfit on social support, due to loss of a spouse, may increase the risk of mortality, and it may take a year for this effect to manifest itself in illness (see review by Kaplan, Cassel, & Gore, 1977). Second, past traumatic events may cause current upset as the person relives them, as in the horrors of war (Figley, 1978). Third, past misfit may reduce the self-perceived resilience of the person to cope with subsequent stressors and may thereby increase current levels of strain. Fourth, past experience, as discussed in much of the relative deprivation literature (e.g., Davies, 1969; Taylor, 1982) may serve as a comparison standard against which the fulfillment of current expectations may be judged. Such past experience may then contribute directly to current feelings of deprivation and to resulting strain. These latter explanations involve potential interaction effects between time frames of misfit, which are addressed briefly in the study. Lastly, this study examines two types of misfit, cognitive and motivational, and their relation to several indicators of strain. Cognitive misfit deals with demands on memory, cognitive organization of material, and related intellectual abilities. Motivational misfit deals with the amount of effort the person is willing to supply to meet demands. No specific hypotheses or predictions are offered regarding the potential differences in how cognitive and motivational misfit, across the time frames, may influence strain. Nor are any specific hypotheses presented regarding the extent to which misfit may affect some types of strains more than others. The analyses, however, are specifically designed to examine these issues as research questions. The strains fall within the following categories: somatic complaints, negative affective states (such as depression and anxiety), and positive affective states (such as satisfaction). Some hypotheses suggested by differences in the effects of cognitive and motivational fit on these strains are examined in the Discussion section. Method Sample Data wre gathered via self-administered questionnaires handed out to 293 arts and sciences undergraduate

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

PAST, PRESENT, AND FUTURE FIT

students of the University of Allahabad, Allahabad, India. Questionnaires were completed in the classes at the time of distribution. Eighty-six of those questionnaires were rejected during preparation for coding because of varying amounts of incomplete data, providing an effective response rate of 71%. The responses of the remaining 207 students were then coded and used for analysis. No information is available on the differences between the respondents and nonrespondents. The respondent set, however, was in no way intended to be random or representative of a particular population. The 207 students had a mean age of 18.6 years (SD = 1.4 years). Males comprised 72.5% of the sample; 27.5% were female. All of the females and 45.3% of the males came from urban backgrounds (high school education in a city); the remainder were classified as coming from a rural background (high school in a village). These distributions are similar to those found in the university. About two-thirds of the students were enrolled in liberal arts and the remainder were enrolled in the sciences. Slightly more than half the students belonged to the first year class, and the remainder belonged to the second year class.

Index Construction

Data Collection

Subjective Perceived Fit (Fits)

The data were collected during 1980, roughly a year after the past and 4 to 5 months before the next exams.3 The subjects were unpaid volunteers who were recruited in their classrooms and asked to complete self-administered, structured questionnaires at the time of recruitment. Students were asked to read the introduction on the cover of the questionnaire and the instructions regarding the manner of responding. Next, any questions regarding the procedure were answered. The cover page described the project as a study of examinations as a part of student life. The questionnaire was group administered and collected personally by one of the Indian investigators. The average time taken for completing the questionnaire (which included other variables not examined here) was 1 '/2 hrs. Past examination performance from university records was collected on a small subset of students. To this end, 30 students were asked if they would write their names on the back of their questionnaires and all agreed to do so.

The Measures Items were written for the first draft of the questionnaire in English on the basis of either theory or validated instruments already in use. Then we created a Hindi translation and back translation. The first draft of the questionnaire was individually administered to 30 undergraduate students to check for clarity of meaning in the items. Following revision, the questionnaire was administered to the current sample. Appendix A presents the item content of up to three items for each of the measures. The full content is available from the authors. Here, and throughout the text, all measures will have their first letter capitalized (e.g., Anxiety) to distinguish them from references to the general construct (e.g., anxiety). English translations that keep as close as to possible to the Hindi questionnaire are used in this report. The full content is available from the authors.

183

Where possible, previously validated measures were used. For the measures of stress, this was not possible, and pools of items were generated on the basis of theory. For all multiitem indices, hierarchical cluster analysis (Johnson, 1967) was used to determine if items hypothesized to form an index were more highly correlated with one another than with items hypothesized to form another index. In general, indices were retained if they had internal consistencies (coefficient alpha) of .60 or higher. As noted later, one exception was made in order to include at least some estimate of a concept thought important to the study. The number of items in each index and the alpha, mean, and standard deviation of each index are presented in Appendix B. Although alphas of .50 or higher were judged to be adequate for research purposes (Nunnally, 1967, p. 226), the range of reliabilities and some variance in standard deviations led us to rely on analyses that examined the standardized slopes (betas) of relations rather than correlation coefficients. Slopes, unlike correlations, are not influenced by the reliabilities of measures.

Directly perceived fit, measured by a direct rating rather than by assessing E, and P, separately, will be abbreviated as Fit (e.g., FitR, File, FitA). The symbol F is reserved only for measures of fit that are derived from separate measures of E and P. As a measure of retrospected fit, respondents were asked to focus on the last examination they had taken in their most difficult subject and answer to what extent

3 At the time of the questionnaire administration, the date of the next examination had not been announced. Its announcement was a matter of negotiation affected by student unrest. Although such circumstances were unusual, their effect was the same for all respondents in the sample. Of more substantive importance, we note that tests of hypotheses about retrospected. current and anticipated misfit will necessarily be limited in any data set to the time intervals explored. This consideration applies here without question. Also of substantive importance, by definition, retrospected and anticipated fit will always be more distant from one another than from current fit regardless of the time intervals involved between these types of fit in different studies. 4 Given that most of the studies cited in this article are based on research in North Amrica, there may be some question as to whether or not respondents from India would give answers comparable to those given by North American students in similar situations. Only replication in North America can answer this question exactly. Note, however, that the item content and structure of indices of emotional strain and of somatic complaints described later in the text are very similar to those produced in North America (e.g., Derogatis, et al., 1974). Note also that the relations between stress and strain in this study are in the same direction as that found in studies of stress in the United States among respondents as diverse as participants in the work force (e.g., Caplan, et al., 1980; French, et al., 1982) and in high school (Kulka, Mann, & Klingel 1980).

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

184

R. CAPLAN, R. TRIPATHI, AND R. NAIDU

they thought they had certain personal characteristics required by the examination. Similar questions were posed with regard to if an examination was held today and with regard to the next examination in their most difficult subject. The content and formats (see Appendix A) for the measures of retrospected, current, and anticipated perceived fit were kept as similar as possible. The question about current fit was stated in hypothetical terms ("If the examination were held today. . . ."). Nevertheless, it referred to a time perspective that is more current than the measures of anticipated and retrospected fit. In this regard, it adequately met the requirements for testing our hypotheses about the relative effects of different time frames. From an initial pool of 14 items for each time frame, two types of fit indices—cognitive and motivational— were constructed. These two dimensions can be thought of as major components determining cognitive performance (i.e., performance = motivation X cognitive ability}. Except for reference to the time perspective itself, the item content was identical for each time perspective, FitR. Fitc, FitA, and the indices are described as follows: Cognitive perceived fa is an example of abilitiesdemands fit. It refers to the degree to which respondents perceive good PE fit between their cognitive abilities (e.g., memory and logic) and the demands made by the essaytype annual examinations that they must pass to continue their studies. Motivational fit refers to how much the person was able to be devoted to academic studies and how much the person was able to work hard in preparation for the examination. As a whole, the measures of motivational fit had much lower reliabilities compared to the measures of cognitive fit (.45 to .72 vs. .93 to .95); the difference in reliabilities underscored the need to examine differences in the effects of motivational versus cognitive fit with statistics, such as slopes. The FitA measure had an internal reliability of only .45, which was judged marginally acceptable for our research purposes (Nunnally, 1967, p. 226). It was retained despite the low alpha so that some examination of motivational fit was possible for the future time perspective. In sum there were six measures of subjective perceived fit used in this study: two types of fit measures, cognitive and motivational, and three time frames, past, present, and future. The measures are labeled retrospected cognitive (Cog. R), current cognitive (Cog. C), anticipated cognitive (Cognitive A), retrospected motivational (Mot. R), current motivational (Mot. C), and anticipated motivational (Mot. A) perceived fit. As a potential indicator of the concurrent validity of the retrospected fit measures, the students were asked to report their grades in the last examinations for their most difficult subject. Comparisons of these self-reports with actual grades were possible for the 30 students who provided their names to us. The correlation between reported and actual marks was .82. This coefficient is almost identical to that found in research on American high school students for overall grade point average (Kulka, 1976). It is likely that reports of past exam scores for all the students, accordingly, were relatively accurate. Reported past grade was positively correlated with retrospected cognitive fit (r = .29, p < .01) but was unrelated to retrospected motivational fit (r - .02). The lack of

relation for motivational fit may be due in part to its lower reliability and may partly represent a substantive effect, but the relative contribution of these two explanations is indeterminate in these data.5

Indicators of Strain and Well-Being The indicators of strain and well-being used in this study have been categorized under three broad headings: (a) somatic complaints, (b) negative aflects, and (c) positive affective outcomes. Items for these indices came from different sources. Items for measuring somatic complaints and for measuring anxiety and depression, were taken from measures developed in India by Spielberger, Sharma, and Singh (1973) and by Wig and Verma (1973). They were similar to the items contained in the Hopkins Symptoms Checklist (Derogatis, et al., 1974). Besides these items, a set of items was written specifically for this study to measure academic satisfaction and satisfaction with extracurricular activities that took place within the university. Items were also written to measure self-esteem in terms of the self-identity (French, 1969; Sherwood, 1970) of students. The formats appear in Appendix A. Somatic complaints. The index influenza included the symptoms that are generally associated with this disease, such as a cold, headaches, breath lessness, and feeling tired and generally ill. Results involving this index should be interpreted with caution as no independent medical diagnosis was obtained as validation. Somatic anxiety contained manifested symptoms of anxiety (Taylor, 1953), such as disposition to crying, nervousness, heart palpitation, and poor appetite. Digestive complaints had items relating to problems pertaining to the digestive system, such as constipation and upset stomach. Negative affect. This subhead includes five indices, four of which relate to negative affect and one of which involves negative self-esteem. The indices are "anger," "depression," "anxiety," "interpersonal sensitivity," and "negative self-esteem." Positive affective outcomes. The positive affect index dealt with feeling good and happy. Academic Satisfaction contained items that expressed satisfaction with academic aspects of the university and its reputation. Extracurricular Satisfaction dealt more with cultural programs, sports, and intolerance of criticism of the university (akin to school spirit). Sense of mastery was dealt with positive evaluations of the self in relation to "mastering a difficult lesson" and to "rekindling interest in the studies whenever one got tired of studying." Correlations among all the measures are presented in Table 1.

5 The overall magnitude of these relations is similar to that in research on current PE fit among American high school students (Kulka, 1976). There, fit explained from less than 2% to 9% in overall grade point average depending on the measure of fit. In this study reported grades are not as good a predictor of strain and wellbeing (average r across 12 measures of strain and wellbeing = .09) compared to reported past fit (average r = .20). Consequently, grades are a poor indicator of fit and of strain (and vice-versa).

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

PAST, PRESENT, AND



|

£ — » r^

1

y »»s BS (So OS SS

,2 E &

S S S

FUTURE FIT

*.

32 g

185

S

s s:

*

i

o o

«tf

: s 8s

-

S 2

186

R. CAPLAN, R. TRIPATHI, AND

1-

-

'*3

od

'd

«I

| |

M

¥ |

11 1 |

S

S