the impacts of positive psychological states on...

Sot Scl Med Vol 27. No 12. pp 1447-1459, 1988 0277-9536188 53 00 + 0 00

Printed m Great Bntam All nghts reserved CopyrIght C 1988 Pergamon Press plc

THE IMPACTS OF POSITIVE PSYCHOLOGICAL

STATES ON PHYSICAL HEALTH: A REVIEW

AND THEORETICAL FRAMEWORK

JEFFREY R EDWARDS’* and CARY L COOPER*

‘Colgate Darden Graduate School of Busmess Admmlstratlon, Umverslty of Vlrgmla, Charlottesvdle, VA 22906, U S A and *Unlverslty of Manchester, Institute of Science and Technology, Manchester M60

lQD, England

Abstract-While much research has focused on the Impacts of negative psychologtcal states, such as stress, on physical health, relatively httle research has exammed the effects of posrtrue psychologlcal states We suggest this Imbalance IS attnbutable to Inadequate theoretlcal and methodological development regardmg the tmpacts of posmve psychologtcal states on health Thts paper presents a framework by which posttlve psychological states may Influence physlcal health Followmg this, we review evidence pertammg to this framework We conclude by dtscussmg methodological Issues associated with this relatively new area of inquiry

Key words-physlcal health, satlsfactlon, personal happmess, stress, posttlve emotions

INTRODUCTION

A merry heart doeth good hke a medrcrne [Proverbs 17 221

A tremendous amount of research has exammed the Impacts of Job stress on psychologtcal and phyno- loglcal disorder This research has ylelded mvaluable evidence concerning the hnk between negative psychologlcal states, such as stress, and mental and physlcal Illness [l-7] This research has provided mvaluable mformatlon regarding the nature of the relatlonshlp between negative psychological states and health However, by focusmg solely on the impacts of negutzve psychological states, this research falls to answer an equally fundamental question What are the impacts of posrtloe psychological states on health?

While this questlon has been raised before [8-121, rt has received httle serious attention m the stress field We beheve this lack of attention IS due to inadequate theoretlcal and methodological development regarding the impacts of posltlve psychological states on health Without an appropriate theoretical and methodological basis to guide mqulry, systematic research m this area IS not possible, and evidence concerning the Impacts of posltlve psychological states on health ~111 remam scattered and anecdotal

The purpose of this paper 1s to estabhsh a theoretical and methodological basis for research mto the

*Requests for repnnts to Dr J R Edwards, Colgate Darden Graduate School of Business Admuustratlon, Umverslty of Virgmta, P 0 Box 6550, Charlottesville, VA 22906-6550, U S A

tThough our onentatlon IS toward stress III orgamzatlons, we will present this dIscussIon m general terms, thereby faclhtatmg the apphcatlon of this dIscusston to the effects of posmve psychological states m both work and nonwork settings

impacts of positive psychological states on health t First, we ~111 present a theoretical framework to guide research m this area Second, we ~111 examine available evidence pertaining to the basic tenants of this framework Finally, we will discuss methodological issues facing the researcher m this relatively new area of inquiry

POSITIVE PSYCHOLOGICAL STATES AND HEALTH. A THEORETICAL FRAMEWORK

In this section, we will present a theoretical framework concerning the impacts of positive psychological states on health This framework will specify major pathways by which posltlve psychological states may influence health This framework draws from a cybernetic theory of stress, coping, and well- being presented elsewhere [13-l 51 Bnefly, this model defines stress as a negative drscrepancy between an mdlvtdual’s percerved state and desired state, prouded that the presence of this ducrepancy IS considered important by the mdnuduaf [cf 16, 171 By perceived state, we Include the mdlvldual’s cogmtlve represent- ation of his or her physical and social environment By desired state, we mean the state which the mdlvldual considers adequate [cf 181 By negatrve dlscrep- ancy, we mean that stress exists when the mdlvldual’s perceived state falls short of his or her desired state According to the model, stress will lead to two classes of outcomes One class includes indices of psychological and physlologlcal functlonmg which, taken together, represent the mental and physical health of the mdlvldual The other class of outcomes consists of copmg, which 1s defined as efforts to prevent or reduce the negative Impacts of stress on mduxdual well-bemg These efforts are directed toward the determinants of stress, 1 e the perceived and desired states mvolved m the discrepancy and/or the amount of importance associated with the discrepancy Thus,

1447

1448 JEFFREY R EDWARDS and CARY L COOPER

stress, copmg, and well-bemg are vtewed as crtttcal components of a negattve feedback loop. where stress damages well-bemg and produces copmg. which m turn mfluences the determmants of stress [cf 19. 201

The reader will note that, consistent wtth most theories of stress, thts model focuses on negatrue dtscrepanctes, or condtttons where the mdtvtdual’s perceived state fulls short of his or her desired state This model may be readily adapted to mclude postttve psychologtcal states by exphcttly acknowledgmg posltwe dtscrepancies, t e condmons where the mdt- vldual’s percetved state exceeds hts or her destred state Borrowing from Selye [2 1,221, we will define a posttwe dzscrepancy between an mdrudual’s percewed state and dewed state, prouded that the presence oj this drscrepancy IS consldered tmportant by the tndwd- ual, as eustress This defimtton 1s analogous with Selye’s [21,22] discusston of eustress, which refers to the pleasant, curattve stress of fulfillment This defimtton is also conststent wtth conceptuahzatton of other postttve psychologtcal states, such as lob satisfaction [l&23], perceived quality of hfe [24] and subjecttve well-bemg [25]

There are two major processes by which eustress may Influence health One process mvolves the dtrect effect of eustress on health That is, eustress may evoke physiologtcal responses which, m the long run, Improve physical health Thts process may be viewed as the converse of stress and 1s therefore hkely to influence health through many of the same under- lying mechamsms A second process involves the effect of eustress on coping In other words, rather than affecting health directly, eustress may Influence health mdtrectly by facihtatmg attempts to cope with existmg stress, such that copmg acts as a medtator of the relationship between eustress and health [I I] In the remamder of this section, we wtll elaborate these processes, m the followmg section, we will consider evidence regardmg then vahdity

Direct effects of eustress on health

Three aspects of the stress and coping process may produce eustress [ 151 Ftrst, eustress may result from the direct appratsal of factors m the environment as meeting or exceedmg desires For example, a worker may vtew hts or her mcome as more than adequate to meet hts or her needs Thts source of eustress IS directly comparable to Locke’s [18,23] defimtton of Job sattsfactton, whtch refers to a postttve psychological state resultmg from the mdtvtdual’s appratsal of the work situation. as meetmg his or her values Second, eustress may result from engaging m copmg acttvtties whtch are vtewed as inherently enloyable For example, an executive may cope wtth quahtattve overload by enrollmg m an executtve education program and find the program inherently enloyable, regardless of its impact on experienced stress Third, eustress may result from the successful accom- phshment of copmg acttvtttes per se That IS, holding constant the Impacts of copmg efforts on the focal stressor and the Inherent properties of coping acttvmes, the mdtvidual may derive pleasure from the sense of accomplishment assoctated with the successful tmplementatton of coping acttvmes For example, a manager may cope wtth the stress of a poor fitness

examination by engaging m an exercise program which. though perhaps averstve, may produce a sense of accomplishment upon Its completton Thus, the stress and copmg process may generate eustress through direct appraisal of the environment as exceedmg desires, engagmg m inherently enloyable copmg acttvtties, or successfully executmg copmg strategies

Given these three baste sources of eustress, what are the effects of eustress on health? One way to deduce these effects IS to constder the health consequences of states where perceptions meet desires, 1 e where neither stress nor eustress are present It 1s likely that this state IS assoctated with gradual detert- oratton m physical health, due to the phystological entropy associated wtth the agmg process This notion 1s supported by studtes mdtcatmg that, after controllmg for soctal and psychologtcal factors, age IS stall predicttve of detenoratton m health [26,27] Now consider the effects of stress on health. Exten- sive evtdence suggests that stress promotes the development of long-term, degenerattve diseases whtch ulttmately cause premature death [l, 3, 51 Given that stress promotes the development of degenerattve diseases and that, m the absence of stress, these dtseases contmue to develop but at a slower rate, we might extrapolate to predict that eustress may decrease the rate of development of degenerative diseases and, tf the extrapolatton were vahd, that htgh levels of eustress may actually reverse the development of disease, thereby tmprovmg health

The notion that any psychological state, mcludmg eustress, may improve physical health IS a radical departure from the current ‘disease-oriented’ view of the role of soctal-psychological factors m health Such a departure requires an explanatton of the phystological mechamsms by which psychological states may benefit health In an mstghtful exam- matton of thts Issue, Karasek et al [28] describe the pathways by which situations mvolvmg high demands combmed wtth high control may produce physiologtcal growth and regeneratton In general, these sttuattons are conststent wtth our vtew of eustress That IS. high control tmphes the abthty to meet the demands placed on the mdtvtdual If the mdtvtdual destres to meet these demands and constd- ers meeting them Important, then eustress wtll result Karasek et al [28] suggest that these sttuattons stimulate the production of anabohc hormones, such as testosterone, msulm, adrenalin, and growth hormone When the balance of these anabohc hormones exceeds catabohc hormones (e g corttsol), phystological growth may occur For example, testosterone and growth hormone may actually enhance protein syntheses m the myocardtum Itself Stmilarly, epi- nephrme response with quick return to baselme may selecttvely protect muscle ttssue (mcludmg the myo- cardmm), while mobihzmg other body reserves for energy production HDL cholesterol also plays a protecttve role by removing hptd depostts from the inner arterial walls [29] Thus, eustress may directly mfluence health, and perhaps improve It, by sttmu- latmg the productton of anabohc hormones, HDL cholesterol, and other health-enhancing biochemical substances Whtle thts process is certainly speculattve, it nonetheless suggests pathways by which eustress

Pos~twe psychological states and health 1449

may mfluence physlologlcal mechanisms which ultimately improve physical health

Indvect effects of eustress on health

As Indicated earlier, eustress may also influence health mdlrectly by faclhtatmg attempts to cope with existmg stress In general, eustress may facilitate coping by enhancmg mdlvldual ablhtles relevant for copmg and/or stlmulatmg increased effort directed toward coping Note that these effects focus on the reduction of physlologlcal damage associated with exrstmg stress rather than the production of physlologlcal benefit associated with eustress

The effects of eustress on coping have been dlscussed by Lazarus et al [I 11 under the rubnc of positive emotions According to Lazarus et al [l 11, posltlve emotions occur when a transaction with the environment 1s appraised as desirable, a sltuatlon which 1s essentially identical to our conceptuahzatlon of eustress Lazarus et al [ll] identify three mechanisms by which eustress may facilitate coping First, eustress may serve as a breather from ongoing stress These breathers, or breaks, presumably facilitate coping by allowmg periods for creative problem- solvmg (see also Ref [30]) Second, eustress may act as a sustarner of ongoing coping, mcreasmg the likelihood that coping efforts will persist Third, eustress may serve as a restorer, replemshmg damaged or depleted resources or developmg new resources For example, positive expenences may bolster damaged self-esteem, which may in turn renew coping efforts While breathers, sustainers, and restorers may mdeed influence coping, Lazarus et al [l I] fall to clanfy several issues First, if breathers operate merely by temporarily removing stressful experiences from awareness, then any alternate cognition, whether Its emotional content 1s posltlve, negative, or neutral, would serve as a breather (obvl- ously, a negative alternative cognition may cause additional stress, thereby nulhfymg Its beneficial effect as a breather) it IS therefore unduly restrictive to classify breathers as necessanly posltlve m emotional content Second, Lazarus et al [l l] describe challenge and hope as two examples of positive emotions which sustain coping However, since both challenge and hope also mvolve a sense of optlmlsm regarding future events, it 1s unclear whether positive affect or an optlmlstlc outlook presumably facilitate coping [cf 31, 321 Third, to function as a restorer, it seems that posltlve expenences must be relevant to the coping efforts m use For example, It IS doubtful that success m artistic endeavors would facilitate coping which requires analytical problem-solvmg Thus, while the process by which breathers, sustam- ers, and restorers may influence coping certainly require further clanficatlon and development, they nonetheless suggest three possible mechanisms by which eustress may faclhtate coping

Eustress may also facilitate coping through moderators identified m earlier stress and coping research [33] For example, the positive affect associated with eustress may improve social interactions by de- creasing social distance and wmnmg social approval, which may m turn faclhtate social support [34-361 Eustress may also discourage the display of Type-A behavior That IS, the posrtlve affect associated with

eustress 1s mconslstent with the negative affect associated with charactenstlc Type-A behaviors (e g ag- gresslveness, hostility, anger, Irntablhty) and may therefore suppress their occurrence In addition, eustress may faclhtate feelings of mastery and control, which may m turn faclhtate coping [37, 381 In fact, the causal relationship between eustress and mastery may be bldlrectlonal, such that eustress generates feelings of mastery, which m turn promote posltlve and perhaps humorous responses to stress [3 1.391 As these examples indicate, rather than playing an ISO- lated role m coping, eustress may be an important component of stress-buffenng factors, which have been the focus of previous research

The mechanisms linking eustress and coping described above may be unified under a more parsl- momous framework, by vlewmg coping as simply a task requiring both effort and ability [cf 40,411 The ultimate objective of this task 1s to alter the determinants of stress, I e the discrepancy between perceived and desired states and/or the amount of importance associated with this discrepancy [ 14,421 This view lmphes that, m order to faclhtate coping, eustress must enhance either the amount of effort directed toward coping or the ablhtles required to alter the determinants of stress [14] Eustress may enhance coping effort by promotmg a sense of self- efficacy and optimism, particularly when the source of eustress 1s relevant to the coping task at hand [cf 11,31,32] For example, a product manager who expenences eustress from a senes of successful product mnovatlons may cope with the stress of a new product from a competitor, by enthuslastlcally ap- plying effort toward addltlonal product mnovatlons Eustress may also enhance coping effort when It has been produced by coping m slmllar circumstances That is, d coping m similar circumstances has been pleasurable, either inherently or based on Its outcome, then this pleasure may act as an incentive to apply increased effort toward the coping task at hand In addition to Its effects on coping effort, eustress may also enhance coping ablhtles For example, stress may hinder copmg by interfering with the mdlvldual’s ability to generate and evaluate coping alternatives [I41 It follows that periods of eustress may facilitate these ablhtles, thereby promotmg effective coping As indicated earlier, eustress may also enhance coping ablhtles by facthtatmg social interaction, which may m turn provide coping resources m the form of social support [34-361 These examples illustrate only a few of the various mechanisms by which eustress may facilitate coping While other mechamsms surely exist, we suggest that they share a common process, lmprovmg coping by enhancing either the effort directed toward coping or the ablhtles reqmred to alter the determinants of stress

Summary

In summary, eustress may improve health directly through hormonal and blochemlcal changes or mdlrectly by faclhtatmg effort and ablhtles directed toward copmg with exlstmg stress Whtle these pathways are certainly speculative, they nonetheless outline potential processes by which eustress may improve health In the followmg section, we ~111


attempt to examme the validity of these processes incongruity or ambiguity IS analogous to the eustress by reviewing evidence pertaining to the effects of associated with successful copmg (I e resolvmg a eustress, as represented by various posmve psycho- discrepancy between perceptlons and desires) There- logical states, on health fore. evidence regardmg the physIological impacts of

humor IS relevant to our analysis of the consequences

EMTRESS, COPING, AND HEALTH A REVIEW OF THE EVIDENCE

Research concernmg the effects of eustress, as defined here, on health IS very scarce However, by broadening our view of eustress to mclude posltlve psychological states m general. we can Identify several areas of relevant research These areas include anecdotal evidence, laboratory studies of positive affect and humor, positive life events, and Job satisfaction We conducted a computerized search of the orgamzatlonal, psychological, and medical literatures to identify relevant studies m these areas In this section, we will review and evaluate this evidence and note questlons left unanswered by existing research

Eustress and health

Anecdotal eurdence Recently, Cousins [8,9] re- counted his recovery from ankylosmg spondylttu, a paralyzmg collagen Illness After a distressing and unsuccessful hospital stay, Cousms relocated hlmself to a hotel, stopped his medlcatlon, and commenced a program of megadoses of vitamin C and humorous movies and books, each of which appeared to mde- pendently improve his condltlon After 4 weeks. Cousins was able to walk, and several months later he returned to nearly normal functlonmg Moody [44] also presented anecdotal evidence from several sources For example, he cited a case where clowns brought a catatomc girl out of her stupor He also described Joseph Gnmaldl, a mneteenth century comic. who apparently returned the hearing and speech of a deaf and dumb sallor during one of his performances The sailor attributed the loss of his faculties to Intense exposure to heat and sun during a recent tnp Durmg Gnmaldl’s performance, the man struggled to express his great amusement over the antics on stage and apparently succeeded Moody also cited a doctor of gerlatrlc medlcme who con- cluded that one attribute which all of his elderly patients had m common was a good sense of humor Furthermore, Moody hlmself claims to have wit- nessed cases where patients are told they have only a few weeks or months to live but, due to a positive outlook and a will to live, survives for years While the vahdlty of these accounts IS obviously limited. due to small sample size, the confoundmg of positive psychological states with other variables (e g vitamin C), and the absence of control groups, they nonetheless Illustrate potential benefits of eustress on health

Laboratory experrments The bulk of the experl- mental evidence regarding the Impacts of positive psychological states on health mvolves the physlologlcal concomitants of humor and laughter Ac- cording to cognitive humor theories [44,45], humor involves the perception and subsequent resolution of some mcongrulty or ambiguity If we assume that the mcongrulty or amblgulty IS undesirable [46,47], then this situation may be considered a source of stress It follows that the humor associated with resolvmg

of eustress This evidence IS hlscussed below Experiments examining the physiological impacts

of humor and laughter have pnmanly focused on indices of arousal. particularly heart rate, skm conductance, muscle tension, and respiratory patterns This evidence has been aptly reviewed by McGhee (481 and Robinson [49], and a comprehensive review here would be unnecessarily redundant However, a summary of these findings will facilitate our dlscusslon Evidence regarding the Impacts of humor on arousal dates back to Spencer [SO], who found altered respiratory patterns and increased heart rate and muscle tension associated with humor Later, Martin [51] found increased heart and respiration m subjects who found cartoons amusing More recently, Jones and Harris [52] presented 20 cartoons to nme undergraduates and measured rated humor and heart rate Results indicated that humor ratings were associated with increased heart rate Langevm and Day [53] presented 12 cartoons to I5 SubJects, measuring heart rate, skm conductance, and rated funniness Heart rate changes and maximum responses, and GSR amplitude and recovery time, were posltlvely related to rated funniness In a series of experiments exammmg the physlologlcal concomitants of laughter, Fry [54-561 found that laughter reduced skeletal muscle tone, increased respiratory activity, oxygen exchange, muscular actlvlty and heart rate, and stimulated the cardiovascular system, the sympathetic nervous system, catecholamme excretion, and endorphm production

Several experiments attempted to assess changes m arousal during the course of a humorous experience For example, Goldstem and his colleagues [57] gathered heart rate. skm conductance, and funniness data from 20 male undergraduates who were presented seven riddles and seven problems Results indicated that heart rate increased during the question portion of both nddles and problems, and decreased shghtly during answers Chapman [58] divided SubJects mto groups who had given high or low funnmess ratings to Jokes Independent of actual laughter, muscle tension was greater among the high funnmess rating groups throughout the course of the Joke Godkewltsch [59] conducted two related experiments In the first experiment, 24 female undergraduates read 24 Jokes m which Joke bodies and punch hnes were separated Measures included skin conductance, heart rate, self-reported arousal, and rated funniness Results mdlcated that skm conductance rose during the Joke body and that, at the punch hne, heart rate. skin conductance responses, and self-reported arousal were posltlvely related to rated funniness The second experiment, using the same stimuli but measuring only self-reported arousal, found that self-reported arousal was posltlvely related to funniness ratings gathered from the first experiment Finally, Bushnell and Scheff [60] examined the physlologlcal consequences of laughter over a longer time frame and found that, while laughter temporarily increased sympathetic nervous system activity,

Posltlve psychological states and health 1451

It ultimately resulted m a net decrease In arousal and the causal pathways between emotion and arousal Improvement m mood for up to 45 mm may be bldlrectlonal

Several studies exammed physlologlcal responses to stlmuh designed to produce multiple affective responses Sternbach [61] examined autonomrc responses (skin conductance, gastnc motlhty, resplra- tlon rate, heart rate, eyebhnk rate, finger pulse volume), among 10 children, to sad, fnghtenrng, pleasant, and humorous scenes m the movie Bumbz After the movie, subjects identified the scenes which were most sad, frightening, pleasant (happy), and funny Results mdtcated a decrease m skm conductance and eyebhnk dunng the saddest scene and a decrease m gastnc motlhty dunng the happiest scene In a later study, Levi [62] exammed mood and urinary catecholammes responses among 20 female office clerks who vlewed neutral, tragic, humorous, and fnghtemng films Levi found that the tragic and humorous films elrclted moderate Increases m adrenalin, while the fnghtemng film ehclted the greatest mcrease m adrenalin In contrast, the humorous and frlghtemng films elicited moderate mcreases m nor- adrenalin, while the tragic film did not ehclt a slgmficant increase m noradrenahn Aver111 [63] examined affective and physIologIca responses (blood pressure, skm temperature, finger pulse volume, skm resistance, respiration) among 54 subjects vlewmg a sad, comedy, and control film Results mdlcated that the sad film produced higher blood pressure, while the comedy films produced greater respiratory actlv- Ity (due to laughter) Both the sad and comedy films produced greater electrodermal actlvlty than the control film Finally, Schwartz et al [64] exammed blood pressure and heart rate changes among 32 sublects engaging m happiness, sadness, anger, fear, control, and relaxation Imagery, both seated and durmg exercise Happmess elicited higher systohc blood pressure and heart rate than relaxation and control, lower heart rate than anger and fear, and higher dlastohc blood pressure than relaxation but lower dlastohc blood pressure than anger

An expenment by Sales [67] focused on changes m serum cholesterol rather than sympathetic arousal Sales presented anagrams at two levels of difficulty to 73 sublects for 12 5-mm sessions Measures Included serum cholesterol. subjective workload, and task enJoyment Results indicated that subjects m the high ObJectlve workload condmon, who perceived low workload, exhlblted an Increase m cholesterol, while subjects m the low obJective workload condltlon, who perceived high workload, exhlblted a decrease m cholesterol Sales surmised that these changes m cholesterol were associated with task enjoyment, such that SubJects m the high objective workload condltlon, who perceived low workload, felt they were doing poorly on an easy task, while subjects m the low ObJective workload condltlon, who perceived high workload, felt they were domg well on a difficult task This reasonmg was supported by a significant negative relatlonshlp (r = -0 24) between task enJoyment and change m cholesterol

Finally, Edwards [68] speclfically examined satls- faction and stress as competing predictors of physlo- logical outcomes Edwards presented anagrams of various levels of difficulty to 36 subJects for three trials Blood pressure, heart rate, and self-reported stress and satlsfactlon were measured for each trial Results mdlcated that satlsfactlon, as a positive mood state, was unrelated to stress, while stress and dlssatlsfactlon were highly related However, neither blood pressure nor heart rate were related to either anagram difficulty, satlsfactlon, or stress

Finally, two expenments conducted by Schachter and his colleagues suggest that, rather than resulting from emotion, arousal may interact with cogmtlon to produce emotion Schachter and Singer [65] Injected 185 SubJects with either epmephnne or a placebo These SubJects were then placed m situations designed to elicit either euphona or anger Results mdlcated that, compared to sublects who received the placebo, SubJects who received epmephrme reacted with moderately greater euphoria m the euphoria condltlon and with moderately greater anger m the anger condltlon Schachter and Wheeler [66] attributed the small size of these effects to the posstblhty that subJects who received the placebo may have nonetheless been aroused through their own sympathetic nervous system To correct this, Schachter and Wheeler [66] injected 132 subjects with either an arousal-reducmg, arousal-mcreasmg, or placebo substance and measured laughter during a slapstick movie Compared to subjects who received the placebo, sublects who recetved the arousal-mducmg mjectlon laughed more, while those who received the arousal-reducmg InJectIon laughed less Note that these experiments do not rule out the notlon that emotion produces arousal, but Instead suggest that

With few exceptions [61,68], experimental evidence mdlcates that eustress, as represented by humor, laughter, and satlsfactlon, 1s associated with m- creased sympathetic arousal However, several studies suggest that Increased arousal IS not umque to eustress. such that the lnrensrty of the emotional experience, regardless of Its affective tone, 1s assoc1- ated with arousal [62,64-66] Furthermore, there 1s some evidence that increased arousal associated with eustress may be transient, such that eustress ultl- mately reduces arousal [60] This notion 1s supported by Cousms [8,9], who reported that a session of hearty laughter faclhtated hours of unmterrupted sleep Slmllarly, Freud [69] and Berlyne [70] have argued that humor and laughter serve a cathartic effect by releasmg anxiety and tension and ultimately reducmg arousal In the single experiment measurmg serum cholesterol [67], eustress apparently reduced cholesterol over a 1 hr Interval From this evidence, we may tentatively conclude that eustress IS assocl- ated with temporary Increases m arousal, but may ultimately produce physlologlcal benefits in the form of decreased arousal and perhaps lower serum cholesterol

Posrtrve h/e events While a number of studies have explored the Impacts of posltlve hfe events on mental health [71], we found only two published studies which examined the impacts of positive hfe events on physical health Chlnboga [72] assessed the impact of posltlve, negative, and total hfe change on health and morale among 179 adults Measures mcluded a 48-Item hfe events scale with each Item rated for affective Impact (happy vs unhappy) and degree of


preoccupatton wtth the event and self-reported phystcal symptoms (energy, Improvement m health, m- crease m health problems) Results Indicated a stgmficant negattve relatlonshlp between posrtrve events and health problems (r = -0 17) In addttton, negattve events were negattvely related to energy level and improvements m health In a later study, Svensson and Theorell [73] mtervtewed 19 hyper- tensive, 16 normotenstve, and 12 hypotenstve young men, wtth questtons focusmg on the occurrence and deslrabthty of 48 life events during each SubJectsI life course No between-group dtfferences were found for total life events or life events appraised as negative However, the hypotensrve group reported progres- sively more positive events with Increasing age than the other two groups Taken together, these studies provide tentative support for a modest negative relatlonshrp between posltlve life events and phystcai health symptoms

Job and Izfi satlsfactzon As mdtcated earlier, our defimtlon of eustress IS consistent with exlstmg con- ceptualizations of satisfaction [ 18, 231 Therefore, studies exammmg the relatlonshlp between satls- faction and health are directly relevant to our dls- cusston Several studtes have assessed the relattonshrp between sattsfactton and health usmg cross-secttonal designs A number of these studies have relied on self-report measures of health For example, Hmkel and hts colleagues [74] collected interview and questionnaire data from 100 Chinese graduate students to examine the relattonshtp between psychologtcal func- ttonmg and tllness dunng the previous 20 years Students reporting a greater number of illness ept- sodes viewed their hves as dtfficult, demandmg, and unsatrsfymg, while those reporting fewer illness epr- sodes viewed then lives as mterestmg, vaned, and relatrvely sattsfymg More recently, Wemtraub [75] examined the relatronshtp between Job satisfaction and nsk of heart attack, musculoskeletal dysfunctton, and psychosomattc dysfunctton among 1279 employed adults Results Indicated a negative relattonshtp between Job satisfaction and nsk of heart attack for female workers, and between Job sattsfactton and psychosomattc dysfunctton for both men and women Near et al [76] collected interview data from 1041 persons, focusing on occupatton. occupattonal prestige, Job tenure, housing, demographics, sattsfactton, and perceived health Results Indicated weak but stgmficant correlations between health and Job sattsfactton (r = 0 09), life sattsfactton (r = 0 24) and improvements m quality of life over time (r = 0 20) Manning [77] exammed the relattonshlp between Job sattsfactton, phystcal health, and physical fitness among 122 parttcrpants m a human performance laboratory and found a negative relattonshtp between mtrmstc Job satrsfactton and both phystcal health and phystcal fitness Lmn and assoctates 1781 examined the relatlonshrp between life sattsfactton and Job satisfaction, stress, coping behaviors, health habits, and mental and phystcal health among 211 practtcmg Internists, and found a negative relattonshtp between life satlsfactton and chronic symptoms and diseases (r = -0 18) Metr and Melamed [79] measured Job sattsfactton, anxiety, somatrc complamts, and vocattonal, avocatronal, and demands-abthttes congruence among 74 female

pnmary school teachers Results mdtcated negative relattonshtps between Job sattsfactton and both anxiety and somattc complamts Finally, Rahman and Sen (801 examined the relattonshtp between Job sattsfactron and stress, performance, and health among 150 workers engaged m self-paced repettttve work and found a posmve relattonshlp between Job sattsfactton and health

Several cross-sectronal studies have used ObJecttve rather than subJective measures of physical health For example, Schar er al [81] measured perceived stress, Job satisfactron, famtly concerns. neuroticism. lung capactty, blood pressure, obestty, serum cholesterol, sugar and protein m urine. smokmg. phystcal activity, flextbthty, history of cardtovascular disease. and family history of cardtovascular dtsease among 885 Swiss factory workers Results indicated negative relattonshtps between Job sattsfactton and percetved stress (r = -0 46) serum cholesterol (r = -0 07) and cardtovascular disease (r = -0 07) More recently, McDonald and Gunderson [82] collected de- mographtc, health, and Job sattsfactton data from 583 1 naval enhsted men, using number of dispensary vlstts as an mdtcator of health Results indicated negative relattonshtps between Job sattsfactton and number of dtspensary vtstts (r = -0 12) Hauenstem er al [83] mtervtewed 508 women, focusmg on the impact of work load, importance of work, strain. and sattsfactton on dtastohc blood pressure A negative relattonshrp was found between sattsfactton and drastohc blood pressure, parttcularly among low SES women and working women who preferred the role of housewrfe Khaleque [84] exammed the re- latronshrp between Job satisfaction, percetved effort, and heart rate during working hours among 22 female factory workers Results mdtcated a negative relatronshlp between Job satrsfactton and heart rate (r = -0 53), but no relattonshtp between perceived effort and either Job sattsfactton or heart rate Jenkins and hts colleagues [27] conducted a more comprehen- stve study, collectmg mtervtew, questtonnatre, and medical data from 204 male angina patients awattmg coronary by-pass surgery Results indicated a negative relattonshtp between hfe sattsfactton and resting angma (r = -0 26) House and hts colleagues [85] exammed the relatlonshtp between stress, satts- faction, and vartous tllness criteria (angma, ulcer, cholesterol, blood pressure) among 1809 white male factory workers Sattsfactton was negatively related to ulcer, hypertenston, and overall CHD risk, oper- attonahzed as scoring high on at least two of smokmg, blood pressure, and cholesterol Fmally, French er al [86] collected mtervtew and phystologtcal data (systohc and dtastohc blood pressure, cholesterol, thyroid functron, uric acid, corttsol) from 390 workers m eight occupattonal groups Intervtews mcluded questions regarding Job satisfaction and person- environment fit on four dtmenstons (quantitative work load, Job complexity, responsrblhty for persons, role ambrgurty) These fit questlons asked workers to indicate both current and preferred levels for each Job dlmennon, thereby provldmg a direct oper- atlonahzatlon of eustress as described m this paper Results Indicated that posltlve dlscrepancles (1 e situ- atlons where perceptions exceeded deslres) for role ambtgutty were negatively related to systolic and


dlastohc blood pressure However, given the large number of relatlonshlps tested (32), these results should be interpreted with caution

Other studies of the relationship between satls- faction and health have employed longltudmal rather than cross-sectional designs For instance, Palmore [87] examined the relationship between satisfaction and longevity by followmg 268 community volun- teers for 15 years Palmore found that, after controllmg for age, the strongest predictor of longevity was work satisfaction (r = 0 29), followed by happiness (r = 0 26). physical functlomng (r = 0 21), and to- bacco use (r = -0 21) More recently, Vadlant [SS] collected questionnaire and interview data from 188 men over a 40-year period, focusing on childhood environment, psychologlcal adJustment (including Job satisfaction), use of alcohol, cigarettes, and mood altering drugs, obesity, and physical health Results Indicated that both overall psychologlcal adJustment and Job satisfaction as an independent predictor were lower among men who were 111, disabled, or had died Verbrugge [89] collected interview data from 589 men and women, assessing health status, health behaviors, stress, anxiety, social roles, and time constraints Followmg this, dally health records were collected for 6 weeks, assessing general health status, symptoms, curative and preventive actlons, mood, and special events Results indicated that people who hked their work reported better health, fewer chronic condo- tlons, fewer dally health problems, less conversation with friends about symptoms, and better daily moods Satisfied people also tended to take more curative and preventive care for reasons unrelated to health status Rosse and Huhn [90] interviewed 42 newly hired hospital employees over a 23-week pe- nod, measuring satisfaction, ldentlficatlon with the company, ahenatlon, avoidance behaviors, attempts at change, somatic health, and mood Results mdlcated a negative relatlonshlp between Job satisfaction and health symptoms Casslleth and associates [91] examined the relatlonshlp between Job and life satisfaction and disease prognosis (relapse and survival rate) among 359 patients diagnosed as having un- resectable cancers, melanoma, or breast cancer No relatlonshlp was found between prognosis and either Job or hfe satisfaction Finally, House and his colleagues [26] examined the relatlonshlp between Job characteristics, Job tension, Job satisfaction, intrinsic and extrmslc rewards, health behaviors, CHD, hypertension, bronchitis, and mortality among 2754 men and women Analysis mvolved cross-sectional com- parisons at time 1 and predictions of subsequent mortality After controllmg for age and health behaviors Job satisfaction was unrelated to CHD. hypertension. bronchitis, or mortality

Several investigators used the case-control method, comparing SubJects with manifest disease to subJects without disease but with otherwise comparable nsk For example, Bruhn et al [92] compared 26 CHD patients who had died with 21 matched controls. using data from open-ended interviews and personality tests (e g MMPI) Analysis focused on differences between the groups m emotlonal dram, characterized by long-term, frustrating struggles with conflicts mvolvmg deep-seated values and beliefs Results indicated that CHD patients exhibited more

emotional dram, particularly dlfficultres concermng work and mamage, while controls expressed greater satisfaction with their achievements and optlmlsm toward the future Wolf [93] compared 65 post-MI patients against matched controls on measures of aspirations, satisfactions, frustrations, personality, and cardiovascular functlomng, and found that de- Jectlon and dlssattsfactlon were more frequent among the post-MI patients Theorell and Rahe (941 used data from various psychologlcal indices to compare 62 MI survivors with 109 men comparable m age and occupation Post-MI men reported less work satisfaction but greater satisfaction with their financial state (possibly due to higher pay generated by their higher reported overtime work) Fnedman and asso- elates [95] admmlstered a 155-item questionnaire to 330 patients who later developed MI and two groups of controls, one matched for demographics and one matched for both demographics and nsk factors Results indicated that Items mdrcatlve of eustress (I e ability to laugh at Jokes, feeling happy) were endorsed less frequently by MI patients than by controls Kasl and Cobb [96] compared 56 male blue-collar workers whose Jobs were terminated by plant closings to 46 steadily employed controls m terms of happiness, affective well-being, self-esteem, blood pressure, and heart rate Results indicated that higher happiness. self-esteem, and affective well-being were associated with lower dlastohc blood pressure In a unique case-control study, LllJefors and Rahe [97] examined 32 pairs of twms who differed m CHD symptoms and occurrence (abnormal ECG during exercise, MI, angina pectons) Interview data indicated that twins with greater CHD reported lower Job and life satlsfactlon, mcludmg level of education, chlldhood expenences, adult personal relations, and working conditions

Some studies did not collect satlsfactlon and physlologlcal data from the same respondents but are nonetheless mformatlve for our discussion Hmkle (981 collected personality, lrfe history, and self-report Illness data from 68 Hunganan immigrants Hmkle rated each year of each SubJect’s adult life on a 5-point need-satisfaction scale Results mdlcated that years rated as highly unsatisfactory were slgmficantly associated with clusters of diseases of all causes Sales and House [99] present three studies comparing Job satlsfactlon scores drawn from various occupations with natlonal standard CHD mortality rates assocl- ated with these occupations The first study, using a single mdlcator of Job satisfaction, found a negative relationshIp between Job satisfaction and CHD nsk for both white-collar (r = -0 63) and blue-collar (r = -0 72) occupations The second study used a 7-item scale to measure mtrmslc, extrmslc, and overall satlsfactlon and found negative relatIonships between CHD nsk and mtrmslc (r = -0 36), extrmslc (r = -0 49). and overall (r = -0 55) satlsfactlon The third study, using Job satisfaction data from the JDI [IOO], agam found negative relatlonshlps between CHD nsk and mtrmslc (r = -0 68). extrrnslc (r = -0 62), and overall (r = -0 64) satisfaction, but only for white-collar workers All three studies were replicated, using other maJor causes of death as outcomes (tuberculosis, cancer, diabetes, hypertension with heart disease, influenza and pneumonia,


and acadents) None of these causes of death was examining the stress-buffering effects of positive life slgmficantly related to Job satisfaction events This evidence 1s reviewed below

With the exceptlon of House et al [26], studies usmg either cardiovascular functlonmg or longevity as a crlterlon generally found a negative relatlonshlp between satlsfactlon and health, while studies usmg other health crlterla typIcally found no such relatlonshlp [91,99] These findings were, for the most part, consistent across studies usmg a variety of measures of satlsfactlon and employmg cross- sectlonal, longltudmal, and case-control designs While one may be tempted to conclude that these results strongly support the beneficial effects of eustress on health, four words of caution are m order First, most measures of satlsfactlon tap both positive and negative affect, ranging from extreme dlssatlsfactlon to extreme satlsfactlon As a result, an overall negative relatlonshlp between satlsfactton and illness may actually reflect a posztlue relatIonship between dlssatlsfactlon and Illness A second and related Issue mvolves the negative relatlonshlp between satis- factlon and stress found m several studies [68,78, 80,811 Given the posltlve relatlonshlp between stress and illness prevalent m the hterature, a negative correlation between satlsfactlon and Illness may be spurious,, m that stress may cause both lower satisfaction and increased illness Third, most m- vestlgators used the Incidence of disease as a crlterla rather than indices of physlologlcal growth and re- generatlon, such as those dlscussed by Karasek et al [28] Therefore, a negative relatlonshlp between satls- faction and disease may simply reflect a reduction m disease rather than the productlon of some physlo- logical benefit (this argument also holds for the relatlonshlp between posltlve life events and health) Fourth, studies usmg global assessments of satlsfactlon [74, 761 may confound satlsfactlon with perceived health, spuriously mflatmg the relatlonshlp between these vanables In sum, while the evidence revlewed IS consistent with the notlon that eustress. as reflected m Job and hfe satlsfactlon, benefits health. these findmgs are far from conclusive

Anecdotal ecldence Two reports mdlcate that eustress, particularly humor, faclhtated social mter- action For example, Coser [34] found that humor on a hospital ward decreased social distance, soclahzed other patients mto the society, provided on outlet for hostlhtles and discontent, and allayed anxiety Kaplan and Boyd [35] also studied humor on a hospital ward and found that humor created a feelmg of mtlmacy and provided a means of wmmng social approval By faclhtatmg social Interaction. humor may also provide an opportunity to receive social support, which may m turn benefit mdlvldual attempts to cope with stress [cf 1041

Posrtwe lzfe events While a number of studies examined the stress-buffering effects of eustress on mental health [38, 1051, we found only one study which exammed its effects on physical health Cohen and Hoberman [106] admmlstered measures of life events (rating the impact of each event on a negative to positive contmuum), social support, and physical and depressive symptoms to 57 college students Results Indicated that negative events were more strongly related to symptoms than total events, while posltlve events were unrelated to symptoms How- ever, both posltlve events and social support buffered the Impact of negative events on symptoms

Summary

Optmum and hope A handful of studies have exammed optlmlsm and hope as predictors of physical health Because both optlmlsm and hope reflect expectations of a posltlve future, they may be re- garded as antlclpated eustress Hence, these studies are relevant to our dlscusslon Udelman [IO]. 1021 conducted two studies of the relationship between hope and Immune system function among psychiatric outpatlents Both studies mdlcated that hope, as measured by content analysis of speech samples. was associated with enhanced immune system function Schemer and Carver [103] examined the relationshIp between opttmlsm (1 e expectation of a posltlve future) and self-reports of physical symptoms Results indicated that optlmlsm prospectively predlcted lower symptoms

The evidence reviewed above IS consistent with the notion that eustress IS associated with improvements m physlcal health This conclusion IS supported by findings from a variety of sources, mcludmg anecdotal evidence, laboratory expenments, and studies of positive life events and Job satisfaction While the bulk of the evidence pertams to the direct effect of eustress on health, scant evidence also suggests that eustress may benefit health indirectly by faclhtatmg copmg with exlstmg stress However, due to the methodological Issues noted above, this evidence 1s merely suggestive rather than conclusive, and with the exceptlon of an experiment by Sales [67], not a single study demonstrated that eustress IS associated with an Improvement m physlologlcal functlomng rather than merely a reduction m physlologlcal damage A conclusive demonstration of the effects of eustress on health requires attention to several methodological Issues In the followmg section, we discuss methodologtcal issues we feel are particularly Important m assessmg the relatlonshlp between eustress and health

METHODOLOGICAL ISSUES IN THE STUDY OF EUSTRESS AND HEALTH

Eustress and coprng

While a reasonable amount of research has exammed the direct effects of eustress on physlcal health, very little research has focused on the effects of eustress on coping AvaIlable evidence falls mto two categories, mcludmg anecdotal evidence and research

As the preceding review mdlcates, evidence regardmg the beneficial effects of eustress on physlcal health IS suggestive but far from conclusive To generate more conclusive evidence, we must consider several methodological issues While many of these issues are relevant m tradltronal stress research, they are partlc- ularly important m eustress research In this section, we hlghhght what we feel are the promment methodological Issues m the study of eustress and health


Measurement

The studies revlewed suggest two approaches to the measurement of eustress One approach mvolves the assessment of perceptlons and desires on commen- surate (I e equivalent) dlmenslons [86], as well as the Importance of these dlmenslons A cntlcal Issue m this approach IS the selectlon of relevant dlmenslons for assessmg eustress Two basic procedures may be Identified One procedure IS to present the respondent with a preselected set of dlmenslons [40,86] This approach rests on the dubious assumption that we have correctly specified the dlmenslons which are relevant to each respondent An alternative procedure 1s to sohclt dlmenslons from each respondent mdlvldually [ 1071 This approach 1s likely to produce dlmenslons which vary across respondents but which are maximally relevant to each mdlvtdual respondent Vanatlon m dlmenslons across subjects IS not mher- ently problematic, because the proposed theory adopts a process approach [108], such that perceptions, desires, and Importance, regardless of the specific dlmenslon m question, constitute eustress However, because the respondent IS likely to provide only those dlmenslons that he or she views as lm- portant, the range of this vanable will be restricted, making statlstlcal tests mvolvmg Imporrance difficult Altematlvely, one may use dlmenslons provided by both the researcher and the respondent, thereby capltahzmg on the strengths of each approach

A second, more general approach to the measurement of eustress mvolves the assessment of posltlve psychological states, such as happiness. satlsfactlon. and posltlve affect, an approach lmphcltly adopted m most of the studies reviewed A maJor issue m this approach IS estabhshmg the presence of posltlve psychological states, rather than merely the absence of negative psychological states This issue 1s crltlcal because the presence of posmve psychological states cannot be Inferred from low scores on measures of negative psychological states There are two primary reasons for this First, measures of negative psychologlcal states may suffer from a floor effect. such that they fall to dlscnmmate varying levels of posltlve psychologlcal states This was found m the study by Edwards reviewed earher [86], m which stress and dlssatlsfactlon were highly related, whereas stress and satisfaction, as a positive psychological state, were unrelated Closer exammatlon revealed that vanatlon in satisfaction was, m fact, compressed at the lower end of the stress scale Second, research mto self- reported mood suggests that measures of posmve and negative affect are often uncorrelated‘[ 109, I IO] This suggests that, rather than representmg opposite ends of a single continuum, posltlve and negative psychological states may represent two distinct constructs. which would require separate Indices for their measurement

Measuring varlatlon m health associated with eustress also presents several challenges For example, the effects described by Karasek et al [28] generally require Invasive measurement procedures, such as drawing blood, which are costly and difficult to conduct Furthermore, previous research suggests that different psychological states produce patterns of physlologlcal response rather than simple. unified

outcomes [I I l] For example, challenge without control may elevate both catecholammes and cortlsol, while challenge combmed with control may elevate catecholammes and suppress cornsol [ 1121 These findings suggest that a comprehensive assessment of physlologlcal functlonmg may be required to detect the effects of eustress on health Finally, a major challenge mvolves demonstratmg zmprooements m physlologlcal functlomng The assessment of such improvements requires more than the measurement of the presence or absence of disease Rather, we must carefully assess a variety of mdlcators which are sensitive to both physlologlcal benefit and damage, such as serum cholesterol and growth hormone (281, and track these mdlcators over time By carefully assessmg these mdlcators, we may be able to place physlologlcal functlonmg on a continuum, ranging from illness and disease to growth and regeneration, thereby provldmg a means to test hypotheses regarding the beneficial effects of eustress on health

Design

While transient effects have been demonstrated [67], the long-term health consequences of chronic eustress have yet to be determmed [cf 113, 1141 To clearly demonstrate these long-term effects, we must adopt longltudmal designs which repeatedly assess eustress and health from a cohort of respondents Retrospective designs common m both stress and eustress research are plagued with a host of problems, such as inaccurate recall, mablhty to assess changes m health over time, and potential confoundmg of stress and eustress with health [I IS] Ideally, of course, we would randomly assign respondents to condltlons designed to generate various levels of eustress and measure changes m an array of physlologlcal mdlcators [cf 1051 However, mampulatmg eustress 1s potentially unethical and lmpractlcal, and causality often must be inferred from quasl- expenmental designs [I 151

Analws

Several issues associated with data analysis deserve mention First, If the researcher assesses eustress as a dlscrepacy between perceived and desired states, he or she may be tempted to create a smgle mdex of eustress by usmg the difference between perceptions and desires across multiple dlmenslons This procedure, and the use of difference scores m general, have been cntlclzed on both theoretlcal and methodological grounds [I 16118], and these cntlclsms need not be repeated here As an alternative, we recommend using separate mdlcators perceptions and desires m analysis, and testmg the Jomt effect of these vanables as predictors of physlologlcal functlonmg [I 16. I 181 Second, the above recommendation regarding multiple mdlcators of physlologlcal functioning requires multlvanate analytical techniques Of the studies reviewed, those which employed multiple crltena either combined them to form a summary index or analyzed each cntenon separately m plece- meal fashion Both procedures are mappropnate, m that the former results m a loss of mformatlon, the latter increases the nsk of type I error, and both fall to account for mterrelatlonshlps among the cntena variables These problems may be avoided by the use

1456 JEFFREY R EDWARDS and CARY L CWPER

of currently avallable multtvartate techniques, such as multtvartate regression [ 1191 and structural equations modehng [120, 1211

A third issue mvolves controlling for nsk factors known to affect health, such as smoking, diet, medl- cal care, and alcohol consumption While this procedure IS common and generally recommended [88], it should be emphasized that It Isolates the lndepen- dent contnbutlon of psychologlcal states to physlcal health. By dorng this, v/e overlook the posslblhty that psychological states may influence health behaviors related to nsk factors, which may m turn Influence physlcal health Furthermore, this technique estl- mates the effects of eustress as If mdlvlduals did not differ on nsk factors, thereby creating an artlficlal sltuatlon [ 1151 Again, multlvarlate procedures, par- tlcularly structural equations modeling, would help uncover the causal pathways by which standard nsk factors and psychological variables combme to Influence physical health

SUMMARY AND CONCLUSION

Compared to the Impacts of negative psychological states, the Impacts of posltlve psychologtcal states on physical health have received relatively little research interest In this article, we have presented a framework for exammmg the effects of posltlve psychological states on health and revlewed evidence regardmg these effects This framework extends exlstmg theoretical work on stress, coping, and well-being [13-l 51 by mcorporatmg eustress, defined as a pow- tlve discrepancy between perceptlons and desires, provided that the presence of this discrepancy IS considered important by the mdlvldual Eustress may improve health directly through medlatmg phystological process or mdlrectly by faclhtatmg copmg with existing stress Evidence regarding the posltlve impacts of eustress on health IS generally supportive, but vanous methodological problems prevent firm conclusions To gam further mslght mto the potential benefits of eustress on health, we recommend longltudmal research mcorporatmg comprehenslve assessment of psychological states and physlologlcal functlomng Hopefully, future research WIII address the effects of both stress and eustress on health, utlmately mcludmg crlterla of both physlcal and mental functioning

Acknowledgemenrs--This research was supported m part by the Darden School Sponsors The authors would hke to thank two anonymous reviewers for their helpful comments on an earher version of this paper

REFERENCES 23

Jenkins C D PsychologIcal and social precursors of coronary disease- NewEngl J Med i84, 244-255. 307-317, 1971 24 House J S Occupational stress and coronary heart disease a review and theoretlcal model J Hlfh sot Behar 15, 12-27, 1974 Cooper C L and Marshall J OccupatIonal sources of stress review of literature relatmg to coronary heart disease and mental 111 health J occup Psycho1 49, I l-28, 1976 Beehr T A and Newman J E Job stress, employee health, and orgamzatlonal effectiveness a facet anal-

5

6

7

8

9

10

II

12

13

14

15

16

17

18

19

20

21

22

25

26

ys~s, model and literature review Personnel Pslchol 31, 665-699. 1978 Jenkins C D BehavIoral nsk factors m coronary artery disease A Ret Med 29, 543-562, 1978 Kasl S V Epldemloiogxal contnbutlons to the study of work stress In Stress ar Work (Edited by Cooper C L and Payne R ), pp 3-48 Wdey. New York. 1978 Fletcher B B and Payne R L Stress and work, a review and theoretical framework Personnel Ret 9, 19-29, 1980 Cousms N Anatomy of an Illness (as perceived by the patlent) New Engl J Med 295, 1458-1463. 1976 Cousms N Anaromy of an Illness Norton, New York, 1979 Lazarus R S , Cohen J B , Folkman S , Kanner A and Schaefer C PsychologIcal stress and adapatatlon some unresolved Issues In Selve’s Guide IO Stress Research (Edited by Selye H ), pp 90-I 17 Van Nostrand-Remhold, New York. 1980 Lazarus R S . Kanner A D and Folkman S Emo- tions a cogmtlve-phenomenologlcal analysis In Emonon Theory, Research, and Experience (Edited by Plutchlk R and Kellerman H ), pp 189-217 Academic Press. New York, 1980 Edwards J R and Cooper C L Research m stress, copmg, and health theoretical and methodological Issues Psycho1 Med 18, 1520, 1988 Cummmgs T G and Cooper C L Cybernetic framework for studying occupational stress Human Relat 32, 395-418, 1979 Edwards J R The determinants and consequences of coping with stress In Causes, Coqmf, and Con- sequences of Stress at Work (EdIted by Cooper C L and Payne R ) Wllev. London. 1988 Edwards J R’ and cooper C L A cybernetic theory of stress, coping, and well-bemg Paper presented at the Intemanonal Congress of Psychology. Sydney, Australia, 1988 French J R P Jr, Rodgers W L and Cobb S Adjustment as person-environment fit In Coprng and Adapfafron (Edlted by Coelho G , Hamburg D and Adams J ), pp 316-333 Basic Books, New York, 1974 Schuler R S Defimtlon and conceptuahzatlon of stress m orgamzatlons Org Behac Human Perform 25, 184-215, 1980 Locke E A What IS Job satisfaction? Org Behal Human Perform 4, 309-336, 1969 Katz D and Khan R L The Social Psychology of Organrzatrons, 2nd edn Wdey, New York, 1978 McGrath J E Stress and behavior m orgamzatlons In Handbook of Industrral and Organrzarronal Psycholog> (Edlted by Dunnette M ), pp 1351-1395 Rand McNally, Chlcago, Ill, 1976 Selye H (Ed ) Selye’s Gurde to Stress Research, Vol I Van Nostrand, New York, 1980 Selye H The stress concept past, present and future In Stress Research (Edlted by Cooper C L ), pp I-20 Wiley. New York, 1983 Locke E A The nature and causes of Job satlsfactlon In Handbook of Indusrrlal and Orgamzanonal Psychol- ogy (Edited by Dunnette M ), pp 1297-1350 Rand McNally, Chlcago, 111 , 1976 Rice R W , McFarhn D B , Hunt R G and Near J P OrganizatIonal work and the perceived quahty of hfe toward a conceptual model kcad Mgm; Reo- 10, 296-310, 1985 Dlener E SubJective well-being Psycho1 Bull 95, 542-575. 1984 House J S , Strecher V , Metzner H L and Robbms C A Occupational stress and health among men and women m the Tecumseh Commumty Health Study J Hlth sot Behav 27, 62-77, 1986

Posmve psychologlcal states and health 1457

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

SSM

Jenkins C D . Stanton B , Klein M D , Savageau J A and Dwight E H Correlates of angma pectorls among men awaiting coronary by-pass surgery Psvchosom Med 45, 141-153, 1983 Karasek R A, Russell R S and Theorell T Physlo- logy of stress and regeneration in Job related cardlo- vascular dlness J Human Sfress 8, 2942, 1982 Brown M S and Goldstem J L Receptor-medlated control of cholesterol metabolism study of human mutants has disclosed how cells regulate a substance that IS both vital and lethal Srrence 191, 150-154, 1976 Robmson V M Humor and the Health Professtons Slack, Thorofare, N J , 1977 Levine J Humor as a form of therapy mtroductlon to symposium In If’s a Funny Thtng, Humour (Edited by Chapman A J and Foot H C ) Pergamon, Oxford, 1977 Schemer M F and Carver C S Optlmlsm, coping, and health assessment and imphcations of generalized outcome expectancies Hlth Psycho1 4, 219-247. 1985 Cohen S and Edwards J R Personahty charactenstlcs as moderators of the relatlonshlp between stress and disorder In Advances tn the Investrgatton of Psycho- logtcal Stress (Edited by Neufeld W J ) Wiley, New York, 1988 Coser R L Some social functions of laughter Human Relat 12, 171-182. 1959 Kaplan H and Boyd I H The social functions of humor on an open psychlatnc ward Psvchtat Q 39, 502-515. 1965 Turner R G Self-monitoring and humor production J Person 48, 163-172, 1980 Johnson J and Sarason I Life stress, depresslon. and anxiety mternal+xternal control as a moderator vanable J pswhosom Res 22, 205-208, 1978 Lefcourt H M and Martm R A Humor and Ltfe Sfress Anttdote to Adt~ersrt~ Springer, New York, 1986 Dixon N F Humor a cogmtlve alternatlve to stress? In Stress and Anxtetv (Edlted by Sarason I G and Spielberger C D ), Voi 7, pp 281-289 Hemisphere. Washmgion. D C , 1980 Porter L W and Lawler E E Managertat Attrtudes and Performance Dorsey. Homewoodr Ill, 1968 Naylor J C , Pritchard R D and Iieen D R A Theor\ of Behat$ror tn Organtzatrons -Academic Press, New York, 1980 Menaghan E G Indlvldual coping efforts moderators of the relatronshlp between hfe stress and mental health outcomes In Psychologtcal Stress Trends tn Theor, and Research (Edlted by Kaplan H B ), pp 157-191 Academic Press, New York, 1983 Moody R Laugh After Laugh The Healrng Power of Humor Headwaters Press. Jacksonville, Fla. 1978 Berlyne D E Laughter, humor, and play In Hand- book of Soctal Psvcholoav (Edited bv Lmdzev G and Aronsbn E ), Vol 3, pp ?95-852 Addlson:Wesley. Reading, Mass, 1969 Suls J M Cogmtlve processes m humor appreciation In Handbook of Humor Research (Edlted by McGhee P E and Goldstem J H ), pp 39-58 Springer. New York, 1983 Frenkel-Brunswik E Intolerance of amblgulty as an emotional and perceptual personahty varlable J Per- son 18, 108-143, 1949 Harnson R V The person+nvlronment fit model and the study of Job stress In Human Stress and Cognrtton tn Organtzattons (EdIted by Beehr T A and Bhagat R S ). pp 23-55 Wiley, New York, 1985 McGhee P E The role of arousal and hemlsphenc laterahzatlon in humor In Handbook of Humor Re-

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69 70

71

search (Edited by McGhee P E and Goldstem J H ), pp 13-37 Sprmger, New York, 1983 Robinson V M Humor and health In Handbook of Humor Research (Edited by McGhee P E and Gold- stem J H ), pp 109-I 12 Sprmger, New York, 1983 Spencer H The physiology of laughter Macmtllan’s Mag 1, 395-402, 1860 Martin L Psychology of aesthetics expenmental pros- pecting in the field of the comic Am J Psvchol 16, 35-116, 1905 Jones J M and Harm P E Psychophyslologlcal correlates of cartoon humor appreciation Proc A Convent Psycho1 Ass 6, 381-382, 1971 Langevm R and Day H 1 Physlolopscal correlates of hum& In The Psychology of &mor<Edlted by Gold- stem J H and McGhee P E ). DD 129-142 Academic

-. Press, New York, 1972 Fry W F Jr The appeasement function of mirthful laughter In It’s a Funny Thtg, Humour (Edited by Chapman A J and Foot H C ), pp 23-26 Pergamon, Oxford, 1977 Fry W F Jr The respiratory components of mirthful laughter J btol Psycho1 19, 39-50, 1977 Fry W F Jr Humor and the human cardiovascular system In The Study of Humor (Edlted by Mmdess H and Turek J ) Antloch Umverslty, Los Angeles, Cahf, 1979 Goldstein J H , Harman J , McGhee P E and Karasek R Test of an mformatlon processmg model of humor physlologlcal response changes dunng problem- and nddle-solvmg J Gen Psycho1 92, 59-68, 1975 Chapman A J Social aspects of humorous laughter In Humor and Luughter Theory, Research, and At&t- cattons (Edlted by~chapman k J and Foot H-e), pp 117-138 Wiley, London. 1976 Godkewltsch M fihyslologlcal and verbal indices of arousal m rated humor In Humor and Laughter Theory, Research, and Apphcattons (Edited by Chapman A J and Foot H C ), pp 117-138 Wiley, London, 1976 Bushnell D D and Scheff T J The cathartic effects of laughter on audiences In The Study of Humor (Edited by Mmdess H and Turek J) Antloch University, Los Angeles, Cahf, 1979 Sternbach R A Assessmg dlfferentlal autonomic patterns in emotions J psychosom Res 6, 87-91, 1962 Levi L The unnary output of adrenalin and nor- adrenalin durmg pleasant and unpleasant emotional states Psychosom Med 27, 80-85, 1965 Avenll J R Autonomic response patterns dunng sadness and mirth PsychophysroIogy 5.399-414, 1969 Schwartz G E, Wemberger D A and Smger J A Cardiovascular dlfferentlatlon of happmess, sadness, anger, and fear followmg imagery and exercise Psy- chosom Med 43, 343-364, 1981 Schachter S and Smger J Cognmve, social, and physlologlcal determmants of emotlonal state Psy- chol Rev 69, 379-399, 1962 Schachter S and Wheeler L Epmephrme, chlor- promazme and amusement J Abnbrm- Sot Psycho1 65, 121-128, 1962 Sales S M Orgamzatlonal role as a risk factor m coronary disease Admtn Set Q 14, 325-336, 1969 Edwards J R Job satisfaction and Job stress as competing predictors of Job outcomes Working Paper No 14-85-86, Graduate School of Industnal Admmlstratlon, Carnegie-Mellon Umversity, 1984 Freud S Humor Inr J Psvchoanal 9. Id. 1928 Berlyne D E Humor and I& km In The Psychology of Humor (Edited by Goldstem J H and McGhee P E ), pp 43-60 Academic Press, New York, 1972 Tholts P A Dlmenslons of hfe events that influence

L861 ‘N-62 I ‘1 ssJJW )I’oM sanssi leaI%olopoqlaru pun lenjda3uoj ~Jom pue aJ’I UI aseaslp 31UOJq3 PUE ssaJls 31uoJq3 s r asnoH

2861 ‘SZ-IZ ‘ZZ sJJ”J’JS J’U &Cl 1E aseasrp daaq qlJuu uE3 h?p e q%nel v H f utalsp[oD

0861 “lJoA MEN ‘mnuald EZZ-fez dd ‘( H uwn pue s aulha? 1(q paups) yl/vb~ pun Surdo3 u[ ssaJ]s aJq 30 wadse lE3do~olqoq3~sd n JasnaequaqueJd

SL61 ‘YJOA MaN ‘UaAE~ lg[-fpl dd ‘( 1 I,taT ,iq pallpa) wJ~&?msv~~ puv

S.lJlJlUV.iVd AIJlfl SUOllOUJ~ III UOllEJ%alUI aIIUaOpUa

JO suJawd III pa]JayaJ se UOI~OUKJ M 1 uose~ 5861 ‘sfz-61z ‘86 vw iw~sd ~00~ 30 awmis

lensuasuo3 e p~rrmol v uaSallal pue <7 UOSlEM P861 ‘LIII-SO11

‘LP lOI/JASd JOS UOS.tJd f i3aue ahlle%au pub? a,tursod JO aaUapUadapU1 aqL V 8 SUOLUIU~ PW 3 JJUaIa

OL6[ ‘TJOA MaN ‘[[IH-MeJf>aH SSJUJWJ&

PUV ‘JJUVU‘~O~~Jd ‘lOlClVl,J&, IVllJ&UV~ 3 )I YJ’a,,,

pue 3 3 J++-Jl ‘ a N allauuna ‘ d f Ilaqdure3

9L61 ‘PLt49P ‘PC IO’@d 305 UosJJd f U~~lae~al~l uoIlEnlls-uowad 30 SIS@?%e aql

01 q3EoJddlr uolldIJ3sap aSUOdSaJ-aaJJ V V 7 UlAJad

f861 ‘521-66 ‘El IOYJ”Sd ~0s /ddv f SsaJls a%uEqa a~11 JO sJaynq se slJoddns

p.?laos pue sluar\a ahIlIsod H Ut?UIJ~OH pua s uaqo3

1861 ‘ZIOI-ZOO1 ‘IV /OYJksd JOS UosJJd f SSaJlSIp

pue uoIlae3sIlEs qlIM sdIqsuoIlElaJ alllos uoIlEsnes

IEuosJad PUE SluaAa aJI1 V eJlnt?Z PUE ,,, r qaIax

P861 ‘ I’ N ‘WPWH ‘mneqlJ3 LIZ-fsz dd ‘t IOA ‘( 3 s JOWL pus a I lasurs * v mnea 1(q palrpg) I~IIV~H puv kSo~oya& Jo yOOqpUVH

UI SISQEW IEaIlaJoaql e slsaqloddq SurJaynq aql pue ‘ssaJis ‘1Joddns pmos E) deyaw pue s uaqo3

5861 ‘LPZ-6lZ ‘p IOI@Sd yll~ saIauelaadxa auroalno pazqEJaua8 30 suoi~eaqdmr pue luautssasse qlleaq pue ‘%urdo:, ‘rusiuwdo s 3 JattJEa pue J R Jaraqgg

5861 ‘ZLOI-6901 ‘OZ. pan 13s 30s i(lIUnIUUJ1 pue adoq

uo slaaya ~11 pue ddEJaq1 )ueSSaJdap-IluE uo 1JOdaJ

dJEurwI~aJd v a H ueuqapn pue 7 a ueuqapn

5861 ‘P81-9Ll ‘Lf sJY U’O=“tJ~Sd lJyfoIfJ&$ r(lIIlIlUIUII PUE SSaJlS 1 a IIWlJlapn

6961 ‘ III ‘OsEaIqD ‘h~NW PUEH lUJlUJ.illJ~ PUV ‘l’OM Ut UOllJV$.llVS /O lUJU,JlnSVJ~

JXL 7 3 W’H Pue k’i 1 IIEPUW ‘ 3 d qlUW IL61 ‘fL8-198 ‘ff W

uoJy3 f aSEaSIp iJEaq LJEUOJ~~ UI ~olaE3 qsu alqrssod

e SE UOIlaE3SllESSIp qOr f aSIIOH PUE m S Sa[I?S

1961 ‘96Z-68Z ‘tTZ PJW “‘OSOYJ~Sd luaw -UOJIAUJ [eraos aql pw ‘ssaulrr leluaru ‘ssaullr lewl(qd JO UOIlElaJ aql JO SUOIlI?AJasqO [EaI%O[oa~ ‘3 1 al?IIIH

OL61 ‘ZPS-EZS ‘ZE PJH ~OsWAsd uapahts u1 aseaslp iJEaq i(JEu0~0a SJ~~~EJ leaosoqsisa

JO bpnlS UlMl [EaIlUapI UV H a aqex PUE 1 SJOJar[lT L6

2861 “lJOA MaN ‘SSaJd

PII

El1

511

III

011

601

801

LO1

901

so1

PO1

EOI

501

101

001

66

86

aaJJ s9f~_sw dd ‘( s zlruza~g pue 7 Ja8Jaqplof)

iq palIp3) SiJJdsv ]vxul~~ puv ]VJIJJ.IOJI~~ ssalrs

_/O yoo9puv~ UI ssol qOr 8uIXIauadxa uatu %IOIUE

Slaa&a SSaJlS 30 6111Iq”IJEA S qqO3 PUE 12 S [SEX 96

PL61 ‘[f~-LZf’s PJw UrOSOl@d UOIlaJE3UI [EIpJEa -Oh 30 kpnlS (EaI%oloIurapIda alIIaUEUIJa&JaSIE~ aql u10~3 sl[nsaJ UOIlaJ%!3UI lElPJ~30iCIU 30 aAl

-aIpaJd aJlEUUOIlSanb Ia%?oloqz&d v s K qnEla%als pue -l V hsWX ’ X H h-l ‘ Cl 9 UEUIPJIJJ 56

ZL6l ‘LPI-6fI ‘SZ W7 uvl’t3 f UO'PJW" [ElpJEaOhI qlIM SlaarqnS qSIPaMS JO SUOllX3SIlES

23’1 PUE l=Jo’,‘=qafl H U Jq’Ql PUE _L IlaJoaqL P6 6961 ‘fg+L ‘op @ lddns uouv[n3~g qieap uappns pue

uo11a~e3u1 IEIPJ~~O~I UI saaJo3 lEIaosoqsbsd s 310~ f6

8961 ‘OP-_b< ‘6Z lSJ%a JVlyAtSd UOIlZJt?3UI ~tJIPJE~O,kU

LUOJJ qleap %uIpaaaJd _UIEJP leuorloura,, 30 aauapr~x

v srssald np puE 3 x i(pe~aw ‘ f) r uqnJfl

5861 ‘SSSI-ISSI ‘ZIf pJfV f ]sUg Ma&’ aSEaSIp lIJI?U%I~EIU PaaUEApE UI [EAIAJnS 30 SJlE[JJJOa [k?IaOSOqal(Sd 3 JaIlIN p;B

7 7 UMOJ~ ‘ s a ~awi ’ r 3 vn7 * a a qwvsw $861 ‘LPf_PZE ‘x dSJJOld SlJJa UVWlH “Vt,J~ 810

a%IEqa pIIt2 IEMEJpqlIM ‘qlp.?aq aal(oldma JO SIS@!uE ue ~JOM 01 uotwdepy 0 3 urlnH pue f) f assoa

5861 ‘Z8Z-29Z ‘L YllH UnwU’o3 f ‘-Waq

leaIsAqd pue uo11ae3sI1es YJOM M 7 a%%nJqJaA

6L61 ‘PSZI-6PZI ‘IOE PJW f l%l “JN ql[eaq [EaIsiqd uo qlleaq [EluauI JO slaaya ql[eaq 318oIoqzkd a[eLu 30 6JolsIq IeJnleN 3 f) luell~e~

696 1 ‘801-EOI ‘6 fSl~O/OlUOlJ~ Jl/J blIAJ%UO( %uilXpaJd UI SJOlaEJ [WaOS PW [ElUaLU ‘[eaIS/(qd 3 aJOU‘[Ed

586 I ‘UoPuo-l ‘Aa[I,Q UlV.US PUV SSJllS qOf /O SWlUVlf3Jfl JqL

i\ a UOSUJI?H PUE Q x ue[dEa ‘Jr d x 1 qaUaJJ

6L6i ‘091-6f I ‘OZ ‘VYJS JOS L/I/H f SJaqJOM I(JOlaE3 %lOUIE ql[Eaq PUE ssaJls [Euoi~edtwo a 7 UEIUJ~~UE~ pue H

a ueidex * v r s11a~ ‘ r v laEqaIwly ‘ s r aSnoH

1861 ‘ZPL-SEL ‘pz SJllUOUOh~ IJOM [EIJlSIIpUI lq%Il UI alI?J 1JEaq

PUE ‘lJO#J pahlaalad ‘~01laEJS1lES qO[ V anbalEqx

LL61 ‘6ff-Pff ‘I 3 UJWOM /OlfJ’Sd UaUIOM alIqM PUE TaElq PalJJElU

%uouIe aJnssaJd poo1q PUE ‘UOllaE3SlleS qJ0~ ‘snlels

‘IJOM 3 %JnqJI?H PW A S ISE)I ‘S 7 UIalSUaIIEH

tL61 ‘ELE-ILE ‘6s l“YJ~~Sd lddv f s]uatuuoJiAua [EA~U UI uo1lae3s1l~s qor 30

SalE[aJJOD x 3 UOSJapUnf) PUI? M 8 p[EUOaaN

EL61 ‘f0958S ‘L PJN ‘3 3oS qaIJnz le ko13e3 e JO uorle[ndod alzur aqL 11 I(pnls

aAIlEJadOOa [‘C?UOIlEUJalUl U‘\I’ ql[k?aq JEl”3sEAOIpJEa

PUE SSaJlS w r Ua?JIa PUE 0 1 Japaaa ‘ H JEqT

L861 ‘IZI-SII ‘6s ~11~ uoliou~ dnmg ywy rul XJOM ahriiladal paced-3Ias UI qlleaq pue ‘~cw?IUJO~J~~ ‘SSaJlS uo

u0113e3~11es q0r30 my3 )I v uas pue m ueruqea 9861 ‘EZE-SIf ‘L “vYJt7 dnmo I

%Illaq-[IaM pue sasuanJSuo3 ]uamuoJli\ua-uos1ad 30 uoIiEInrun~z aql s pawelaN pue I 3 Ja~m

9861 ‘Lf8FN8 ‘Pz J~KJ PJW SlsIUJaluI %UIaIlaEJd %UOUIE UOIlaE3SllES J311 qllM PalWaOSSE

SJOlaEd 8 aqeq PUE ,+, a adO3 ‘ 1 Ja%A ‘ S 1 UU1-I

f861 ‘6lISJaAlIIn alElfj El%JOag

‘uow3npg 30 a%allo3 ‘uo~lel~ass~p pzrolaop paqsqqnd

-un ssauly leaIsbqd JO [aAa[ pue ql[eaq pzaIsKqd

JO UOIlaUIlJ E SE UOIlaIlJSIlI?S qO[ S w %UIUUE~

8L6l ‘t9Z-8PZ ‘IZ I- lu%4 pVJV U0113~3SIlES qOr PUE a311 30 SalE[aJJOD qJ0.WEJlXa

PUE ‘1JoM E) a 1unH pue M a XIX ‘d r JeaN

ELfj[ ‘blISJaAIUn alElS UaaJf) %UllMOa

‘UOIlElJaSSIp [EJOlaOp paqsqqndun UOIlE%IlSaAUl

alI?lJEAll~nUl 8 ql[eaq JO SalElS PaAIaaJad PUE UOlla’c?3

411~s q0r uaawaq dIqsuoIlE~aJ aqL a r qneJluIaM

8561 ‘S6Z-8LZ ‘OZ PJW ~OSV~Sd SSaUllI 01 blqIqrlda3sns lEJaUa% PUE ‘SaIlSlJalaEJk?qa

h[euosJad ‘sacwauadxa 931[ UaaMlaq UOllE[aJ aqi 30 UOIlr?%IlSJAUI UV t) H B[OM PUR N ,Q PJOJlJql ’ V piagso ‘ a J auex ( N M uosuaisiJq3 ‘ 3 7 al7urH

f 861 ‘9Sb-SW7 ‘LZ say wosoyJnid [ uazu %UnOr( UI aJnSSaJd poo[q palEAala pUE SlUaAa JJIT J_ IlaJOaqL pUE r UOSSUaAs

LL6l ‘zz+SIp ‘Iz SJU wosoymd f SIS/([EUE aAIlEJEdUlOa

E SUIalSk %Ullq~IaM luaha a311 v a e%oqlJIq~

Egf,I ‘IJOA MaN ‘SSaJd aIUlapEaV

EOI-EE dd ‘( 8 H UEldEx Aq palIps) ~/lJvJsJu Puv tdoay~ UI S~U~JL ssarls p~So~otp&d us aJnleJalq aql

30 SIsaqluk pue uollerqeha ue ssa~lsrp le3l%oloqzisd

56

I6

06

68

88

L8

98

S8

P8

f8

Z8

I8

08

6L

8L

LL

9L

S-L

PL

EL

ZL


115 Kessler R C The Interplay of research design strate- gles and data analysis procedures m evaluatmg the effects of stress on health In Stress and Health Issues m Research Methodology (Edlted by Kasl S V and Cooper C L ), pp 113-140 Wiley, New York, 1987

116 Cronbach L J and Furby L How should we measure “change’‘-or should we? Psycho1 Bull 74, 68-80, 1970

117 Johns G Difference score measures of orgamzatlonal behavior variables a critique Org BehaL Human Perform 27, 443463, 1981

118 Wall T D and Payne R Are deficiency scores deficient? J appl Psycho1 58, 322-326, 1973

119 Dwyer J H Starutlcal Models for Ihe Social and Behacroral Sciences Oxford Umverslty Press, New York, 1983

120 Joreskog K G and Sorbom D LISREL VI National EducatIonal Resources, Chlcago. Ill, 1986

121 Long J S Coaarrance Structure Models An Inrro- ductron to LISREL Sage, Beverly Hills. Cahf. 1983

the impacts of positive psychological states on...

Documents