Age Differences in Stress, Coping, and Appraisal: Findings From the

78254_351B_4_P179.pdf

Journal of Gerontology: PSYCHOLOGICAL SCIENCES

1996
, Vol. 5IB. No. 4, PI79-P188Copyright 1996 by The Gerontological Society of America Ag e Difference s i n Stress , Coping , an d Appraisal :

Finding

s Fro m th e Normativ e Agin g Stud y

Caroly

n M . Aldwin, 1 Kare n J . Sutton, 1 Gin a Chiara, 2 an d Avro n Spir o II P 'Departmen t o f Huma n an d Communit y Development , Universit y o f California , Davis . 2

Normativ

e Agin g Study , Departmen t o f Veteran s Affair s Outpatien t Clinic , Boston .

Controversies

exist concerning the influence of age on the stress and coping process, in part due to differences in methods across studies. We examined age differences in stress, appraisal, and coping, using both semistructured interview questions and a coping checklist in middle-aged, young-old, and old-old men. Despite extensive probing, nearly a quarter of the old-old reported having had no problems and they expended less coping effort even when they did have problems. The types of problems reported varied systematically with age. Middle-aged men were more likely to appraise their problems both as challenges and as annoyances than the older men. Different age patterns emerged from the coping interviews vs the checklists, but controlling for type of problem significantly attenuated age differences. However, there were no age differences in perceived stressfulness of the problem, appraisals of harm/loss, or helpless appraisals, number of emotions reported, or coping efficacy. One interpretation of these results is that the nature of stress changes with age, from episodic to chronic, which in turn affects appraisal and coping processes.T H E manne r i n whic h ag e influence s th e stres s an d copin g proces s remain s poorl y understoo d (Folkman , Lazarus ,

Pimley

, & Novacek , 1987)
. I n general , debat e focuse s o n whethe r youn g an d ol d adult s diffe r i n th e way s i n whic h the y experienc e an d cop e wit h stress . I n part , thi s ma y b e du e t o th e fac t tha t ther e ar e differen t component s o f th e stres s an d copin g proces s a s wel l a s differen t way s o f assessin g thes e components . Fo r example , olde r peopl e ma y (o r ma y not ) diffe r fro m younge r peopl e i n eithe r th e amoun t o f stres s the y report , th e type s o f stressor s reported , o r th e appraisa l o f ho w stressfu l a proble m is , dependin g upo n th e typ e o f measur e used . Further , th e us e o f copin g strategie s ha s variousl y bee n viewe d a s decreasin g wit h age , increas - in g wit h age , o r showin g littl e o r n o age-relate d differences , again , dependin g o n th e conceptua l framework s an d meth - odologica l technique s use d (Aldwin , 1991)
. Th e purpos e o f thi s stud y wa s t o examin e ag e difference s i n th e stres s an d copin g process , contrastin g in-dept h interview s wit h mor e standardize d measure s i n par t t o determin e th e exten t o f age - relate d bia s i n self-repor t instruments . Age and the Stress Process Lat e lif e i s ofte n see n a s a tim e o f great , ofte n uncontrolla -ble, stress (Rodin, 1986). Older people are often coping with chroni c illnes s an d disability , th e los s o f friend s an d famil y members , an d thei r ow n impendin g mortality . However , Payke l (1983 ) foun d tha t th e elderl y repor t fewe r stressfu l lif e events tha n d o th e young . Aldwi n (1990 ) argue d tha t mos t lif e even t inventorie s sampl e problem s whic h ar e mor e relevan t t o younge r ag e groups , suc h a s marriage , divorce , startin g ne w jobs , o r havin g children . Inventorie s whic h includ e item s mor e relevan t t o olde r adults , suc h a s retire - men t o r divorce o f children , ten d t o sho w fe w ag e differ - ence s i n th e amoun t o f stres s reported . Thus , olde r peopl e ma y repor t th e sam e numbe r o f lif e events , althoug h th e typ e o f even t ma y var y accordin g t o lif e stag e (se e Murrell ,

Norris

, & Hutchins , 1984
; Ter i & Lewinsohn , 1982)
.The types of hassles, or daily stressors, experienced also var y a s a functio n o f ag e (Lazaru s & DeLongis , 1983)
. Olde r peopl e repor t fewe r hassle s an d rat e thes e hassle s a s les s stressfu l tha n d o younge r adult s (Aldwin , 1990
; Folkma n e tal., 1987). Life events may also be appraised as less stressful i n late r lif e (Aldwin , 1991
; Silverman , Eichler , & Williams , 1987)
.

Hassle

s ar e ofte n a functio n o f involvemen t i n socia l role s (Lazarus , 1991
; Pearlin , 1989)
. Thus , th e decreas e i n th e numbe r o f role s ma y resul t i n fewe r hassle s i n lat e life . Fo r example , relinquishin g activ e parentin g an d wor k role s ma y wel l decreas e th e number s o f dail y stressor s on e experi - ences . However, why older adults appraise both life events an d hassle s a s les s stressfu l i s no t clear . A possibl e reaso n fo r thi s discrepanc y ma y b e a bia s i n th e reportin g o f stress , jus t a s ther e i s a bia s i n th e willingnes s o f olde r adult s t o repor t psychologica l symptom s (Cohen , 1990)
. I n addition , th e elderl y ofte n appea r t o repor t fewe r negativ e emotion s tha n younge r individual s (Labouvie-Vief , Hakim-Larson , & Ho - bart , 1987
; Lawton , Kleban , & Dean , 1993)
. Bot h questionnaire s an d interview s rel y o n th e individ - ual' s accurat e acces s t o episodi c memor y fo r th e event , includin g it s occurrenc e an d degre e o f stressfulness . Ther e i s som e concer n tha t olde r individuals , wit h thei r highe r inci - denc e o f memor y problems , ma y provid e les s accurat e information , particularl y o n questionnaires . However ,

Rodger

s an d Herzo g (1987 ) foun d tha t olde r peopl e provid e informatio n tha t i s jus t a s accurat e (i f no t mor e so ) tha n younge r individual s o n surve y measures . Nonetheless , inter - vie w technique s ma y provid e a mor e thoroug h an d accurat e accoun t o f stres s (Brown , 1989
) an d thu s ma y provid e th e opportunit y fo r mor e accurat e conclusion s abou t th e rol e tha t ag e play s i n stres s reportin g an d appraisal . Thi s ma y b e especiall y relevan t fo r th e elderly . Caserta , Lund , an d Dimon d (1985 ) randoml y assigne d bereave d elderl y subject s t o group s tha t wer e eithe r interviewe d o r aske d t o complet e a maile d questionnair e whic h measure d P17

9Downloaded from https://academic.oup.com/psychsocgerontology/article/51B/4/P179/650788 by guest on 25 August 2023

P180ALDWINETAL.

depressio n an d lif e satisfaction . Althoug h n o effec t o f grou p assignmen t emerged , thos e wh o wer e interviewe d wer e consistentl y mor e likel y t o provid e complet e information . Thus , interview techniques may yield more accurate infor- matio n i f ther e ar e age-relate d biase s i n th e reportin g o f stress . Ther e ma y als o b e developmenta l reason s wh y olde r adult s apprais e problem s a s les s stressful . Olde r people , throug h thei r greate r rang e o f experience , ma y hav e devel - ope d mor e copin g resource s an d thu s apprais e problem s a s les s stressful . Further , t o th e exten t tha t olde r adult s hav e experience d extremel y stressfu l events , suc h a s th e deat h o f love d ones , everyda y problem s ma y pal e i n compariso n (Aldwin , 1994a
; Aldwin , Levenson , & Spiro , 1994)
. How - ever , th e decreas e i n reserv e capacit y i n ver y lat e lif e (Baltes , 1987
) ma y rende r problem s mor e stressfu l fo r th e old-old . T o ou r knowledge , n o on e ha s compare d stres s betwee n th e young-ol d an d old-old .

Beside

s degre e o f stressfulness , anothe r aspec t o f ap - praisa l concern s th e genera l typ e o f proble m - tha t is , a situatio n ca n b e appraise d a s a threat , a s involvin g har m o r loss , as a challenge, or as benign (Lazarus & Folkman, 1984)
. Given that older people are often coping with be- reavemen t an d chroni c healt h problems , i t i s ofte n assume d tha t the y wil l b e mor e likel y t o apprais e problem s a s involv - in g har m o r los s (Folkma n & Lazarus , 1980
; McCrae , 1982
, 1989
; Rodin, 1986), but, surprisingly, this has not been teste d empirically . Age and the Coping Process A n alternativ e explanatio n fo r ag e difference s i n th e amoun t o f reporte d stres s ma y b e tha t olde r individual s cop e i n a differen t wa y tha n thei r younge r counterparts . Variou s theorist s hav e posite d changes i n copin g wit h age . Fo r instance , Gutman n (1974 ) suggeste d tha t master y style s shif t fro m activ e t o passiv e fro m yout h t o midlife , the n t o "magical " master y i n lat e life . I n contrast , Vaillan t (1977 ) believe d tha t a positiv e chang e occurre d i n th e us e o f defens e mechanisms . H e propose d tha t a n incrementa l developmen - ta l proces s occur s acros s th e lif e span , characterize d b y a decreas e i n th e us e o f neuroti c o r immatur e defensiv e style s couple d wit h a n increas e i n mor e matur e defensiv e style s i n midlife . Alternatively , other s (Folkma n e t al. , 1987
; Mc - Crae , 1982, 1989) have suggested an absence of systematic chang e i n copin g strategie s wit h ag e excep t thos e du e t o situationa l demands .

Empirica

l evidenc e fo r thes e position s appear s t o b e mixed. McCra e (1982 ) foun d tha t olde r adult s use d fewe r escapis t an d hostil e strategie s whe n copin g wit h problems . Yet , fe w difference s i n problem-focuse d strategie s emerge d onc e h e controlle d fo r th e typ e o f problem . Indeed , mos t studie s whic h examin e ag e difference s i n copin g focu s o n specifi c stressfu l episodes , du e t o th e stron g evidenc e o f situationa l effect s o n individua l coping .

Studie

s usin g th e Way s o f Copin g Scal e (Folkma n &

Lazarus

, 1980
, 1985
) an d it s derivative s hav e generall y confirme d McCrae' s (1982 ) findings. Indeed , severa l re - searcher s hav e show n tha t olde r adult s us e les s escapis m o r avoidan t copin g bu t a simila r o r highe r leve l o f problem -focused coping as do younger adults (Blanchard-Fields,

Sulsky

, & Robinson-Whelen , 1991
; Felto n & Revenson , 1987
; Irion & Blanchard-Fields, 1987). An exception to this tren d i s a stud y b y Folkma n e t al . (1987 ) whic h foun d tha t olde r peopl e use d les s planfu l proble m solvin g an d mor e escap e avoidance . However , i n thei r stud y the y examine d th e relativ e us e o f copin g strategies , o r th e rati o o f thes e strategie s t o overal l strategie s used , althoug h Aldwi n (1991 ) wa s unabl e t o replicat e ag e difference s i n thes e copin g ratios . Finally, regardless of the overall pattern of strategies used , olde r individual s d o see m t o us e fewe r strategies , whil e remainin g a s effectiv e a s th e youn g i n thei r abilit y t o cop e (Aldwin , 1991
; Meeks , Carstensen , Tamsky , Wright , & Pellegrini , 1989)
.

Contradictor

y findings ma y als o b e du e t o th e limitation s o f th e measure s bein g used . Fo r instance , standar d copin g inventorie s ma y oversampl e maladaptiv e strategie s whil e undersamplin g adaptiv e ones , a s indicate d b y th e genera l tendenc y o f suc h scale s t o correlat e positivel y wit h poo r outcomes suc h a s psychologica l o r physica l symptom s (Al - dwi n & Revenson , 1987
) rathe r tha n mor e benig n outcome s suc h a s master y o r positiv e affec t (Aldwin , 1994b)
. Further , studie s whic h hav e use d standar d checklist s ma y no t b e sensitiv e t o aspect s o f th e copin g proces s whic h chang e ove r time , especially positive or adaptive coping which may increas e wit h maturit y (Labouvie-Vie f e t al. , 1987)
. Thes e inventorie s hav e no t bee n designe d fo r th e examinatio n o f developmenta l difference s i n copin g and , a s a result , ma y no t b e tappin g appropriat e strategies . Usin g qualitativ e techniques , som e researcher s hav e iden - tifie d copin g processe s tha t appea r t o change . Vaillan t (1977 , 1993) documented age-related increases in mature defense s i n thre e longitudina l samples : th e Gran t men , th e Terma n me n an d women , an d a sampl e o f poo r inner-cit y resident s onc e a t ris k fo r juvenil e delinquency . Thes e find- ing s wer e partiall y confirme d usin g cross-sectiona l dat a fro m a self-repor t inventor y o f defens e mechanism s (Diehl , Coyle , & Labouvie-Vief , i n press) . I t i s als o possibl e tha t ther e ar e qualitativ e (a s oppose d t o quantitative ) difference s i n th e us e o f copin g strategies . Usin g a mixtur e o f qualitativ e an d quantitativ e data ,

Labouvie-Vie

f etal . (1987 ) foun d tha t olde r individual s use d copin g strategie s simila r t o thos e o f younge r individuals , bu t tha t the y di d s o fo r differen t reasons . Fo r example , whil e bot h youn g an d ol d subject s sough t socia l suppor t i n copin g wit h problems , younge r subject s di d s o primaril y fo r self - validation , whil e olde r individual s sough t socia l suppor t a s a wa y t o obtai n feedbac k abou t th e appropriatenes s o f th e strategie s the y wer e using . I n general , qualitativ e procedure s ar e mor e likel y t o revea l age-relate d chang e tha n ar e quanti - tativ e methods . I n summary , i t i s clea r tha t ther e ar e age-relate d change s i n socia l role s an d healt h statu s whic h affec t th e number s an d type s o f stressor s experienced , whic h i n tur n influenc e th e way s i n whic h peopl e ma y cop e wit h stress . However , wh y th e elderl y repor t les s stres s i n th e fac e o f increasin g disabili - tie s an d los s o f famil y an d friend s i s no t clea r an d ma y reflec t reportin g biase s and/o r developmenta l processes . Further , ther e ar e difference s betwee n quantitativ e an d qualitativ e studie s concernin g whethe r ther e ar e direc t age-relate d in

-Downloaded from https://academic.oup.com/psychsocgerontology/article/51B/4/P179/650788 by guest on 25 August 2023

STRESS, COPING, AND APPRAISALP181

fluences o n copin g processes , i.e. , whethe r o r no t mor e matur e strategie s emerg e wit h age .

Present

Study I n thi s study , ag e difference s i n severa l aspect s o f th e stres s an d copin g proces s wer e examined , usin g bot h quali - tativ e an d quantitativ e dat a fro m olde r men , rangin g fro m th e middle-age d t o th e old-old . W e aske d fo r th e mos t stressfu l proble m i n th e prio r wee k an d ho w the y cope d wit h thi s specifi c problem , usin g bot h open-ende d an d close-ende d formats , t o examin e a numbe r o f questions . (1 ) Give n prio r wor k (Aldwin , 1991)
, i t i s likel y tha t w e wil l se e age-relate d difference s i n th e type s o f problem s experienced , wit h work - an d parenting-relate d stressor s re - porte d mor e b y thos e i n mid-lif e an d health-relate d stressor s reporte d mor e frequentl y i n lat e life . (2 ) T o th e exten t tha t th e ag e difference s i n stres s report - in g an d appraisa l ar e a functio n o f respons e bias , w e hypoth - esiz e tha t th e us e o f interview s wil l minimiz e thi s age-relate d respons e bia s i n th e reportin g an d appraisa l o f stress , a s wel l a s negativ e emotions . However , i f ther e ar e developmental / maturationa l effect s o n th e appraisa l o f stress , the n ag e difference s shoul d emerg e eve n wit h extensiv e probing . (3 ) W e hypothesiz e tha t th e pattern s o f ag e difference s i n copin g strategie s typicall y foun d wit h self-repor t question - naire s wil l b e attenuate d whe n copin g strategie s ar e elicite d usin g semi-structure d intervie w question s fo r som e prob - lems . In other words, probing should decrease the tendency o f th e elderl y t o repor t fewe r strategies , unles s ther e ar e direc t age-relate d influence s o n th e copin g process . (4 ) I f th e age-related influences on coping are indirect and mediate d b y change s i n th e typ e o f problem s experience d (McCrae , 1982
, 1989)
, w e woul d expec t tha t controllin g fo r type s o f problem s reporte d woul d attenuat e an y age-relate d difference s i n coping . (5 ) Conversely , base d o n wor k b y Vaillan t (1993) , w e hypothesiz e tha t copin g interview s ma y sho w tha t th e elderl y us e mor e adaptiv e strategie s tha n middle-age d individuals . A t th e ver y least , ther e shoul d b e n o difference s i n self-rate d copin g efficac y betwee n th e tw o groups . (6 ) Finally , w e wil l explore stres s an d copin g processe s amon g th e old-old , an d hypothesiz e tha t som e o f th e age - relate d decrement s hypothesize d i n copin g wil l primaril y b e see n i n tha t group . METHO D

Sample

and Procedure Th e Normativ e Agin g Stud y (NAS ) i s a longitudina l stud y o f biomedica l an d psychosocia l factor s i n norma l agin g tha t ha s followe d 2,28 0 me n fo r ove r 3 0 years . Al l me n wer e screene d fo r th e absenc e o f seriou s chroni c disease , a s wel l a s bloo d pressur e greate r tha n 140/9
0 befor e participatin g i n th e study . I n addition , me n wer e selecte d wh o wer e likel y t o b e geographicall y stable , assesse d b y th e exten t o f famil y ties . As such, they reflect the demographics of the Boston populatio n o f male s i n th e lat e 1950
s an d earl y 1960s
, an d ar e primaril y Whit e an d ar e equall y divide d betwee n blue - an d white-colla r workers . (Fo r mor e informatio n o n NA Ssample characteristics, see Bosse, Ekerdt, & Silbert, 1984; Spiro , Aldwin, Ward, & Mroczek, 1995.) Th e presen t sampl e consiste d o f 1,06 5 me n wh o reporte d fo r thei r triennia l physica l examinatio n betwee n th e year s 198
9 an d 1991
. Onl y 1 3 me n refuse d t o participat e i n th e interview , leavin g a sampl e o f 1,05 2 me n an d a respons e rat e o f 98.8%
. The N varies slightly across analyses, due to missin g data . Th e me n range d i n ag e fro m 4 8 t o 9 1 a t th e tim e o f th e interview , wit h a n averag e ag e o f 65.6

3 (SD = 7.68). Since

th e NA S disproportionatel y sample d middle-age d subject s a t th e beginnin g o f th e study , w e divide d th e me n int o fou r ag e groups : earl y mid-lif e (45-54 , n = 76)
; late r mid-lif e (55 - 64
, n = 405); young-old (65-74, n = 438); and old-old (7 5 + , n = 132)
. Thi s als o allowe d u s t o investigat e nonlin - ea r ag e differences .

Measures

Th e me n wer e administere d th e Stres s an d Copin g Inter - vie w o n th e da y o f thei r medica l examination . Th e origina l desig n calle d fo r self-administere d questionnaire s t o b e com - plete d throughou t th e cours e o f th e day . However , pilo t testin g o f th e instrumen t indicate d tha t ove r a quarte r o f th e me n di d no t repor t havin g ha d a proble m i n th e pas t wee k o n th e self-administere d questionnaires . Confirmator y inter - view s wit h thes e me n suggeste d tha t som e wh o di d no t repor t havin g a proble m o n th e questionnair e actuall y ha d ha d problem s i n th e pas t week , whic h wer e ofte n quit e serious . Fo r example , tw o o f th e me n wh o participate d i n pilo t interview s wer e primar y caretaker s fo r dyin g wives . Ther e ha d bee n n o flare-ups i n th e wives ' medica l condition s durin g th e prio r week , s o the y di d no t repor t problem s o n th e questionnaire . Nonetheless , i t wa s clea r fro m th e interview s tha t thi s wa s a ver y stressfu l tim e fo r them . W e als o observe d tha t man y o f th e oldes t me n denie d havin g "problems, " pe r se . On e 84-year-ol
d ma n explaine d tha t h e ha d "concerns " bu t tha t h e wa s no t "th e sor t o f perso n wh o ha d problems ." (H e explaine d that , durin g hi s formativ e year s durin g th e Depression , on e simpl y di d no t admi t t o "havin g problems." ) Hi s "concern " tha t wee k wa s tha t h e wa s tryin g t o convinc e hi s 91-year-ol
d siste r t o ente r a nursin g home . Sh e live d alon e i n a majo r metropoli - ta n are a an d ha d falle n severa l times . H e wa s ver y worrie d abou t he r an d trouble d b y hi s inabilit y t o convinc e he r t o ente r a nursin g home . Thi s "concern " seeme d t o b e ver y stressfu l fo r him . Thus , w e aske d fo r "concerns " a s wel l a s problem s i n ou r interviews . Usin g th e intervie w format , w e aske d participant s t o identif y th e mos t seriou s proble m o r concer n the y ha d ha d i n th e pas t wee k an d t o describ e i t briefly . W e aske d respon - dent s fo r a specifi c proble m i n th e pas t week , fo r severa l reasons . First , w e wer e concerne d abou t memor y problem s an d wante d t o elici t a ver y recen t episod e t o minimiz e this .

Second

, ther e i s a fai r amoun t o f evidenc e tha t copin g varie s greatl y b y situation , an d littl e ca n b e inferre d fro m question s abou t genera l copin g style s (se e Aldwin , 1994a
, fo r a re - view) . Third, most studies which have examined coping in th e elderl y hav e focuse d o n a specifi c problem , an d w e wishe d ou r result s t o b e comparable . I f a responden t di d no t immediatel y respon d wit h a prob

-Downloaded from https://academic.oup.com/psychsocgerontology/article/51B/4/P179/650788 by guest on 25 August 2023

P182ALDWINETAL.

lem , w e assure d the m tha t th e proble m di d no t hav e t o b e "major, " bu t coul d b e anythin g tha t ha d bothere d the m i n th e pas t week . I f thi s di d no t elici t a problem , w e the n sai d tha t the y coul d tal k abou t problem s a t wor k o r a t home , wit h thei r famil y o r thei r health , o r wit h hom e o r ca r mainte - nance . Thi s migh t elici t severa l problems , an d w e woul d as k th e me n t o pic k th e on e tha t the y foun d th e mos t troubling . Th e men' s response s wer e recorde d verbati m b y th e inter - viewer . Respondents wer e give n a 0 i f the y di d no t repor t a problem , an d th e res t o f th e appraisa l an d copin g section s wer e no t completed . Fo r th e proble m selected , respondent s wer e the n aske d t o rat e thi s proble m accordin g t o it s stressfulness , compare d t o othe r problem s the y ma y hav e ha d i n th e past . Th e standar d

7-poin

t scal e wa s used , wit h 1 = "not troubled at all" and 7 = "th e mos t trouble d I'v e eve r been. "

Respondent

s wer e the n presente d a lis t o f appraisal s o r "feeling s abou t th e problem, " an d aske d t o indicat e whic h wer e appropriat e t o th e proble m jus t described . W e ex - pande d o n Folkma n an d Lazarus ' (1980 ) origina l thre e stres s appraisal s o f threat , harm/loss , o r challenge , becaus e pilo t interview s indicate d tha t thes e appraisal s di d no t cove r al l o f th e way s i n whic h th e me n appraise d thei r problems . Th e ne w appraisal s were: a t a los s fo r wha t t o d o nex t (i.e. , helpless) , annoyed , and/o r worrie d abou t others . Further , th e respondent s ofte n fel t strongl y tha t mor e tha n on e ap - praisa l fi t th e problem , s o w e allowe d the m t o indicat e multipl e appraisals . Fo r example , a ma n caregivin g fo r a terminall y il l wif e migh t simultaneousl y fee l threatene d a t th e prospectiv e los s o f hi s wife , worrie d abou t he r comfor t an d well-being , an d challenge d i n hi s caretakin g tasks . Th e mea n numbe r o f appraisal s reporte d wa s 2. 2 (SD = 1.33) . Whe n w e asked , i n th e pilo t interviews , ho w the y ha d cope d wit h th e problem , w e foun d tha t thes e olde r me n ha d a tendenc y t o disregar d ho w the y cope d wit h thei r feelings , instea d tellin g u s ho w the y manage d th e situation . Thus , w e firs t asked , "Ca n yo u tel l u s ho w yo u cope d wit h o r trie d t o manag e thi s problem?, " usin g genera l probe s suc h a s "Di d yo u d o anythin g else? " t o elici t a s man y strategie s a s possible . W e the n aske d specificall y abou t emotion s tha t the y ha d experienced . I f the y ha d mentione d a n emotio n i n th e previou s question , w e prompte d thei r recal l b y th e phrase , "Yo u sai d yo u wer e 'X ' [e.g. , frustrated] . Ca n yo u tel l m e a littl e bi t mor e abou t that? " I f th e me n wer e no t forthcomin g abou t thei r emotions , w e asked , "Wer e yo u angr y o r frustrated , sa d o r a littl e blue? "

Finally

, w e aske d specifically , "Ho w di d yo u tr y t o cop e wit h o r manag e you r feelings?' ' I f th e me n wer e to o vagu e o r general , th e interviewe r woul d tr y t o ge t mor e specifi c information , fo r example , b y mentionin g a n emotio n elicite d b y th e previou s questio n an d askin g ho w the y deal t wit h tha t specifi c emotion . Al l informatio n o n copin g strategie s an d emotion s wa s recorde d verbati m b y th e interviewer . Th e Brie f Way s o f Copin g (BWOC ; Folkma n & Lazarus , 1988
) questionnair e wa s the n administere d orall y regardin g th e sam e problem . Th e BWO C i s a lis t o f 2 5 copin g strate - gie s rate d fro m 0 t o 3 , i n whic h 0 indicate s tha t th e respon - den t di d no t us e thi s strateg y a t al l an d 3 indicate s tha t h e use d i t a lot . Th e scal e wa s derive d b y selectin g th e to p item s fro m th e eigh t factor s o f th e revise d BWO C (se e DeLongis , 1986)
. Unlike the interview questions, the BWOC was not administere d separatel y fo r problem s an d emotions . W e factore d th e item s usin g principa l axi s analysi s wit h bot h varima x an d obliqu e rotations . Th e scre e tes t indicate d tha t ther e wer e betwee n five an d eigh t factors ; w e opte d fo r th e eight-facto r solutio n wit h varima x rotatio n t o enhanc e bot h interpretabilit y an d comparabilit y t o othe r samples . Thi s solutio n accounte d fo r 60.1
% o f th e variance . Sub - scale s wer e compute d b y summin g item s scores , usin g uni t weights , an d include : instrumenta l action , escap e avoid - ance , social support, self-blame, cognitive reframing, interpersonal/hostile , threa t minimization , an d self - isolation . Onl y tw o o f th e item s differe d fro m th e DeLongi s (1986 ) facto r solution : praye r loade d o n escap e avoidanc e rathe r tha n cognitiv e reframing , an d th e ite m "Di d yo u tr y no t t o ac t to o hastily . . . " di d no t loa d o n an y factor . Thi s make s sense , give n th e restricte d natur e o f ou r sampl e (olde r males) . Coping effort scores were created by summing the numbe r o f strategie s reporte d bot h i n th e interview s an d o n th e BWOC .

Finally

, th e participant s wer e aske d t o indicat e ho w effec - tiv e the y believe d the y wer e i n managin g th e problem , usin g a scal e fro m 1 t o 5 , wit h 1 = "no t wel l a t all " an d 5 = "ver y well " (Aldwi n & Revenson , 1987)
, an d als o rate d ho w effectiv e the y though t the y woul d b e wit h a simila r proble m i n th e future .

Content

Analyses

Conten

t analyse s wer e use d t o identif y type s o f problems , copin g strategies , an d emotions . Al l interview s wer e code d b y th e first an d thir d authors , an d agreemen t rate s fo r th e tota l sampl e range d fro m .7 5 t o .97 . Nonetheless , al l differ - ence s wer e reconciled . Te n majo r type s o f problem s wer e identified . Thes e in - clude d healt h problem s (sel f an d others) , marita l relation s an d wife' s problems , childre n an d grandchildren , social , work , retirement , finances, bereavement , an d genera l has - sle s (e.g. , ca r repair , hom e maintenance , gettin g lost , etc.) . I n addition , w e include d a "multipl e problems " category .

Althoug

h w e wer e successfu l i n gettin g th e me n t o focu s o n a particula r episode , som e problem s cu t acros s domain s an d i t wa s occasionall y impossibl e t o disentangl e them . Fo r exam - ple , one respondent's son-in-law was believed to have mo- leste d a grandchild , an d h e wa s simultaneousl y helpin g hi s daughte r cop e wit h th e divorce an d th e molestatio n trial , a s wel l a s helpin g th e grandchildren . W e als o conten t analyze d th e copin g strategie s reporte d fo r bot h th e problem - an d emotion-focuse d questions , iden - tifyin g 11 9 differen t copin g strategie s tha t wer e groupe d int o 1 0 majo r categories . (W e ha d hope d t o identif y strategie s whic h w e considere d mor e developmentall y mature , suc h a s transformationa l copin g - e.g. , th e abilit y t o transfor m negativ e situation s int o positiv e ones . However , spontane - ou s reportin g o f thes e type s o f strategie s wa s s o rar e tha t w e opte d t o us e a groupin g schem e tha t mor e closel y resemble d tha t normall y foun d wit h copin g inventories , whic h woul d als o b e mor e usefu l i n thi s comparativ e study. ) Th e problem-focuse d copin g strategie s include d instru - menta l action , interpersona l action , an d restraine d action .

Instrumenta

l actio n include d strategie s suc h a s "Mad e a pla

nDownloaded from https://academic.oup.com/psychsocgerontology/article/51B/4/P179/650788 by guest on 25 August 2023

STRESS, COPING, AND APPRAISALP183

o f actio n an d followe d it, " whil e interpersona l actio n wa s directe d specificall y a t individual s withi n th e situation , e.g. , "Trie d t o ge t th e othe r perso n t o se e you r poin t o f view. "

Restraine

d actio n include d strategie s suc h as , "Decide d t o defe r actio n o n problem, " an d wa s ofte n use d whil e await - in g result s fro m medica l tests . Th e emotion-focuse d strategie s included : threa t minimi - zation/acceptance , redefinitio n o f th e problem , anxiet y re - duction , distraction , escapism , an d emotiona l expression . I n addition , th e socia l suppor t categor y include d bot h provid - in g an d receivin g socia l suppor t (th e codin g schem e i s availabl e o n reques t fro m th e first author) . Threa t minimiza - tio n include d strategie s suc h a s "Accepte d situatio n o r wa s resigne d t o it" ; anxiet y reductio n include d usin g medica - tions , alcohol, food, or exercise to reduce stress; distraction include d strategie s suc h a s watchin g televisio n o r playin g wit h grandchildren , specificall y t o avoi d thinkin g abou t th e problem . Escapis m include d strategie s suc h a s "Fantasize d abou t unrealisti c solution s t o problems " o r goin g o n vaca - tion , whil e expresse d emotio n include d ventin g emotion s t o others , swearing , crying , an d s o on . Ther e wer e individua l difference s i n th e numbe r o f copin g strategie s elicited ; however , mos t peopl e reporte d five o r fewe r strategie s fo r eac h o f th e copin g questions . Thus , fo r eac h person , w e code d u p t o five strategie s fo r th e problem - focuse d question s an d five fo r th e emotion-focuse d ques - tions . The number of strategies were summed within each of th e 1 0 majo r copin g categorie s separatel y fo r bot h th e problem - an d th e emotion-focuse d questions . W e compare d th e abilit y o f th e question s abou t managin g th e proble m v s managin g th e emotion s t o differentiall y elici t problem - an d emotion-focuse d coping . Paire d /-test s o n th e copin g strategie s elicite d b y thes e tw o question s confirme d tha t score s o n th e problem-focuse d an d socia l suppor t sub - scale s wer e highe r fo r th e questio n abou t managin g th e proble m (/' s range d fro m 2.4 0 t o 19.18) , whil e th e emotion - focuse d subscal e score s wer e highe r fo r th e question , "Ho w di d yo u cop e wit h you r emotions? " (f' s range d fro m -5.1 6 t o -16.16) . Interestingly , escapis m wa s th e onl y strateg y tha t wa s use d equall y a s (in)frequentl y fo r bot h questions .

Negativ

e an d positiv e emotion s wer e als o coded . Rathe r tha n groupin g th e emotions , w e use d conten t analysi s base

dupon the exact emotion terms used. Thus, we coded for themention of specific emotion terms, such as sad, blue, miser-

able , angry, etc. We identified 89 negative terms and 23 positiv e terms . Th e numbe r o f emotion s wer e summe d fo r bot h th e positiv e an d negativ e dimensions .

RESULT

S

Stress

Analyses Usin g extensiv e probing , 90
% (n = 952
) o f th e me n reporte d a proble m i n th e pas t week , compare d t o th e typica l 70-80
% of community residents who typically report prob- lem s o n self-repor t questionnaire s (e.g. , Aldwi n & Reven - son , 1987)
. Nonetheless , th e old-ol d subject s wer e stil l les s likel y t o repor t havin g ha d a proble m i n th e pas t week , x 2 ( N = 1,052) = 17.69, p<. 001. Nearly one fifth of the old-old me n (17.4% ) sai d the y ha d ha d n o problem s i n th e pas t week , compare d t o onl y tw o me n (2.6% ) i n th e earl y mid - lif e group . Percentage s fo r th e tw o othe r group s wer e 6.9 % an d 10.7% , respectively . A s expected , ther e wer e ag e difference s i n th e type s o f problem s me n reported , x 2 (n = 952
) = 121.18
, p < .0001. A s Tabl e 1 indicates , th e oldest-ol d wer e mos t likel y t o repor t healt h problem s (27.5% ) an d genera l hassle s (27.5%) , while the early mid-life group were most likely to repor t problem s a t wor k (43.2% ) an d wit h children , wh o wer e ofte n adolescent s o r youn g adult s (20.3%) . I n term s o f th e type s o f problem s reported , th e tw o oldes t group s wer e mos t simila r t o eac h other , whil e th e middle-age d group s wer e als o wrestlin g wit h simila r problems . Consisten t wit h ou r earlie r wor k o n retiremen t stress , ver y fe w o f th e NA S me n reporte d problem s wit h retiremen t (fo r a revie w se e Bosse , Levenson , Spiro , Aldwin , & Mroczek , 1992
; Bosse , Spiro , & Levenson, in press). Note that the problems re- ferre d t o a s "wife " coul d includ e th e wife' s healt h problem s a s well , whic h i s probabl y wh y thi s i s reporte d slightl y mor e frequentl y i n th e olde r groups . Amon g th e me n wh o reporte d havin g a problem , w e foun d tha t ther e wa s onl y a tren d fo r th e olde r group s t o rat e thei r problem s a s les s stressfu l [F(3,947 ) = 2.19 , p = .09] ; th e tw o middle-age d group s ha d mean s o f 3. 9 an d 4.1 , an d th e tw o olde r group s ha d mean s o f 3.8 . We investigated whether thi s relationshi p wa s linea r o r nonlinear . A margina l cubi c Tabl e 1 . Ag e Difference s i n th e Type s o f Problem s Reporte d (Percentages ) Typ e o f Proble m Healt h Wil e

Childre

n Socia l Wor k

Retiremen

t

Finance

s

Genera

l hassle s

Bereavemen

t

Multipl

e problem sEarly Mid-Life 45-5
4 (n = 74
) 5.4 % 6. 8 20. 3 5. 4 43.
2 0. 0 0. 0 14. 9 1. 4 2. 7 Lat e Mid-Lif e 55-6
4 (n = 377
) 15.1 % 9. 8 13. 0 10. 3 23.
9 1. 6 5. 6 15. 9 1. 3 3.

4Young-Old

65-7
4 ( n = 392
) 21.2
% 11. 5 13. 3 11. 0 7. 1 1. 5 4. 3 25.
0 2. 8 2.

3Old-Old

7 5 + (n = 109
) 27.5
% 11. 9 9. 2 10. 1 3. 7 0. 9 4. 6 27.
5 1. 8 2. 8

Note.,

X2 (l,n = 952
) = 121.18,/
?

< .0001.Downloaded from https://academic.oup.com/psychsocgerontology/article/51B/4/P179/650788 by guest on 25 August 2023

P184ALDWINETAL.

ter m [F( 1,947) = 2.96, p = .09] indicated a nonlinear trend . However , thes e difference s appea r t o b e trivial .

Separat

e chi-square s wer e use d t o examin e age-grou p difference s fo r eac h appraisa l item . Ther e wer e n o ag e difference s i n appraisal s commonl y though t t o increas e wit h age , such as threat, harm/loss, or at a loss for what to do next , no r wer e ther e an y difference s i n bein g worrie d abou t other s (se e Tabl e 2) . However , ther e wer e significan t ag e difference s i n tw o appraisals : feelin g challenged , x 2 (3 , n = 949
) = 20.77
, p < .001 , an d annoyed , x 2 (3 , n = 935
) = 15.65 , p = .001. The early mid-life group was most likely t o conside r th e proble m a challeng e (64.9%) ; bu t 40.6
% o f thos e i n lat e lif e als o reporte d bein g challenge d b y th e problem . Thos e i n earl y mid-lif e wer e nearl y universall y annoye d b y th e proble m (87.5%) , bu t two-third s (67.0% ) o f th e old-ol d wer e als o annoye d b y th e problem . Finally , olde r individual s als o endorse d fewe r appraisal s [F(3,946 ) = 6.05 , p< .001]; mean number of appraisals endorsed ranged fro m 1.9 7 fo r th e old-ol d t o 2.6 6 fo r thos e i n earl y mid-life . I n explorator y analyses , w e foun d tha t al l o f th e appraisals varie d significantl y b y proble m type , wit h chi-square s rang - in g fro m 30.8
3 t o 105.2
9 (ally' s < .001) . However , thi s di d no t totall y explai n th e ag e difference s foun d i n th e previou s analysis . A s woul d b e expected , wor k problem s wer e mos t likel y t o elici t bot h challenge d (74% ) an d annoye d appraisal s (85.2%) ; the absence of work problems in late life would explai n wh y thes e appraisal s decrease . I n contrast , healt h problem s wer e mos t likel y t o elici t bot h threa t (28.1% ) an d harm/los s appraisal s (22.5%) , an d ag e increase d th e likeli - hoo d tha t th e proble m reporte d wa s a healt h problem . How - ever , thi s environmenta l "pull " wa s apparentl y no t stron g enoug h t o resul t i n ag e difference s i n threa t an d harm/los s appraisals . Amon g al l ag e groups , wif e problem s univer - sall y elicite d annoyanc e (92.8%) . I n summary , som e interestin g ag e difference s emerge d i n th e propensit y t o repor t stressor s an d i n th e type s o f problem s tha t occurred , whic h i n tur n affec t appraisal s o f challeng e an d annoyance . Despit e repeate d probing , olde r adults , espe - ciall y th e old-old , wer e les s likel y t o repor t havin g ha d a proble m i n th e pas t week . Ther e wer e als o clea r ag e differ - ence s i n type s o f problem s reported , wit h th e middle-age d grapplin g wit h childre n an d work , an d olde r adult s facin g healt h problem s an d genera l problem s o f everyda y living . Whil e n o significan t ag e difference s appeare d i n th e stressful - nes s ratings , ther e wer e difference s i n th e specifi c appraisals . Olde r me n wer e les s likel y t o fee l challenge d o r annoyed , probabl y du e t o th e lac k o f a wor k role . However , the y wer enot more likely to feel threatened, harmed, that they had suffere d a loss , o r wer e a t a los s a s t o wha t t o d o next .

Coping

Analyses Ag e difference s i n copin g strategie s wer e analyze d usin g th e multivariat e analysi s o f varianc e (MANOVA ) progra m i n SPS S (SPSS, 1986)
, whic h compensate s fo r unequa l cel l sizes . Analyses on the semi-structured interview questions reveale d tha t onl y wea k ag e difference s emerge d i n th e type s o f strategie s use d t o dea l wit h th e problem , Wilks ' lambd a = .955,F(30,2757 ) = 1.46 , p = .05 . N o significan t difference s betwee n ag e group s emerge d fo r strategie s use d t o dea l wit h emotions , Wilks ' lambd a = .959 , F(30,2757 ) = 1.34 , n.s .

Inspectio

n o f th e univariate Fs for the strategies elicited by th e problem-focuse d questio n reveale d tha t onl y tw o strate - gie s differed , interpersonal , F(3,948 ) = .273 , p < .05 , an d escapism , F(3,948 ) = 4.53 , p < .01 . Whil e a Scheffe' s post-ho c rang e tes t showe d n o group s differe d significantl y a t th e .0 5 leve l fo r interpersona l strategies , a significan t weighte d linea r relationship , F(l,948 ) = 6.033 , p < .01 , indicate d tha t olde r individual s appeare d t o us e fewe r inter - persona l strategies . A s expected , olde r peopl e di d us e les s escapism , an d examinatio n fo r nonlinea r trend s usin g a one - wa y analysi s o f varianc e showe d a significan t cubi c relation - shi p [F ( 1,948) = 8.25, p < .01]. Only a few of the early mid-lif e grou p reporte d thes e strategies ; virtuall y non e o f th e othe r thre e group s did . (Th e mea n fo r th e old-ol d grou p wa s .000 , indicating that none of the 97 men over the age of 74 reporte d usin g an y for m o f escapis m t o dea l wit h thei r problems. ) A MANOV A examinin g ag e difference s i n th e BWO C subscale s reveale d a differen t patter n o f results . There was a significan t overal l effec t o f ag e fo r copin g strategies , Wilks ' lambd a = .954 , F(24,2715 ) = 1.88 , p< .01 . A s Tabl e 3 indicates , fou r o f th e eigh t subscale s showe d ag e differ - ences : instrumental action, social support, cognitive refram- ing , an d interpersona l hostil e strategies . Whil e al l showe d linea r decrease s wit h age , Scheffe' s post-ho c rang e test s suggeste d tha t th e middle-age d group s wer e significantl y differen t fro m th e old-old . A s indicate d b y paire d subscript s o n Tabl e 3 , th e old-ol d wer e significantl y differen t fro m bot h middle-age d group s o n th e instrumenta l action , cognitiv e reframing , an d interpersona l hostil e subscales . Th e young - ol d grou p wa s als o significantl y differen t fro m th e lat e mid - lif e grou p o n th e latte r strategy .

However

, th e strategie s use d b y differen t ag e group s migh t b e influence d b y th e type s o f problem s the y encountered . A s Tabl e 2. Age Differences in the Appraisals of Problems Reported (Percentages) Typ e o f

Appraisa

lEarly Mid-Life ( " = 74
)Late Mid-Life ( n = 377
)Young-Old ( n = 392
)Old-Old (n = 109
) Threa t

Harm/los

s

Challenge

d A t a los s

Annoye

d

Worrie

d18.9 24.
3 64.
9 21.
6 87.
5 51.
414.3
24.
2 58.
9 24.
7 80.
0 39.
314.0
20. 4 47.
4 24.
7 72.
6 37.
815.0
16. 8 40.
6 20. 6 67.
0 38.
7 1.2 4 3.5 4 20.77
* 1.1 5 15.65 * 4.8 9 .001

.Downloaded from https://academic.oup.com/psychsocgerontology/article/51B/4/P179/650788 by guest on 25 August 2023

STRESS, COPING, AND APPRAISALP185

Tabl e 3 . Ag e Difference s i n Copin g Strategie s Usin g th e Brie f Way s o f Copin g Copin g Strateg y

Instrumenta

l actio n Escap e avoidanc e Socia l suppor t Blam e sel f

Cognitiv

e reframin g

Interpersona

l hostil e Threa t minimizatio n Sel f isolatio nEarly Mid-Life ( n = 74
) 5.51 " 3.2 1 3.8 8 2.3 0 2.53 ° 2.62 ° 3.1 1 1.9

1Late Mid-Life

( " = 376
) 5.32 ° 3.5 8 3.8 7 2.0 4 2.81 °

2.46°

c 3.0 2 1.8

5Young-Old

(n = 390
) 4.8 5 3.4 5 3.3 5 2.0 9 2.2 9 1.97 d 2.9 9 1.8

3Old-Old

(n = 107
) 4.35 " 3.1 5 3.3 2 1.5 7 2.00 " 1.81 " 2.8 9 1.5

6Univariate

F(3,943

) 4.90* * 0.8 4 2.77 * 1.7 9 4.73* *

7 29***

0.1 7 0.9 2 Note. Superscript s denot e pair s whic h ar e differen t fro m eac h other , i.e. , as ar e differen t fro m bs, an d c s diffe r fro m ds. *p < .05 ; **p < .01 ; ***p < .001 . indicate d earlier , th e group s differe d mos t o n problem s wit h wor k an d children ; no t surprisingly , th e tw o ol d groups , wh o wer e mos t likel y t o hav e relinquishe d bot h wor k an d activ e parentin g roles , wer e leas t likel y t o repor t thes e type s o f problems . Thus , w e create d tw o dumm y indicato r variable s fo r thes e tw o proble m type s i n orde r t o contro l fo r proble m type . MANCOVAs revealed that, once we controlled for problem s wit h childre n an d work , th e multivariat e F fo r th e effec t o f ag e grou p o n copin g strategie s elicite d b y th e proble m questio n intervie w dat a wa s n o longe r significant , Wilks ' lambd a = .959 , F(30,2751 ) = 1.31 , p = .12 .

Further

, th e multivariat e F fo r th e effec t o f ag e grou p o n th e BWO C subscale s als o becam e nonsignificant , Wilks ' lambd a = .966 , F(24,2710 ) = 1.37,/ ? = .11. W e als o examine d whethe r ther e wer e ag e difference s i n copin g effort , i.e. , th e numbe r o f strategie s reported . Fo r th e intervie w data , ther e wer e n o difference s i n th e numbe r o f strategie s elicite d b y eithe r th e problem-focuse d questio n

F(3,948

) = 2.00 , n.s. , o r th e emotion-focuse d question ,

F(3,948

) = 1.15 , n.s . Th e mean s fo r th e problem-focuse d an d emotion-focuse d strategie s wer e 2.6 0 an d 2.27 , respec - tively . However , ther e wer e highl y significan t difference s fo r th e numbe r o f item s reporte d o n th e BWOC , F(3,944 ) = 9.05 , p < .0001. The weighted linear term was significant, F( 1,945 ) = 25.02, p < .001, but a Scheffe's post-hoc range tes t (p < .05 ) showe d tha t thos e i n th e old-ol d grou p ( M = 11.00 ) wer e mos t differen t fro m thos e i n th e middle-age d group s (M s = 13.2 4 an d 13.39 , respectively) , whil e th e young-ol d wer e intermediat e (M = 11.99) . Not e tha t thi s remaine d significan t eve n afte r controllin g fo r typ e o f prob - lem , F(3,946 ) = 6.27 , p< .001 .

Emotions

and Efficacy Analyses I n thes e interviews , ther e wa s n o significan t effec t o f ag e grou p o n th e reportin g o f eithe r negativ e emotions , F(3,926 ) = 1.22 , n.s. , o r positiv e emotions , F(3,923 ) = 0.31 , n.s .

Finally

, a s expected , th e elderl y fel t the y wer e jus t a s effectiv e i n dealin g wit h th e proble m a s di d th e middle-age d men , F(3,923 ) = .77 , n.s. , an d fel t the y woul d b e jus t a s effectiv e i n dealin g wit h simila r problem s i n th e future ,

F(3,923

) = 0.57 , n.s .

DISCUSSIO

N Thi s stud y provide d a n in-dept h examinatio n o f th e type s o f stressor s experience d b y olde r men , ho w the y ar e ap -praised, and the coping strategies used in dealing with everyda y problems . Give n th e complexit y o f th e results , w e wil l divid e th e discussio n int o tw o sections , on e o n stres s an d th e othe r o n coping .

Stress

A s note d i n th e introduction , ther e i s a discrepanc y be - twee n th e widesprea d assumption s abou t th e severit y an d type s o f stressor s though t t o b e commo n i n lat e lif e an d empirica l studies , whic h typicall y find tha t olde r peopl e repor t fewe r stressors . W e hypothesize d tha t thi s findin g migh t b e a n artifac t reflectin g a reportin g bia s amon g th e elderly , whic h woul d b e attenuate d b y th e us e o f intensiv e intervie w techniques . T o a certai n extent , th e hypothesi s o f a reportin g bia s wa s supported . B y usin g extensiv e probing , 90
% o f th e NA S me n reporte d havin g ha d a proble m i n th e pas t week . Thi s contrast s wit h 25-30
% wh o d o no t repor t problems , eve n ove r th e pas t month , i n mos t communit y survey s (e.g. , Ai d wi n & Revenson , 1987)
. Amon g thos e tha t di d repor t a problem , ther e wer e n o significan t ag e difference s i n th e stres s rating s o f th e problem .

Nonetheless

, w e di d observ e age-relate d difference s i n man y aspect s o f th e stres s process . Despit e extensiv e prob - ing , th e old-ol d wer e stil l les s likel y t o repor t havin g ha d a proble m i n th e pas t wee k (17.4 % a s oppose d t o 2.6 % i n mid - life) . In addition, the type of problem did vary by age, as expected . Th e middle-age d me n wer e mos t likel y t o repor t problem s a t wor k an d wit h thei r children , wh o wer e primar - il y adolescent s an d youn g adults . A t thi s lif e stage , man y o f th e me n reporte d problem s wit h thei r childre n usin g drug s an d alcohol , gettin g int o troubl e wit h th e law , o r ha d con - cern s abou t launchin g the m int o adulthood . I n addition , man y o f th e NA S me n hel d fairl y responsibl e position s an d ha d t o supervis e employee s o r subordinates , placat e thei r ow n supervisors , dea l wit h th e declinin g economy , lawsuits , an d s o on . T o a certai n extent , th e olde r me n substitute d variou s project s fo r thei r work-relate d roles , suc h a s mainte - nanc e an d repai r tasks , voluntee r work , an d participatio n i n th e communit y o r churc h governance . Thes e type s o f task s wer e code d a s "genera l hassles, " whic h i s wh y these , and , a s expected , healt h problems , wer e reporte d mor e frequentl y i n lat e life . Ther e wer e als o ag e difference s i n ho w stres s wa s ap - praised . Th e old-ol d wer e les s likel y t o apprais e thei r prob

-Downloaded from https://academic.oup.com/psychsocgerontology/article/51B/4/P179/650788 by guest on 25 August 2023

P186ALDW1NETAL.

le m a s a challeng e (althoug h som e 40
% stil l di d so) , an d the y wer e les s likel y t o repor t bein g annoye d b y thei r proble m (althoug h two-third s wer e annoyed) . A s thes e appraisal s wer e largel y associate d wit h wor k problems , thi s age-relate d findin g wa s largel y du e t o difference s i n th e typ e o f proble m bein g faced . I n man y ways , th e lac k o f difference s i n som e appraisal s wa s mor e interesting . Contrar y t o commo n opinion , olde r me n wer e no t mor e likel y t o apprais e problem s i n term s o f los s o r threat , despit e th e fac t tha t the y wer e dealin g wit h healt h problems . Thus , althoug h w e kno w tha t olde r peopl e ar e mor e likel y t o experienc e losse s throug h bereavemen t an d healt h problem s (Aldwin , 1990)
, thei r everyda y live s ar e no t characterize d i n thes e terms . Wh y th e old-ol d reporte d fewe r problems , despit e exten - siv e probing , i s no t clear . I n part , thi s decreas e ca n b e attribute d t o a declinin g involvemen t i n som e socia l role s suc h a s wor k an d activ e parenting . Fewe r problem s i n th e old-ol d migh t reflec t a narrowin g o r constrictio n o f everyda y life . A s

Johnso

n an d Bare r (1993 ) pointe d out , th e live s o f th e oldes t ol d ofte n hav e muc h narrowe r foci , a s activitie s whic h ar e n o longe r possibl e o r physicall y to o taxin g ar e eliminated .

Baltes

' (1987 ) construc t o f selectiv e optimizatio n implie s tha t th e ol d narro w thei r rang e o f activitie s t o thos e whic h ar e mos t possibl e an d rewarding . Further , i t shoul d b e emphasize d tha t w e wer e examinin g problem s i n th e pas t week . Majo r lif e event s ma y no t decreas e wit h age , a s lon g a s th e samplin g o f item s i s age-appropriat e (Aldwin , 1990)
. I n addition , th e natur e o f th e sampl e mus t b e considered . Th e NA S sampl e consist s largel y o f White , middle-clas s men , mos t o f who m ar e marrie d an d ha d job s wit h pensions .

Further

, the y ha d t o b e health y enoug h t o mak e a tri p int o downtow n Boston . Som e o f them , especiall y thos e wit h dementia , wer e accompanie d b y thei r wives , bu t al l o f the m wer e ambulatory . Thus , ou r old-ol d mus t b e considere d a s relativel y healthy , althoug h mos t di d hav e chroni c illnesse s o f som e sort , suc h a s arthritis , hypertension , an d diabetes .

Nonetheless

, i f ou r sampl e ha d bee n elderl y Africa n Ameri - ca n widows , 80
% o f who m fal l belo w th e povert y lin e (Smeeding , 1990)
, ou r finding s migh t hav e bee n ver y differ - ent . However , th e fac t tha t Africa n America n elder s typi - call y repor t highe r satisfactio n level s tha n d o Whit e elder s (Jackson , Antonucci , & Gibson , 1990
) suggest s tha t simila r processe s ma y b e occurrin g i n othe r demographi c group s a s well . I n contrast , a developmenta l interpretatio n i s tha t th e old-ol d ma y simpl y hav e develope d a mor e "philosophical " outloo k i n late r life , whic h mean t tha t the y wer e les s likel y t o regar d situation s a s problematic . T o th e exten t tha t Erikson' s (1950 ) eg o integrit y involve s acceptanc e o f lif e a s i t is , th e ol d ma y becom e d e fact o stoics .

However

, thi s "philosophical " outloo k ca n als o b e attrib - ute d t o a shif t i n th e tempora l characteristic s o f stres s fro m middl e t o lat e lif e - tha t is , whethe r stres s i s episodi c o r chronic . T o a larg e extent , th e middle-age d me n wer e dealin g primaril y wit h episodi c stressor s - a particula r proble m a t wor k whic h neede d resolvin g o r specifi c inci - dent s wit h thei r children , whic h immediatel y cam e t o min d whe n aske d abou t a proble m i n th e pas t week . Healt h problem s i n lat e life , o n th e othe r hand , ten d t o b e chronic , an d thu s thei r managemen t become s par t o f th e fabri c o fdaily living. To the extent that consistent management strat- egie s ar e develope d whic h bot h forestal l problem s an d miti - gat e th e negativ e impac t o f a chroni c stressor , the n tha t chroni c stresso r n o longe r i s a n immediat e proble m pe r se . Fo r example , diet , exercise , an d medication s ar e typicall y use d t o manag e chroni c illnesse s suc h a s hypertensio n an d diabetes . Unles s ther e wa s a n episod e i n whic h metaboli c contro l wa s lost , suc h a s a faintin g spel l i n diabetes , chance s ar e quit e likel y tha t peopl e wit h chroni c illnesse s ma y no t perceiv e havin g ha d a proble m i n th e pas t week . Fo r exam - ple , we suspect that hypertension was only cited as a prob- le m i f i t wa s a recen t diagnosi s o r i f ther e wa s a sudde n chang e i n status , eve n thoug h nearl y 40
% o f th e NA S me n ar e hypertensive . (Fo r mor e discussio n o f thi s issue , se e Aldwi n & Brustrom , i n press. ) I n summary , th e result s supporte d a respons e bias , i n tha t ag e difference s wer e attenuate d b y th e intervie w process .

However

, age-relate d difference s di d emerge , primaril y i n th e typ e o f proble m reported , whic h i n tur n affecte d ho w problem s wer e appraised . However , th e old-ol d wer e stil l les s likel y t o repor t problems , whic h w e attribute d i n par t t o a shif t fro m episodi c t o chroni c stressors . I t mus t b e empha - sized , however , tha t al l o f th e difference s reporte d her e wer e cross-sectional , whic h mean s tha t w e canno t distinguis h ag e fro m cohor t effects .

Coping

W e hypothesize d tha t interview s woul d attenuat e th e usua l age-relate d pattern s see n i n copin g strategie s an d tha t an y difference s whic h di d occu r woul d reflec t age-relate d difference s i n th e type s o f problem s bein g faced . Further , an y residua l ag e effec t woul d b e see n primaril y i n th e copin g interview s rathe r tha n th e checklist . Th e expecte d ag e difference s di d emerg e i n bot h th e intervie w an d BWO C copin g subscale s bu t wer e stronge r fo r th e checklis t tha n fo r th e interviews , contrar y t o expecta - tions . On the interviews, older men were less likely to report usin g interpersona l hostilit y an d escapism . Th e us e o f thi s latte r strateg y wa s rare , eve n amon g thos e i n earl y mid-life , an d nonexisten t i n th e old-old . Thi s support s simila r findings i n othe r population s (se e Aldwin , 1991
, fo r a review) .

Escapis

m i s a strateg y whic h i s probabl y th e leas t effica - cious , and includes using drugs, alcohol, and wishful think- ing . Hopefully , b y mid-lif e mos t hav e learne d tha t thes e ar e ineffectiv e strategies ; thos e wh o hav e no t ar e probabl y functionin g fairl y poorl y (se e als o Vaillant , 1993)
. O n th e BWOC , ag e difference s wer e see n i n fou r o f th e eigh t subscales : instrumenta l action , socia l support , cognitiv e reframing , an d interpersona l hostile . Th e old-ol d wer e mos t differen t fro m th e othe r group s i n thei r coping . Their decrease i n us e o f interpersona l hostil e strategie s i s consisten t wit h th e abandonmen t o f escapis m a s copin g strategie s - bot h strategie s ten d t o b e rathe r ineffective . The y als o reporte d fewe r instrumenta l actions , ye t rate d thei r efficac y a s hig h a s th e middle-age d adults . I t i s interestin g t o not e tha t the y als o use d slightl y les s copin g effort - the y average d tw o strategie s les s tha n th e middle-age d o n th e BWOC , althoug h th e inter - vie w question s di d no t revea l ag e differences . I n part , thi s differenc e betwee n th e intervie w an d th e BWO C ma y b e du e t o th e fac t that , despit e repeate d probing

,Downloaded from https://academic.oup.com/psychsocgerontology/article/51B/4/P179/650788 by guest on 25 August 2023

STRESS, COPING, AND APPRAISALP187

th e me n reporte d ove r twic e a s man y copin g strategie s usin g th e checklis t (M = 12.49 ) tha n i n th e interview s ( M = 4.88 ) fo r th e combine d problem - an d emotion-focuse d questions . I n part , th e checklist s ma y serv e t o jog people's memories as t o wha t the y di d an d though t durin g th e event , an d thu s ma y provid e mor e complet e reporting , contrar y t o th e usua l expec - tation s abou t interview s v s checklists . Often , peopl e ma y b e s o use d t o ho w the y cop e wit h problem s tha t i t doe s no t occu r t o the m t o repor t th e strateg y unles s the y ar e reminde d o f it .

Sometime

s th e checklist s ma y overcom e reluctanc e t o tal k o r eve n lo w self-esteem . Fo r example , one interview with one of th e fe w inne r cit y elderl y i n th e stud y wa s mos t instructive . H e wa s dealin g wit h a particularl y troublesom e long-standin g problem . I n th e semi-structure d interview , h e kep t insistin g tha t h e di d littl e o r nothing , becaus e ther e wa s ver y littl e tha t h e coul d d o excep t tr y no t t o thin k abou t th e problem . Wit h th e checklist , however , h e wa s abl e t o repor t fou r o r five strategie s tha t h e ha d actuall y used .

However

, thi s ma y als o reflec t a semanti c shift . Th e middle-age d me n immediatel y kne w wha t w e mean t b y stres s an d coping , bu t th e term s seeme d mor e foreig n t o th e olde r men . W e ha d t o defin e stres s i n term s o f bein g "bothered " o r "troub