Individual Differences in Perceived Information Processing Styles in ...

10 downloads 0 Views 254KB Size Report
perception of his or her dominant mode of information processing in stress and ... measuring three different perceived processing styles: (1) rational processing, ...
Cogniti ve Th erapy and Research , Vo l. 23, No. 4, 1999, p p. 345-371

Individ ual D ifferences in Perceived Inform ationPro cessing Styles in Stress and Coping Situations: D evelopm en t and Validation of the Perceived Modes of Pro ce ssing Inven tory1 Law re nce R . B urns 2,3 an d Thom as J. D ’Zurilla 2

A new self-repo rt in strum ent w as constructed to assess a person’ s aw aren ess an d perceptio n of his or her dom in an t m ode of in form atio n processing in stress an d copin g situ atio ns. The item s w ere based on E pstein’ s (1990) cogn itive± experiential self-th eory, w hich d istin gu ish es betw een ratio nal an d experiential in form atio n processing. E xp lo rato ry an d con® rm ato ry facto r an alyses found that this in strum ent is m easu rin g three different perceived processing styles: (1) ratio nal processing, (2) em otio nal processing, an d (3) au tom atic processing. Collecti vely, these m easu res w ere nam ed the Perceived Modes of Processing In ventory (PMPI). D ata supportin g the reliab ility an d criterio n valid ity of the PMPI w ere presented. In gen eral, percei ved ratio nal p rocessing w as m ost strongly an d consisten tly related to ad ap tive problem focused copin g an d positi ve psych olo gical w ell-b eing. Perceived emotio nal processing w as related to ad ap ti ve em otio n-fo cused copin g (expressin g em otio ns, seekin g social support), but it w as also related to m ore psych olo gical distress an d less positi ve w ell-b eing. K E Y WOR D S: stress; coping; information processing; asse ssme nt; social problem solving.

In mode rn stress the ory, coping is a major e xplanatory construct that is fre que ntly used to account for individual diffe re nces in re sponse to stre ss. V ie wed as a mediator of the re lationship be twe e n spe ci® c stressful life eve nts and the ir e motional and physical outcom e s, coping has be e n de ® ne d by Lazarus and Folkman (1984) as the cognitive and be havioral activitie s by which a pe rson atte mpts to 1

This study is based on a dissertation that was submitted by the ® rst author in partial ful® llme nt of the re quirements for the Ph.D. in Clinical Psychology at the State University of New York at Stony B rook. 2 State Unive rsity of Ne w York at Stony B rook. A ddress correspondence to Thomas J. D’ Z urilla, Departme nt of Psychology, State Unive rsity of New York at Stony B rook, Stony B rook, Ne w York 11794-2500, or Lawre nce R. B urns, De partme nt of Psychology, Grand V alle y State University, 1 Campus Drive , A llendale, Michigan 49410. 3 Now at Grand V alle y State Unive rsity. 345 0147-5916/99/0800-0345$1 6.00/0 Ó

1999 Ple num Publishing Corporation

346

B u rns and D’Zurilla

manage spe ci® c stre ssful situations as well as the e motions that they ge ne rate . Rese arch on the structure of coping has found it to be a comple x, multidime nsional phe nome non that e ncompasse s a range of diffe re nt cognitive and be havioral activitie s (A ldwin, 1994; Folkman & Lazarus, 1980; Pe arlin & Schoole r, 1987; Carve r, Sche ier, & We intraub, 1989; Tobin, Holroyd, Reynolds, & Wigal, 1989) . Most atte mpts to organize and study diffe rent coping activitie s have be e n base d on a functional analysis of coping in which coping strate gie s are classi® e d according to the function or purpose that they se rve in spe ci® c stre ssful situations, such as changing the situation for the bette r (proble m-focuse d coping) and re ducing or controlling the e motions produce d by the stressor (e motion-focuse d coping) . In contrast, re lative ly fe w inve stigators have atte mpte d to study coping from an information-proce ssing pe rspe ctive , which focuse s on the cognitive proce sse s by which coping re sponse s are ge ne rate d and se le cted. A lthough the functional ± analytic approach has ge ne rate d important data concerning the re lations be twe en coping and adaptation (A ldwin, 1994; Lazarus & Folkm an, 1984) , the study of individual diffe re nce s in proce ssing style s or pre fere nces might also contribute to our unde rstanding and pre diction of adaptive coping and well-be ing (B re win, 1989; Epste in, 1994; Ingram, 1986) . In his cognitive -e xpe rie ntial se lf-the ory (CE ST), E pstein (1990, 1994) distinguishe s be twe e n two diffe re nt information -processing syste ms that are assum e d to play an important role in de te rmining e veryday coping behavior: an experien tial system and a ratio nal system. The e xpe rie ntial syste m is fast (orie nte d toward immediate action) , ef® cie nt (minim al atte ntional re source s), e motional, and intuitive ( i.e ., immediate knowing of some thing without the conscious use of re asoning) . In contrast, the rational syste m is slowe r (orie nted toward de laye d action) , e ffortful, de libe rate , analytic, and logical (i.e., require s validation through logic and e vide nce). It is assum ed in CE ST that the re are individual diffe re nce s in the e xtent to which pe ople rely on the se two mode s of proce ssing across a range of life situations, and that the se diffe re nce s might have important implications for adaptation and we llbe ing (E pstein & Me ie r, 1989; E pste in, Lipson, Holste in, & Huh, 1992) . The major purpose of the pre se nt study was to assess individual diffe re nce s in pe rceive d inform ation-proce ssing style s in stre ssful life situations and e xam ine the ir re lations to measure s of coping and psychologic al well-be ing. The study was conducte d in two parts. The major obje ctive of Part I was to construct a ne w se lfre port instrume nt spe ci® cally de signe d to asse ss a person’ s aware ne ss and pe rception of his or he r dominant mode of proce ssing across stre ssful situations. The ite m pool for this instrum e nt was base d on particular de ® ning prope rtie s of E pste in’ s (1994) rational and e xpe rie ntial proce ssing systems that are acce ssible to aware ness and, thus, can be se lf-re porte d on a que stionnaire . It was assum ed that subje cts’ re sponse s to these items would re liably re ¯ e ct the e xtent to which the y re ly on these two processing modalitie s whe n attempting to cope with stre ss. To e stablish structural validity, we e xam ine d the dime nsionality of the instrum ent using both e xploratory and con® rmatory factor-analy tic methods. We pre dicted that two factors would be found corre sponding to pe rceive d rational processing and perceive d expe riential processing. The set of scale s re sulting from the se analyse s was name d the Pe rceived Mode s of Proce ssing Inve ntory (PMPI) .

Pe rce ived Proce ssing

347

The obje ctive of Part II of the study was to conduct a preliminary e valuation of the reliability and conve rge nt and discrim inant validity of the PMPI. A ssuming that the re sults of Part I would con® rm that the PMPI is measuring the two hypothe size d processing style s, we pre dicted that pe rceive d rational proce ssing would be more strongly re late d to adaptive coping and positive psychologica l well-be ing than would pe rceived e xpe rie ntial proce ssing.

PA R T I: D E V E LOPME NT A ND FA CTOR A NA LY SIS OF THE PMPI Method Su bjects The subje cts for this part of the study consiste d of two sample s of unde rgraduate stude nts from the State Unive rsity of New York at Stony Brook. Sam ple 1 (N 5 403) was obtaine d by merging two subsam ple s of stude nts who were enrolle d in psychology course s at the time of the study. The ® rst subsample (N 5 216) was comprise d of volunte ers who comple te d the PMPI during Spring 1994. The second subsam ple ( N 5 187) consiste d of subje cts e nrolle d in the introductor y psychology course during Summer 1994 who participate d in the study in orde r to ful® ll a course re quire me nt. The total sample include d 234 wome n and 169 me n, with a me an age of 21.6 years. The e thnic /racial composition of this sample was approxim ate ly 63% White , 10% A frican A merican, 17% A sian A me rican, 4% Hispanic A merican, and about 6% unknown (not re porte d by the subje cts). Sample 2 consiste d of 865 unde rgraduate s e nrolle d in the introductor y psychology course during Fall 1994 who complete d the PMPI in orde r to ful® ll a course re quire me nt. This sample include d 523 wome n and 342 men, with a mean age of 19.5 years. The e thnic /racial composition was approximat ely 50% White , 13% A frican A merican, 20% A sian A merican, 8% Hispanic A me rican, and 9% unknown. Item Selectio n an d Content Valid atio n A n initial pool of 104 ite ms was writte n by the pre sent inve stigators to represe nt de ® ning prope rtie s or by-products of e ithe r rational or e xpe rie ntial proce ssing, as de scribe d by E pstein ( 1994) . In addition, se veral rational items were spe ci® cally base d on the rational proble m solving construct of D’ Zurilla and his associate s (D’ Z urilla & Goldfrie d, 1971; D’ Z urilla & Nezu, 1990; Mayde u-O livare s & D’ Z urilla, 1996) , which is subsume d within E pste in’ s rational proce ssing construct. O ur sampling of the prope rtie s of the two proce ssing constructs was limite d to conscious e xpe rie nce s that a person could se lf-re port on a questionnaire . Hence , the rational ite ms focused on such feature s as the e xpe nditure of time and mental e ffort and the use of logical re asoning, creative thinking (trying to think of new ways to cope ), and spe ci® c proble m-solving te chnique s (e .g., ge ne rating alte rnative coping re sponse s, e valuating alte rnative s). In contrast, the e xpe rie ntial ite ms focused on such qualitie s as the spe e d and e f® ciency of processing (minim al time and mental effort); the re liance on fe e lings, vibe s, hunche s, and instincts; and the re call

348

B u rns and D’Zurilla

of past coping e xpe rience s and familiar coping re sponse s that have worke d well in the past. The initial ite m pool consiste d of 35 rational items and 69 e xpe rie ntial ite ms. A conte nt validity study was the n conducte d in orde r to e valuate the e xte nt to which the two se ts of ite ms de scribe d pre viously unam biguously represe nte d the re le vant prope rtie s of the two conce ptual domains that they we re de signe d to tap. Five male and ® ve female advance d graduate stude nts in clinical psychology at the State Unive rsity of New York at Stony B rook were provide d with se parate de ® nitions and descriptions of rational proce ssing and e xpe riential proce ssing and aske d to sort each item into one of the se two cate gorie s, as appropriate . The items were arrange d in random orde r. O nly ite ms that were found to have 100% inte rrate r agre e ment among all 10 graduate stude nts were retaine d for the initial PMPI. Se ve nty-e ight ite ms me t this crite rion. Test Fo rm at an d In structions These 78 ite ms were the n distribute d at random in a self-report questionnaire using a dispositional format. The instruction s aske d subje cts to re port how they typically re spond to stre ssful situations in ge neral on the following 5-point Likerttype scale: (1) ``not at all true of me,’ ’ (2) ``slightly true of me,’ ’ (3) ``mode rate ly true of me,’ ’ (4) ``ve ry true of me,’ ’ (5) ``e xtre mely true of me.’ ’ The spe ci® c instructions were as follows: B e low is a series of state me nts that describe the way some pe ople think, fe el, and act when faced with STRE SSFUL SITUA TIO NS that occur in e ve ryday living. A stressful situation is any situation in which strong de mands are made on you that are important for your we ll-being and put a severe strain on your coping abilities and resource s. These demands might be from the e nvironment (for e xample , other people ’ s e xpe ctations, course re quireme nts, job demands, etc.) or from yourself (for example, your own goals, values, and pe rformance standards) . Ple ase re ad e ach state ment carefully, and then sele ct the numbe r from the scale be low that best describe s the e xte nt to which you fe e l the stateme nt is true of you. Consider yourself as you TYPICAL L Y cope with stressful situations in your life.

Procedure A ll subje cts in Sample s 1 and 2 comple te d the PMPI in group testing se ssions lasting approxim ate ly 1 hour. In addition to the PMPI, se ve ral othe r questionnaire s were comple te d that provide d data for othe r studie s. To prote ct the subje cts’ anonymity, only subje ct numbe rs were place d on these que stionnaire s. In addition, all subje cts were provide d with consent forms that stated that their te st re sponse s would be kept strictly con® de ntial. We initially inve stigate d the dime nsionality of the PMPI by pe rforming e xploratory factor analyse s on the data obtaine d from Sample 1. B ase d on the results of the se analyse s, con® rmatory factor analyse s were then pe rforme d on the Sample 1 data and the re sults were cross-valid ate d in Sample 2. R esu lts E xp lo rato ry Facto r A nalysis Unwe ighte d least square s e xploratory factor analysis was pe rforme d on the 78-ite m PMPI for Sam ple 1. O blique rotations (Promax) were use d in inte rpreting

Pe rce ived Proce ssing

349

the multifactor solution because correlate d factors were e xpe cted. The ® rst six e ige nvalue s were : 17.07, 8.35, 3.13, 2.22, 1.71, and 1.52. B ase d on the ratio-ofe ige nvalue -diffe re nce s crite rion ( also known as ``scree plot’ ’ ) and our inte rpre tation of the multifactor solutions, we ide nti® e d thre e factors as be ing most appropriate for the data. A fourth factor was also examine d but it was discarde d be cause the rotate d solution was not inte rpre table substantive ly. B ased on a conte nt analysis of the item clusters, the three factors were name d (1) rational proce ssing, (2) e motional proce ssing, and (3) autom atic processing. A s pre dicted, the rational proce ssing factor corre sponds to E pste in’ s (1994) rational proce ssing construct. Contrary to our hypothe sis, howe ve r, the original e xpe rie ntial ite m subse t forme d two diffe rent factors. The items forming the emotional processing factor de scribe a reliance on fee lings, e motions, vibe s, hunche s, and instincts to guide one ’ s coping efforts, where as the ite ms associate d with the autom atic proce ssing factor focus on the spe e d and e f® cie ncy of processing (minimal time and effort) and the immediate knowing of how to cope base d on past coping e xpe rience s. A t this point a total of 46 items were de le te d from the inve ntory base d on conce ptual and /or empirical conside rations. Seve ral ite ms in e ach cluste r were found to be very similar in wording or content to othe r ite ms and, thus, were conside re d re dundant or supe r¯ uous. In addition, some ite ms appe are d to load on two diffe re nt factors be cause the ir wording or conte nt focuse d on a common prope rty rathe r than a distinctive one. Finally, seve ral ite ms were discarde d be cause the y loade d most highly on some othe r uninte rpretable factor. A fter screening out the se 46 items, we were le ft with 32 ite ms in the PMPI. A nothe r unwe ighte d least square s exploratory factor analysis with oblique rotations (Promax) was the n pe rforme d on the re duce d 32-ite m PMPI to de termine if the de letion of ite ms might have affe cted the factor structure of the inve ntory. The ® rst six eigenvalue s obtaine d in this analysis were 7.86, 4.49, 2.46, 1.50, 1.09, and 0.99. A gain, the ratio-of-e ige nvalue -diffe rence s crite rion sugge sted that a thre efactor solution was most appropriate for the data, which corre sponde d to the same thre e processing constructs de scribed pre viously. The pe rcentage s of variance accounte d for by the two- and thre e-factor solutions were 38% and 46%, re spe ctively. The rotate d four-factor solution was e xam ine d, but it mere ly divide d the autom atic processing factor into two-ite m subse ts that were not clearly distinguishab le conce ptually. More ove r, most of the ite ms in this fourth cluster also had re lative ly high loadings on the autom atic proce ssing factor. He nce , the ® nal 32-ite m ve rsion of the PMPI was organize d into three pe rceive d processing scale s as follows: a 12-ite m Rational Proce ssing (RP) scale, a 10-ite m E motional Proce ssing (E P) scale , and a 10-ite m A utom atic Proce ssing (A P) scale. The ite ms in e ach of these scale s are pre se nted in A ppe ndix A . Con® rm ato ry Facto r A nalysis To establish the structural validity of the PMPI, we the n pe rforme d con® rmatory factor analyse s on the instrume nt in two sample sÐ Sample 1 (the same sample use d in the exploratory factor analysis) and an inde pende nt cross-valida tion sample drawn from the same population ( Sample 2). In orde r to con® rm that our thre e -

350

B u rns and D’Zurilla

dime nsional model is supe rior to a two-dim e nsional mode l, we compare d goodne ssof-® t indice s for the two- and thre e-factor mode ls in both sample s. Maxim um like lihood con® rmatory factor analysis was pe rforme d on the covariance matrix of the ite ms in both sample s. A n oblique two-factor inde pende nt cluste rs solution and an oblique thre e -factor inde pe nde nt cluste rs solution were ® tted to these data. E ach mode l was obtaine d by e stimating all factor loadings found in the e xploratory factor analysis to be large r than .2 while ® xing all othe r loadings at zero. B ecause the asse ssment of goodne ss-of-® t of a mode l is a multiface ted and controve rsial issue (Tanaka, 1993; Loehlin, 1987) , multiple indice s should be use d (se e B e ntle r, 1990; B olle n & Long, 1993; Tanaka, 1993; McDonald & Marsh, 1990) . In particular, the use of the chi-square statistic alone can be misle ading be cause it is ve ry se nsitive to sample size and ofte n statistically reje cts a mode l in re lative ly large sample s (B rowne & Cude ck, 1993; Loe hlin, 1987) . Hence , in addition to chi-square , we also use d the following indice s to asse ss goodne ss of ® t: the Root Me an Square d E rror of A pproximatio n (RMSE A ; Steiger, 1990) , the Root Mean Square d Residual ( RMSR; JoÈ re skog & SoÈ rbom , 1993) , the Goodne ss-of-Fit Inde x (GFI: JoÈ re skog & SoÈ rbom , 1985) , the Comparative Fit Inde x (CFI: B entle r, 1990) , and the E xpe cted Cross-V alidation Inde x (E CV I; Browne & Cudeck, 1989) . A de quate to good ® t is sugge ste d by RMSE A and RMSR value s close to .05 and le ss. For the GFI and CFI indice s, value s be twe e n .80 and 1.00 indicate ade quate to e xce llent ® t. Whe n using the E CV I to choose the best model, one would se lect the mode l with the lowe st E CV I value . The re sults for both sample s are pre se nte d in Table I. For e very goodne ss-of® t inde x, the value s cle arly show that the two-factor mode l is inappropriat e for the data in both sample s. In contrast, the oblique thre e -factor inde pe nde nt cluste rs solution ade quate ly ® ts the data in both sample s. In Sample 1, the CFI and the GFI for the thre e-factor mode l both indicate a relative ly good ® t (CFI 5 .85, GFI 5 .83) . The RMSE A value also shows that this mode l can be conside re d a re asonable approxim ation to the se data, RMSE A 5 .064, and the RMSR value

Table I. Goodne ss-of-Fit Indices for Two- and Three -Factor Mode ls for the Pe rceive d Mode s of Proce ssing Inventory Sample 1a

x

Inde x 2

d.f. p -value GFI RMSE A RMSR CFI E CV I

Sample 2 b

2 Factors

3 Factors

2 Factors

3 Factors

2262.43 463 , .01 .62 .098 .139 .64 5.98

1225.14 461 , .01 .83 .064 .083 .85 3.41

4512.93 463 , .01 .62 .101 .153 .60 5.38

1993.91 461 , .01 .86 .062 .071 .85 2.47

Notes. GFI 5 Goodne ss-of-Fit Index; RMSE A 5 Root Me an Squared E rror A pproximation; RMSR 5 Root Mean Square d Re sidual; CFI 5 Comparative Fit Index; E CV I 5 E xpe cte d Cross-V alidation Index. a N 5 403. b N 5 865.

Pe rce ived Proce ssing

351

approache s the acceptable range (RMSR 5 .083) . In the cross-valid ation sample , the diffe re nce s in favor of the thre e-factor mode l are eve n large r than in Sample 1 on most goodne ss-of-® t indice s. In this sample , the CFI and GFI for the three factor mode l were .85 and .86, re spe ctive ly, whe re as the RMSE A and RMSR value s improve d to .062 and .071, re spective ly. With re gard to the crite rion of e xpe cted cross-valida tion, the E CV I value is lowe r for the three -factor mode l in both sample s, with the large st diffe re nce occurring in the cross-valida tion sample (2.47 vs. 5.38) . Not une xpe ctedly, the chi-square statistic did not provide support for e ithe r mode l in Sample 1 or in Sample 2. Howeve r, a formal te st of the supe riority of the thre e-factor mode l ove r the two-factor mode l in each of these sample s was obtaine d by pe rforming a neste d chi-square te st, which was found to be signi® cant in Sam ple 1, x 2dif(2) 5 1037.29, p , .001, as well as in Sample 2, x 2dif(2) 5 2519.02, p , .001. These re sults provide additional support for the thre e -factor mode l. The factor loadings obtaine d afte r ® tting an oblique , thre e-factor inde pende nt cluste rs solution in the cross-valid ation sample ( Sam ple 2) are presente d in Table II, while the inte rfactor corre lations along with the ir e stimated standard errors are pre se nted in Table III. A s Table II shows, e ach of the PMPI ite ms is loade d by only one factor and all loadings are quite large . The inte rcorre lations in Table III indicate that rational proce ssing is positive ly relate d to autom atic proce ssing, but ne gative ly re late d to e motional proce ssing. A s e xpe cted, e motional proce ssing and automatic proce ssing are positive ly re late d. E motional proce ssing and autom atic processing share about 20% of the ir variance . In contrast, rational proce ssing has only 4% and 9% of its variance in common with e motional processing and autom atic processing, re spe ctive ly.

PA R T II: R ELIA B ILITY A ND V A LID ITY OF THE PMPI The purpose of Part II of the study was to conduct a pre liminary e valuation of the re liability and conve rgent and discriminan t validity of the thre e PMPI scale s. With re gard to re liability, both inte rite m homoge ne ity and temporal stability were asse sse d. V alidity was e valuate d by examining conve rge nt and discriminant corre lations with a numbe r of e xte rnal criterion variable s, including othe r measure s of pe rceive d rational and e xpe rie ntial processing, as well as me asure s of affe ctivity, pe rsonality, ge ne ral inte llige nce, situational coping strate gie s, and psychologic al well-be ing. Method Su bjects an d Procedure The de scriptive statistics, inte rscale corre lations, and reliability analyse s reporte d as follows are base d on the PMPI data obtaine d from Sample s 1 and 2 in Part I. Most of the validity analyse s are base d on subgroups from Sam ple 1 or Sample 2. When new sample s were use d, they were obtaine d arbitrarily from the same population of unde rgraduate stude nts from which Sample s 1 and 2 were

352

B u rns and D’Zurilla Table II. Matrix of Loadings of the Thre eFactor Mode l for the Perce ive d Mode s of Processing Inventory Using Maximum Like lihood Con® rmatory Factor A nalysis Item #

RP

2 4 8 11 15 16 18 19 24 26 28 31 1 3 5 6 12 14 20 22 27 30 7 9 10 13 17 21 23 25 29 32

.510 .609 .707 .567 .754 .804 .818 .897 .830 .640 .626 .374 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

EP 0 0 0 0 0 0 0 0 0 0 0 0 .575 .448 .744 .524 .816 .759 .727 .486 .686 .738 0 0 0 0 0 0 0 0 0 0

AP 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 .577 .380 .649 .635 .682 .599 .458 .415 .660 .687

Notes. N 5 865; RP 5 Rational Proce ssing Scale ; E P 5 Experie ntial Processing Scale ; A P 5 A utomatic Processing Scale . E stimate d standard errors range from .032 to .041.

drawn; the re fore, the y have the same ge ne ral characte ristics (e .g., age , sex, e thnicity) . More ove r, the data colle ction proce dure s were the same as those use d with Sample s 1 and 2 (se e Part I). Criterio n Measu res an d Predictio ns The alphas re porte d for the crite rion me asure s de scribe d as follows are all from the cite d publications e xce pt for the me asure s of affectivity and psychologic al well-be ing, which are base d on the pre se nt data. To inve stigate conve rge nt and discrim inant corre lations with othe r measure s of perceive d inform ation processing, the pre se nt study e mploye d the Rational ± E xpe rie ntial Inve ntory (RE I) de ve lope d by E pste in e t al. (1996) , the Social Prob-

Pe rce ived Proce ssing

353 Table III. Interfactor Correlations in the Thre e -Factor Model for the Perceive d Mode s of Proce ssing Inventory RP RP EP AP

Ð

EP

AP

2 0.30

0.20 (0.04) 0.45 (0.03) Ð

(0.04) Ð

Notes. N 5 865; the e stimated standard e rrors are provide d in pare nthesis. RP 5 Rational Processing Scale; E P 5 E xperie ntial Proce ssing Scale ; A P 5 Automatic Proce ssing Scale .

lem-Solving Inve ntory ± Revise d (SPSI-R ; D’ Z urilla e t al., in pre ss; Mayde uO livare s & D’ Z urilla, 1996) , and three scales from the dispositiona l ve rsion of the CO PE (Carve r et al., 1989) . The RE I consists of two unipolar scale s, the Need for Cognition Scale and the Faith in Intuition Scale , which are de scribe d as me asuring rational proce ssing and e xpe rie ntial processing, re spe ctive ly. The alphas for the se scales are .87 ( Need for Cognition) and .77 (Faith in Intuition) . B ase d on the re sults found in Part I, we pre dicte d that the PMPI’ s Rational Proce ssing Scale would be correlate d with the Nee d for Cognition Scale , and that the PMPI’ s E motional Proce ssing and A utomatic Processing scale s would be positive ly re late d to Faith in Intuition. The SPSI-R is a 52-ite m que stionnaire consisting of ® ve scale s that measure two constructive proble m-solving dime nsions and three dysfunction al dime nsions. The constructive scale s are Positive Proble m O rie ntation and Rational Proble m Solving, which have alphas of .80 and .95, re spe ctive ly. The dysfunction al scale s are Negative Proble m O rientation, Impulsivity /Carele ssne ss Style , and A voidance Style . The alphas are .92, .89, and .89, re spe ctive ly. We pre dicte d that the Rational Proce ssing Scale of the PMPI would be positive ly relate d to both Positive Proble m O rie ntation (posse ssion of a proble m-solving cognitive set) and Rational Proble m Solving (se lf-re porte d use of constructive proble m-solving skills) . B ase d on some re se arch ® ndings that sugge st that proce ssing in the e xpe rie ntial mode may subve rt or disrupt rational thinking and proble m solving (D’ Z urilla & Chang, 1995; E pstein e t al., 1992) , we pre dicte d that both E motional Proce ssing and A utom atic Processing would be positive ly re late d to all three dysfunctiona l proble m-solving dime nsions measure d by the SPSI-R. The dispositional CO PE is a theoretically drive n multidime nsional coping inventory consisting of 13 scale s that measure a varie ty of diffe rent coping strategies and te chnique s. The pre sent study focuse d on thre e scale s that appe ar to be measuring aspe cts of perceive d rational proce ssing: Planning, Suppre ssion of Compe ting A ctivitie s (tendency to avoid proce ssing competing inform ation in orde r to conce ntrate more fully on the stre ssor), and Restraint Coping (te nde ncy to inhibit premature action until an appropriate opportunity occurs). The alphas for the se three

354

B u rns and D’Zurilla

scales are .80, .68, and .72, re spe ctive ly. We e xpe cted the Rational Processing Scale to be re late d to all thre e of these coping scale s. To measure affe ctivity, we use d the Positive and Negative A ffe ct Sche dule (PA NA S; Watson, Clark, & Telle gen, 1988) and a 7-ite m re duce d version of the A ffe ct Intensity Scale (Larsen & Die ner, 1987) . The PA NA S consists of a 10-ite m Positive A ffect Scale and a 10-ite m Negative A ffe ct Scale . The pre se nt study use d ``trait’ ’ rathe r than ``state’ ’ instructions; that is, subje cts were aske d to re port how the y gen erally fe e l. The alphas are .82 and .86 for Positive A ffe ctivity and Negative A ffe ctivity, re spective ly. The A ffe ct Intensity Scale asse sse s emotionality as a trait. The alpha for this scale is .72. O n the basis of E pste in’ s (1994) assum ption that the e xpe rie ntial syste m is intimate ly associate d with the e xpe rie nce of affe ct, we pre dicted that the E motional Processing Scale would be signi® cantly relate d to all thre e affe ctivity measure s. In contrast, Rational Proce ssing and A utom atic Processing were expe cted to have only nonsigni® cant or low corre lations with affectivity. Pe rsonality was asse sse d by the NE O -Five Factor Inve ntory ( NEO -FFI; Costa & McCrae , 1985, 1989) . This instrum ent consists of ® ve 12-ite m scale s that measure the following personality dime nsions: Neuroticism, E xtrove rsion, O penne ss, A gree able ne ss, and Conscie ntiousne ss. The alphas for the se scale s are .86, .77, .73, .68, and .81, respe ctively. A lthough it is possible that the mode s of proce ssing measure d by the PMPI might be tapping one of the se personality dimensions, we had no basis for making any spe ci® c pre dictions. Hence , we adopte d the null hypothe sis that none of the PMPI scale s is re late d to any particular pe rsonality dimension. Gene ral inte llige nce was measure d by the Wonde rlic Pe rsonne l Test (WPT; Wonde rlic, 1992) . The WPT is a 50-ite m, time d, short te st of gene ral inte llige nce that has bee n found to have high corre lations (be twe e n .92 and .96) with traditional IQ te sts such as the We chsler A dult Inte llige nce Scale (Dodrill, 1983; Hawkins, Faraone , Pe pple , & Seidman, 1990) . The alpha for this measure is .88. Theory and rese arch in the areas of practical proble m solving and constructive thinking (D’ Z urilla, 1986; E pste in & Meie r, 1989; Ste rnbe rg & Wagne r, 1986) have sugge sted that none of the proce ssing mode s tappe d by the PMPI are like ly to ove rlap signi® cantly with the traditional conce pt of gene ral inte llige nce . He nce , it was pre dicted that the PMPI scale s would have only nonsigni® cant correlations with the WPT. To asse ss situational coping strategies, we use d the 72-ite m Coping Strate gie s Inve ntory (CSI; Tobin e t al., 1989) , which use s a format similar to the Ways of Coping Checklist ( Folkman & Lazarus, 1980) . Subje cts are aske d to de scribe a re cent stressful e ve nt and the n report the e xte nt to which the y use d spe ci® c cognitive and behavioral coping strategie s on a 5-point Likert-type scale ranging from ``none ’ ’ to ``ve ry much.’ ’ B ase d on a hie rarchical factor analysis, the CSI consists of 14 scales that assess e ight primary coping dime nsions, four se condary dimensions, and two tertiary dimensions. The te rtiary dime nsions are engage me nt (approach) and disengage me nt (avoidance ) . The present study focuse d only on the se condary and primary dime nsions. The se condary scales with the ir alphas are Proble m E ngage ment (.87) , E motion E ngage ment (.92) , Proble m Dise ngage ment (.81) , and Emotion Dise ngage ment (.90) . The primary scale s within Proble m E ngage ment are Proble m

Pe rce ived Proce ssing

355

Solving (.82) and Cognitive Restructuring (.83) . Within E motion E ngage ment, the scales are E xpre ss Emotions (.89) and Social Support ( .89) . The scale s subsume d unde r Proble m Dise ngage ment are Proble m A voidance (.72) and Wishful Thinking (.78) . The scale s within E motion Dise ngage ment are Se lf-Criticism (.94) and Social Withdrawal (.81) . In ge neral, the dimensions of proble m e ngage ment and dise ngage ment corre spond to Lazarus and Folkman’ s (1984) conce pt of proble m-focuse d coping, whe re as e motion e ngage ment and dise ngage ment correspond to the ir e motion-focuse d coping. Previous re se arch has shown that engage me nt coping strate gie s te nd to be associate d with be tte r adaptation and well-be ing than dise ngage ment strategies ( Carve r et al., 1989; Tobin et al., 1989) . We hypothe size d that Rational Proce ssing would be re late d to the use of more adaptive proble m-focuse d coping, whe re as E motional Processing would be relate d to more e motion-foc use d coping. To asse ss psychologic al well-be ing, we use d thre e measure s of psychologic al distress (depression, anxie ty, and ange r) and two me asure s of positive well-be ing (se lf-e stee m and life satisfaction) . Spe ci® cally, the se measure s were the B e ck De pre ssion Inve ntory (B e ck, Ward, Me nde lson, Mock, & E rbaugh, 1961) , the Trait A nxie ty Scale from the State -Trait A nxie ty Inve ntory (Spie lbe rger, 1983) , the A nge r subscale from the E A SI-III Te mperame nt Surve y (B uss & Plomin, 1975) , the Rose nbe rg Se lf-E ste em Scale ( Rose nbe rg, 1965) , and the Satisfaction With Life Scale (Die ne r, E mmons, Larsen, & Griffe n, 1985) . The alphas for the se me asure s are .90 (de pre ssion) , .92 (anxie ty), .65 (ange r), .92 (se lf-e ste em), and .87 (life satisfaction ). We pre dicted that Rational Proce ssing would be re late d to less distre ss and gre ate r positive well-be ing, whe re as E motional Proce ssing and A utomatic Processing would be associate d with more distre ss and less positive well-be ing. R esu lts D escrip ti ve Statistics an d In terscale Correlatio ns The means, standard deviations, and inte rscale corre lations for Sam ple s 1 and 2 are pre se nted in Table IV . Note that the se statistics are fairly stable across sample s. More ove r, the inte rscale corre lations are ve ry similar to the inte rfactor corre lations

Table IV. Means, Standard De viations, and Interscale Corre lations for the Perce ive d Modes of Processing Inve ntory in Two Sample s RP RP EP AP M SD

2 0.22 0.19 37.33 8.65

EP

AP

M

SD

2 0.38

0.00 0.35

37.02 26.41 26.45

9.18 7.65 6.08

0.37 28.33 7.03

26.85 6.37

Notes. The re sults for Sample 1 (N 5 403) are above the diagonal, whereas the re sults for Sample 2 (N 5 865) are be low the diagonal; RP 5 Rational Proce ssing Scale ; E p 5 E xpe riential Processing Scale ; A p 5 Automatic Processing Scale.

356

B u rns and D’Zurilla

re porte d in Table III. Be cause the se two sample s were virtually ide ntical, the y were combine d into a total sample of 1268 subje cts (757 wome n and 511 men) for an e xam ination of ge nde r diffe re nce s in PMPI scale score s. Signi® cant ge nde r diffe re nce s were found on the E motional Proce ssing and A utom atic Proce ssing scale s. Me n (M 5 28.09) scored highe r than wome n (M 5 25.89) on A utom atic Proce ssing {t(1267) 5 6.42, p , .0001}, whe re as women (M 5 28.11) scored highe r than men (M 5 27.18) on E motional Processing {t(1267) 5 2.26, p , .05}. Reliab ility A nalyses Two type s of re liability were assessed: inte rnal consiste ncy and te st± re te st re liability. The forme r is a me asure of inte rite m homoge ne ity, and the latte r is a measure of te mporal stability. The coe f® cients for these two type s of re liability in Sample s 1 and 2 are pre se nted in Table V . With regard to inte rnal consiste ncy, the PMPI scale s were found to have good re liability in both sample s. In Sample 1, the Cronbach alphas for Rational Processing, E motional Proce ssing, and A utomatic Proce ssing were .90, .88, and .80, re spe ctive ly. In Sam ple 2, the alphas for the se scales were .88, .86, and .82, respectively. The te st± re te st coe f® cie nts in Table V are base d on a subse t of 147 subje cts from Sample 2. The se subje cts complete d the PMPI twice with an inte rval of two weeks. The coe f® cients for Rational Processing, E motional Proce ssing, and A utom atic Proce ssing were .56, .61, and .56, respective ly, indicating modest le ve ls of te mporal stability that are lower than would be e xpe cted for stable pe rsonality traits or inte lle ctual abilitie s. In the corre lational analyse s re porte d in the se ctions following, individual correlations with a signi® cance le vel gre ate r than .001 should be inte rpre te d with caution because of the numbe r of corre lations compute d be twe en the PMPI scale s and othe r measure s. The B onfe rroni adjustm ent was not used to de te rmine statistical signi® cance be cause we were more inte re ste d in showing the patte rn and consiste ncy of signi® cant re lations with e xte rnal variable s than the absolute le ve l of signi® cance of any individual corre lations. In all analyse s, gende r was partialle d out of all correlations base d on combine d sample s and se parate correlations for male s and fe male s were compare d to de te rmine if the re were any signi® cant diffe re nce s. Corre lations corre cted for atte nuation (in pare nthe se s) are reporte d in all table s

Tab le V . Internal Consistency and Te st± Re test Re liability Coef® cients for the Perce ive d Modes of Processing Inve ntory Scale RP EP AP

A lpha a

A lpha b

Test ± Re testc

.90 .88 .80

.88 .86 .82

.56 .61 .56

No tes. RP 5 Rational Processing Scale ; E P 5 E xperiential Processing Scale; A P 5 A utomatic Proce ssing Scale. a N 5 403. b N 5 865. c N 5 147.

Pe rce ived Proce ssing

357

in addition to the uncorre cted corre lations in orde r to permit comparisons among correlations that are not distorte d by diffe re nce s in re liability. Correlatio ns w ith O ther Measu res of Percei ved Ratio nal an d Exp eriential Pro cessing Corre lations betwe en the PMPI scale s and othe r measure s of pe rceive d rational and e xpe rie ntial processing are pre se nte d in Table V I. The patte rn of corre lations is consiste nt with some , but not all, of our pre dictions. A s e xpe cted, the Rational Proce ssing Scale was consiste ntly positive ly re late d to othe r rational proce ssing measure s and was unre late d to e xpe rie ntial proce ssing as measure d by the RE I. More ove r, Rational Proce ssing was found to be negative ly correlate d with impulsive /carele ss proble m solving. Consiste nt with e xpe ctations, the E motional Proce ssing Scale was positive ly re late d to e xpe rie ntial processing as measure d by the RE I and was found to have only ne gative or nonsigni® cant corre lations with measure s of rational proce ssing. In addition, as pre dicted, E motional Processing was positive ly re late d to all thre e dysfunction al proble m-solving dime nsions.

Table VI. Corre lations B e tween the Perceive d Modes of Proce ssing Inve ntory and O ther Me asures of Rational and E xpe rie ntial Processing Criterion

RP

EP

AP

.22*** (.24) 2 .04 ( 2 .05)

2 .28*** ( 2 .32) .27*** (.33)

2 .05 ( 2 .06) .25*** (.32)

.34*** (.40) .02) .02) .65*** (.70) .37*** .41) .13 .15)

2 .14 ( 2 .16) .33*** (.37) 2 .17* ( 2 .19) .49*** (.55) .33*** (.37)

.28*** (.35) .04 (.05) .15* (.17) .24*** (.28) .17* (.20)

.59*** (.70) .42*** (.54) .42*** (.52)

2 .09 ( 2 .11) .05 (.06) .07 (.09)

.02 (.03) .18** (.24) .15* (.20)

a

REI Ne ed for Cognition Faith in Intuition SPSI-R b Positive Problem O rie ntation Ne gative Proble m O rie ntation

(2 (2

Rational Proble m Solving Impulsivity/Care lessne ss Style A voidance Style CO PE c Planning Suppress Competing A ctivities Re straint Coping

2

(2

2

(2

Notes. RP 5 Rational Processing Scale; E P 5 E xpe riential Processing Scale ; A P 5 A utomatic Processing Scale; RE I 5 Rational± Expe rie ntial Inventory; SPSI-R 5 Social Problem-Solving Inve ntory± Re vise d; COPE 5 Carve r et al.’ s (1989) dispositional coping inventory. Gende r is partialled out of all corre lations; corre lations in parenthe ses are corre cted for atte nuation. *p , .05, **p , .01, ***p , .001. a N 5 315. b N 5 205. c N 5 351.

358

B u rns and D’Zurilla

A s e xpe cted, the A utomatic Proce ssing Scale was positive ly correlate d with e xpe rie ntial processing as me asure d by the RE I. In addition, like E motional Processing, A utom atic Proce ssing was positive ly re late d to two dysfunction al proble msolving measure s, name ly, Impulsivity /Care le ssne ss Style and A voidance Style . Une xpe ctedly, howe ver, this scale was also found to be positive ly relate d to seve ral measure s of rational processing, including Positive Proble m O rie ntation and Rational Proble m Solving. Two ge nde r diffe rence s were found in the data reporte d in Table V I and both were associate d with E motional Proce ssing. First, the positive correlation be twe e n E motional Proce ssing and Negative Proble m O rie ntation was signi® cant only for wome n (r 5 .43, p , .001) ; more ove r, it was signi® cantly diffe re nt ( p , .01) from the corre lation for men (r 5 .15) . Second, the ne gative corre lation betwe e n this scale and Rational Proble m Solving was signi® cant only for men (r 5 2 .41, p 5 .001) and it was signi® cantly diffe re nt (p , .01) than the correlation for wome n (r 5 2 .05) . Correlatio ns w ith A ffectivity, Personality, an d Intelligen ce Corre lations be twe e n the PMPI scale s and the crite rion measure s of affectivity, pe rsonality, and gene ral inte llige nce are pre se nte d in Table V II. A s the table shows, Tab le V II. Correlations B etwe en the Pe rce ived Modes of Processing Inventory and Affectivity, Pe rsonality, and Intelligence Criterion

RP

Affectivity Positive A ffectivity a Ne gative A ffectivity a Affect Intensity b Perso nality Ne uroticism c E xtrove rsion c Openne ss c Agre eable ne ss c Conscientiousc Intelligence IQ d

2

(2

2

(2

.19*** (.22) .02 .02) .03 .04)

EP .00 (.00) .18*** (.21) .21*** (.26)

AP .08 (.09) 2 .06 ( 2 .07) 2 .00 (.00)

2 .08 ( 2 .09) .08 (.10) .14 (.17) .08 (.10) .35*** (.41)

.14 (.16) .22** (.27) 2 .04 ( 2 .05) .04 (.05) .02 (.02)

2 .09 ( 2 .11) .15 (.19) .08 (.10) .09 (.12) .23** (.29)

.02 (.02)

.02 (.02)

.05 (.06)

No tes. RP 5 Rational Processing Scale ; E P 5 E xpe riential Processing Scale; A P 5 A utomatic Proce ssing Scale . Ge nder is partialled out of all correlations; corre lations in pare nthese s are corrected for atte nuation. *p , .05, **p , .01, ***p , .001. a N 5 351. b N 5 315. c N 5 177. d N 5 170.

Pe rce ived Proce ssing

359

our pre dictions concerning affectivity were large ly con® rmed, although the re we re a few notable exce ptions. Consiste nt with our pre dictions, E motional Proce ssing was positive ly correlate d with ne gative affe ctivity and affe ct inte nsity, but it was unre late d to positive affe ctivity. Une xpe ctedly, Rational Proce ssing was the only PMPI scale that was found to have a signi® cant positive corre lation with positive affe ctivity. A s pre dicte d, A utom atic Processing was unre late d to affe ctivity. With re gard to pe rsonality, E motional Proce ssing was found to be signi® cantly correlate d with E xtrove rsion, whe re as Rational Processing and A utomatic Proce ssing were both relate d to Conscie ntiousne ss. A s pre dicted, none of the PMPI scale s was signi® cantly relate d to IQ as measure d by the WPT, which supports the vie w that none of these mode s of proce ssing is relate d to the traditional construct of gene ral inte llige nce . O nly one ge nde r diffe re nce was found in the data re porte d in Table V II. The positive corre lation be twee n Rational Proce ssing and Positive A ffe ctivity was found to be signi® cant only for wome n (r 5 .31, p , .001) and it was signi® cantly diffe rent (p , .01) from the corre lation for men ( r 5 .02) . Co rrelatio ns w ith Situ atio nal Copin g Strategies Corre lations betwe en the PMPI scale s and the situational coping strate gie s measure d by the CSI are pre sente d in Table V III. In ge ne ral, the patte rn of corre lations is consiste nt with our pre dictions. A s the table shows, Rational Proce ssing was found to be most strongly relate d to adaptive proble m-focuse d coping (Proble m Solving and Cognitive Restructuring), although it was also correlate d with one disengage me nt strate gy (Wishful Thinking) . In contrast, E motional Proce ssing was most strongly relate d to adaptive e motion-focuse d (E xpre ss E motions and Social Support) . In addition, howeve r, this scale was also re late d to two dise ngage ment coping strategies (Proble m A voidance and Wishful Thinking) . Like Rational Processing, the A utom atic Processing Scale was most strongly re late d to adaptive proble m-focuse d coping (Proble m Solving and Cognitive Restructuring), although the correlations were somewhat lowe r. In addition, this scale was also found to be re late d to one adaptive e motion-focuse d strate gy (E xpre ss E motions) and one disengage me nt coping strate gy (Social Withdrawal) . No ge nde r diffe re nce s were found in these data. Correlatio ns w ith Psych olo gical Well-B eing Corre lations betwe e n the scales of the PMPI and the criterion measure s of psychologic al we ll-be ing are pre se nted in Table IX. The patte rn of corre lations is consiste nt with our initial predictions as well as the re sults re porte d pre viously conce rning the re lationship be twee n the PMPI and situational coping strate gie s. The re sults indicate that individuals who re porte d proce ssing more in the rational mode also re porte d le ss depre ssion and gre ate r se lf-e ste em and life satisfaction , whe re as those who re porte d proce ssing more in the e motional mode re porte d more psychologic al distre ss and less se lf-este e m. A utomatic Processing, howeve r, was found to have mixe d results. This scale was ne gative ly relate d to anxie ty but positive ly re late d to ange r. O nly one ge nde r diffe re nce was found in the data re porte d in Table IX. A l-

360

B u rns and D’Zurilla Tab le VIII. Corre lations B etwe en the Pe rce ived Modes of Proce ssing Inve ntory and Situational Coping Strategie s Criterion Prob lem E n gage m en t Proble m Solving Cognitive re structuring E mo tion E ngage me nt E xpress E motions Social Support Prob lem D ise ngagem e nt Proble m Avoidance Wishful Thinking E mo tion D isen gagem en t Se lf-Criticism Social Withdrawal

RP

EP

AP

.43*** (.49) .33*** (.39) .43*** (.49) .14 (.15) .12 (.13) .14 (.15) .19* (.22) .12 (.15) .20** (.24) .08 (.09) .03 (.03) .13 (.16)

.14 (.16) .14 (.16) .11 (.13) .40*** (.44) .37*** (.42) .33*** (.37) .22** (.26) .18* (.23) .18* (.22) .14 (.16) .13 (.14) .10 (.12)

.33*** (.39) .24** (.30) .34*** (.42) .18* (.21) .23*** (.27) .11 (.13) .14 (.17) .17 (.22) .08 (.10) .10 (.12) .01 (.01) .20** (.24)

No te. N 5 177; RP 5 Rational Processing Scale ; E P 5 E xperie ntial Proce ssing Scale; A P 5 A utomatic Proce ssing Scale. Ge nde r is partialled out of all correlations; corre lations in pare nthese s are corre cted for atte nuation. *p , .05, **p , .01, ***p , .001.

though the negative corre lation be twe e n Rational Processing and ange r was nonsigni® cant in the combine d sample , the corre lation for me n was signi® cant (r 5 2 .28, p , .01) ; more ove r, it was signi® cantly diffe rent (p , .001) from the correlation for wome n (r 5 .04, p . .05) .

G E NE R A L D ISCU SSION Mu ltip le Processin g Mode s The purpose of Part I of this study was to construct a new se lf-re port instrume nt that was spe ci® cally de signe d to me asure two diffe re nt pe rceived information processing style s in stre ss and coping situations, nam ely, rational processing and e xpe rie ntial proce ssing, which are base d on information -proce ssing constructs de scribe d in E pste in’ s (1994) cognitive -e xpe rie ntial self-the ory (CE ST). E xploratory and con® rmatory factor analyse s were use d to empirically inve stigate the factor structure of this instrum ent. Contrary to our hypothe sis, the re sults showe d that the instrum ent is measuring thre e diffe re nt, albe it relate d, pe rceive d

Pe rce ived Proce ssing

361

Table IX . Corre lations B etwe en the Pe rceived Mode s of Proce ssing Inventory and Psychological We ll-Be ing Criterion D istress Depression a A nxie tyb A nge rc Po sitive Well-Being Self-Este em d Life Statisfactione

RP

2 .16** ( 2 .18) 2 .11 ( 2 .12) 2 .09 ( 2 .12) .20*** (.22) .14** (.16)

EP .14** (.16) .15** (.17) .27*** (.37)

2 .24*** ( 2 .27) .04 (.05)

AP

2 .05 ( 2 .06) 2 .16** ( 2 .19) .18* (.25) .01 (.01) .06 (.07)

Note. RP 5 Rational Proce ssing Scale; E P 5 E xpe riential Processing Scale ; A P 5 A utomatic Proce ssing Scale . Gende r is partialled out of all correlations; correlations in pare nthese s are corre cte d for atte nuation. *p , .05, **p , .01, ***p , .001. a N 5 345. b N 5 343. c N 5 351. d N 5 350. e N 5 403.

processing modalitie s: ( 1) rational proce ssing, (2) e motional proce ssing, and (3) automatic processing. A s e xpe cted, the rational proce ssing factor corre sponds to E pste in’ s (1994) rational proce ssing construct. Howe ve r, the results indicate that our original e xpe rie ntial ite m se t is measuring two diffe rent factors that we have labe led e motional proce ssing and autom atic processing. This three factor structure was cross-valida ted in an inde pendent sample . Colle ctive ly, the scales that measure the se three factors were nam ed the Pe rceive d Mode s of Proce ssing Inve ntory ( PMPI) . The ® nding that the PMPI is measuring thre e partially inde pe nde nt factors indicate s that individuals may use more than one proce ssing modality on a re gular basis, e ithe r in the same situations or in diffe rent situations. The inte rcorre lations among the factors sugge st that people who re port more autom atic processing are somewhat more like ly to use rational proce ssing and e motional processing as well, whe re as those who report more rational proce ssing are some what less like ly to use e motional proce ssing (se e Table III). B ased on a conte nt analysis of the ite ms in e ach scale , individuals who score high on Rational Processing re port that the y spend much time and mental e ffort trying to ® gure out how to cope by using logical re asoning; creative thinking ( trying to ® nd ne w ways to cope) ; and spe ci® c proble m-solving methods, such as inform ation gathe ring, bre aking down a comple x proble m into more manage able parts, ge ne rating alte rnative solutions, and conside ring the possible conseque nces of diffe re nt solutions. Pe ople with high score s on E motional Proce ssing re port that the y re ly pre dominantly on their fee lings, e motions, vibe s, hunche s, and instincts to guide the ir coping e fforts. Individuals who score high

362

B u rns and D’Zurilla

on A utomatic Proce ssing re port that the y make coping de cisions quickly and e f® cie ntly ( minim al atte ntion, time , and e ffort) by re lying on the ir past coping e xpe rie nce s and the use of familiar, well-re he arse d coping re sponse s. There are two viable inte rpretations of the ® nding that our original e xpe rie ntial item se t is measuring two diffe re nt perceive d processing modalitie s. O ne inte rpre tation is that E pstein’ s (1994) e xpe rie ntial proce ssing is not a single processing syste m, but inste ad is be st viewed as two partially inde pendent processing modalitie s corre sponding to e motional proce ssing and automatic processing. The othe r inte rpre tation is that e motional proce ssing and automatic processing re pre se nt two partially inde pe nde nt compone nts of a broade r expe rie ntial processing construct that was not ade quate ly re pre se nted by our original e xpe rie ntial ite m se t. B ecause our data cannot de te rmine which of the se inte rpre tations is corre ct, this is an important issue for future research. A utomatic processing (or ``autom aticity’ ’ ) has ofte n bee n described in the lite rature as be ing fast and e f® cient, outside of aware ne ss, uninte ntional, and uncontrolle d (B argh, 1994; Smith, 1994; Shiffrin & Schne ide r, 1977) . A ccording to Logan (1988) , the basic me chanism in this proce ssing modality is automatic memory re trie val. Pe rformance is base d on the dire ct-access, single -step re trie val of past re sponse s from me mory. Give n this de ® nition, one might que stion how a se lf-re port que stionnaire can possibly be measuring percei ved automatic processing. Howeve r, theory and re search in recent ye ars has made it incre asingly cle ar that automatic processing rare ly ® ts a strict all-or-none de ® nition consisting of all of the pre viously mentione d crite ria (B argh, 1989, 1994; Smith, 1994) . For e xam ple , a pe rson may not only be aware of the particular stimulus or cue that se ts the occasion for automatic proce ssing, but he or she may also have a conscious inte ntion or proce ssing goal. B argh (1989) has name d this partially conscious and controlle d processing syste m go al-d ependent au tom aticity. In this form of autom atic proce ssing, furthe r conscious control may be accomplishe d by manipulating re trie val cue s (e .g., focusing on the most rele vant inform ation, changing goals) and /or by controlling the response s syste m (Logan, 1989; Logan & Cowan, 1984) . For e xam ple, the pe rson may inhibit imme diate re sponding base d on memory retrie val and scree n or e dit pote ntial re sponse s be fore ® nally exe cuting a response . A ccording to Smith (1994) , the only prope rtie s that are central to all forms of autom aticity are probably spe ed and e f® ciency. A person is like ly to be aware of these prope rties in goal-de pe nde nt automaticity, whe re he or she is alre ady attending to a particular stimulus and proce ssing goal. In the instruction s for the PMPI, individuals are aske d to focus on a particular type of stimulus situation (i.e ., stre ssful situations) and a particular goal (i.e., to cope e ffe ctive ly with stre ss). In addition to these conditions, the ite ms in the A utom atic Processing Scale focus on the qualitie s of spee d, e f® ciency, and the imme diate knowing of how to cope base d on past coping e xpe rience s. Howe ve r, the content of the se ite ms sugge st ve ry little , if any, conscious e valuation or control of the re sponse syste m. Hence, the A utomatic Proce ssing Scale of the PMPI may be vie wed as tapping a pe rceive d proce ssing construct that is linke d to a goal-de pende nt, se micontrolle d form of autom aticity.

Pe rce ived Proce ssing

363

R e liab ility an d V alid ity of the PMPI The purpose of Part II of this study was to conduct a pre liminary e valuation of the reliability and validity of the PMPI. A ll three scales of the PMPI were found to have high inte rnal consiste ncy in two diffe re nt sample s and mode rate le ve ls of te mporal stability in one sample . In ge neral, the patte rn of conve rge nt and discrim inant corre lations with e xternal variable s provide s support for the construct validity of the PMPI as well as the hypothe sis that perceive d rational proce ssing is most strongly and consiste ntly relate d to adaptive proble m-focuse d coping and positive psychologic al we ll-be ing. Individuals with highe r score s on the Rational Proce ssing Scale reporte d a more positive orie ntation toward proble ms in living, more conscie ntiousne ss, more planning and de libe ration, be tte r proble m-solving skills, less impulsive /care le ss proble m solving, and more adaptive proble m-focused coping in a spe ci® c stre ssful situation. Consiste nt with the me ntione d re sults, these individuals also re porte d le ss psychologic al distre ss and more positive psychologic al well-be ing. Une xpe ctedly, pe rceive d rational processing was also found to be re late d to one pote ntially maladaptive coping strate gy Ð name ly, wishful thinking. This ® nding might re ¯ e ct a te nde ncy for individuals who pe rceive themse lve s as proce ssing more in this mode to be ove rly optimistic. The Rational Processing Scale was found to be highly correlate d with the Rational Proble m Solving Scale of the Social Proble m-Solving Inve ntory± Revised (se e Table V I), which has pre viously bee n found to be relate d to a measure of ``naive ’ ’ (ove rge ne ralize d) optim ism, which may have adaptive as well as maladaptive effe cts (D’ Z urilla & Chang, 1995) . In contrast with the results for Rational Proce ssing, individuals with highe r scores on the E motional Proce ssing Scale re porte d a more ne gative orie ntation toward proble ms in living, more impulsive /carele ss proble m solving, more proble msolving avoidance , more de ® cits in proble m-solving skills, and more maladaptive proble m-focuse d coping (proble m avoidance and wishful thinking) . In addition, the se individuals also re porte d more affe ct inte nsity, more psychologic al distre ss, and le ss positive psychologic al we ll-be ing. O n the positive side , the se individuals also re porte d more e xtrave rsion and more adaptive e motion-foc used coping (e xpre ssing e motions and se e king social support) . The se re sults sugge st that people who re ly on e motional proce ssing are more outgoing and emotional individuals who te nd to cope with stre ss by using catharsis and se eking out supportive inte rpersonal re lationships. Compare d to pe ople who re ly more on rational proce ssing, these individuals may have closer e motional re lationships, as well as a coping advantage in some stre ssful situations that are highly e motional and /or unchange able . Howe ve r, the y may have a disadvantage in proble matic situations that are more manage able or controllable , thus resulting in more psychologica l distre ss and le ss positive psychologic al we ll-be ing in general. Individuals with highe r score s on the A utomatic Processing Scale were found to be similar in some respe cts to pe ople who re porte d more rational proce ssing, and similar in othe r re spe cts to those who re porte d more e motional proce ssing. Like rational proce ssors, these individuals re porte d more conscie ntiousne ss, a more positive orie ntation toward proble ms in living, bette r proble m-solving skills, more

364

B u rns and D’Zurilla

adaptive proble m-focuse d coping, and less anxie ty. Like e motional proce ssors, these pe ople also re porte d more adaptive e motion-focuse d coping ( expre ssing emotions) . Howe ve r, the y were also similar in re porting more impulsive /care le ss proble m solving, avoidance of proble m solving, social withdrawal, and ange r. The ® nding that pe rceive d autom atic processing has mixe d adaptational outcome s is not surprising. More than any othe r mode of processing, the coping outcom e s of autom atic processing are like ly to vary, de pending on the type of situation and the nature of the pe rson’ s coping history. In familiar situations, individuals who re ly on this mode of processing will probably continue to cope adaptive ly and e f® cie ntly by using the same coping strate gie s (both proble m focused and e motion focuse d) that have worke d well for the m in the past. In more nove l, ambiguous, and /or comple x situations, howeve r, autom atic proce ssing alone (base d on familiar cue s in the situation) is not like ly to produce adaptive outcome s consiste ntly be cause old coping habits will not always be appropriate or e ffe ctive in the se situations. The more ve rsatile and adaptive individuals probably switch to e ithe r rational proce ssing or emotional proce ssing in these situations. Those who do not are like ly to have more unre solve d proble ms and possibly e xpe rie nce more frustration and ange r as a re sult. The issue of situational in¯ uence s on the re lative dominance or frequency of the thre e processing mode s is discusse d in the following paragraphs. Signi® cant ge nde r diffe re nces were found in mean score s on two PMPI scale s. Me n score d highe r than women on A utom atic Proce ssing, where as wome n score d highe r than men on E motional Proce ssing. These results sugge st that men are more like ly than wome n to respond quickly and automatica lly with ready coping re sponse s base d on similar past coping e xpe rie nce s, whe re as wome n re ly more on the ir fee lings, hunche s, and instincts to guide and direct the ir coping. Despite the se mean diffe re nce s, only a fe w gender diffe rence s were found in the corre lations be twee n PMPI scale s and e xternal variable s. Two of these gender diffe re nce s were associate d with the re lationship be twee n pe rceive d e motional proce ssing and proble m-solving ability. E motional processing was relate d to a more negative orie ntation toward proble ms in living in wome n, but not in me n. Howeve r, e motional proce ssing was relate d to de ® cits in proble msolving skills only in men. Despite the spe ci® c ge nde r diffe re nce s, the se ® ndings sugge st that e motional proce ssing is relate d to lower proble m-solving ability in both men and wome n. Two additional ge nde r diffe rence s were associate d with the relationship be twee n pe rceive d rational proce ssing and e motionality. Rational processing was associate d with more positive affe ctivity only in wome n and with less ange r only in me n. The similarity in the se re sults is that rational proce ssing se e ms to be associate d with more favorable e motionality in both men and wome n. The PMPI V ersus Sim ilar E xistin g Me asu re s Whe n introducing a ne w me asuring instrume nt, inve stigators should conside r whe ther it re pre se nts a contribution be yond e xisting measure s of the same or similar constructs or is mere ly redundant. The only othe r e xisting se lf-re port instrum ent that was spe ci® cally designe d to asse ss E pste in’ s (1990, 1994) conce pts of rational

Pe rce ived Proce ssing

365

and e xpe riential proce ssing is the Rational± E xpe rie ntial Inve ntory (RE I; E pstein e t al., 1996) . The two major scales of the RE IÐ Nee d for Cognition Scale and Faith in Intuition Ð are assum e d to measure perceive d rational processing and pe rceive d e xpe rie ntial processing, respectively. Factor analyse s of the RE I in two diffe re nt sample s provide d support for a two-factor structure corre sponding to the se two processing modalitie s. A n important format diffe re nce betwe en the PMPI and the RE I is that the forme r instrum e nt focuses spe ci® cally on the context of coping with stre ss, whereas the latte r cove rs a more ge neral domain. This diffe re nce could change the nature of the constructs be ing tappe d by these two inve ntorie s. In addition, diffe re nt factor ± analytic methods we re use d in the deve lopm e nt of the se instrume nts, which could account for the diffe re nt factor structure s that were found. The prese nt study use d unweighte d least square s e xploratory factor analysis and maximum like lihood con® rmatory factor analysis, whereas E pstein et al. (1996) use d principal compone nts e xploratory factor analysis. A n inspe ction of the signi® cant but re lative ly low corre lations be twee n the two RE I scale s and the thre e PMPI scales (se e Table V I) indicate that the se two instrum ents are measuring constructs that are similar or ove rlapping, but not re dundant. In support of this vie w, the PMPI’ s Rational Processing scale has a much highe r corre lation with the Rational Proble m Solving Scale of the SPSI-R (D’ Z urilla et al., in pre ss) than with the RE I’ s Nee d for Cognition Scale , which sugge sts that the PMPI’ s rational proce ssing construct focuse s more on the pe rceive d use of constructive proble m-solving skills. A n inspe ction of the ite ms in the Rational Proce ssing Scale (se e A ppe ndix) supports this inte rpre tation. In addition, the signi® cant but re lative ly low positive corre lations be twe e n the PMPI’ s E motional Proce ssing and A utom atic Proce ssing scales on the one hand, and the RE I’ s Faith in Intuition scale on the othe r, indicate that e motional proce ssing and autom atic processing both ove rlap with E pstein et al.’ s ( 1996) e xpe rie ntial proce ssing construct, but that ne ithe r one is e quivale nt to this construct. Despite the re lative ly low correlations be twe e n corresponding RE I and PMPI scales, se ve ral ® ndings for the Rational Processing and E motional Processing scales regarding the ir re lations with coping and adaptation are consiste nt with re sults re porte d for the Nee d for Cognition and Faith in Intuition scale s. B ased on the ir ® ndings, E pste in e t al. (1996) conclude d that individuals with highe r scores on Nee d for Cognition were be tter proble m-focused cope rs and be tte radjuste d people in ge ne ral than those with highe r score s on Faith in Intuition. In addition, the pre se nt ® nding that the E motional Proce ssing Scale is relate d to the use of adaptive e motion-foc use d coping (e xpre ssing e motions and se e king social support) is consiste nt with E pste in e t al.’ s (1996) ® ndings that individuals with highe r score s on the Faith in Intuition Scale were more like ly to report having close emotional re lationships. Howe ve r, these same two scale s we re found to have a diffe re nt patte rn of correlations with se ve ral me asure s of psychologica l well-be ing that supports the vie w that the y are measuring constructs that have important diffe re nces as well as similaritie s. E motional Proce ssing was found to be ne gative ly correlate d with se lfe stee m and positive ly corre late d with anxie ty and de pre ssion, whereas Faith in

366

B u rns and D’Zurilla

Intuition was found to be positive ly corre late d with se lf-e ste em and ne gative ly correlate d with anxie ty and de pre ssion. A nothe r appare nt diffe re nce be twe en the PMPI and the RE I is that the Nee d for Cognition Scale of the RE I and the Rational Proce ssing Scale of the PMPI se em to be measuring constructs that have diffe re nt relations with ge neral inte llige nce in colle ge stude nts. Epste in e t al. (1996) re porte d that the Need for Cognition Scale was signi® cantly corre late d with both SA T score s and grade point ave rage (E pste in e t al., 1996) , where as the Rational Processing Scale was found to be unre late d to gene ral inte llige nce as measure d by the Wonde rlic Pe rsonne l Te st. This diffe re nce may once again re ¯ ect the highe r corre lation that the Rational Proce ssing Scale has with the Rational Proble m Solving Scale of the SPSI-R , which has be e n found to be unre late d to SA T score s and grade point ave rage in colle ge stude nts (D’ Z urilla e t al., in pre ss). It is inte resting and pote ntially signi® cant that all of the ite ms in the Rational Proce ssing and Rational Proble m Solving scale s are state d in positive , constructive te rms (e .g., ``I usually put a lot of mental e ffort into ® guring out what is the be st thing to do,’ ’ ``Whe n I am trying to solve a proble m, I think of as many options as possible until I cannot come up with any more ide as’ ’ ) , whe reas most of the ite ms in the Nee d for Cognition Scale are state d in negative , dysfunction al te rms (e .g., ``The notion of thinking abstractly is not appe aling to me,’ ’ ``I don’ t re ason well unde r pre ssure’ ’ ) . It is possible that a ne gative or dysfunctiona l thinking style dime nsion may be more highly relate d to SA T score s and grade s than a positive or constructive thinking style dime nsion. In support of this view, the Impulsivity / Carele ssne ss Style Scale of the SPSI-R, which measure s dysfunction al proble msolving thinking, has bee n found to be signi® cantly ne gative ly corre late d with both SA T score s and grade point ave rage (D’ Z urilla e t al., in press). D isp ositio nal V e rsus Situ atio nal Processin g The dispositiona l format of the PMPI is base d on the assum ption in CEST that there are important individual diffe re nces in the exte nt to which pe ople re ly on diffe re nt inform ation processing modalitie s across a range of life situations (E pste in, 1994) . Howeve r, CE ST also assum es that situational variable s in¯ ue nce the re lative dominance of the diffe re nt processing modalitie s. For e xam ple , automatic proce ssing might be most common in familiar, routine situations; e motional processing might occur most ofte n in highly emotionally situations (e .g., fear, love , ange r), and rational proce ssing might be most like ly to occur in situations that are nove l, ambiguous, or comple x. The de gre e of ge ne rality vs. situational spe ci® city of the proce ssing modalitie s is an important issue for future re search be cause the utility or adaptive ne ss of certain proce ssing mode s may de pe nd signi® cantly on the type of situation. The PMPI can easily be adapte d for use in this re se arch by changing the instructions and the wording of the ite ms from a dispositional format to a situational format. It is note worthy that the ge ne rality vs. situational spe ci® city issue is still an unre solve d and controve rsial one in ® e ld of stre ss and coping in ge ne ral. A lthough the re is e vide nce for the existe nce of re lative ly stable coping style s or prefe re nce s

Pe rce ived Proce ssing

367

(Carve r e t al., 1989) , the re is also conside rable evide nce supporting the situational spe ci® city of coping strate gie s and te chnique s (A ldwin, 1994) . The ove rall ® ndings have led A ldwin ( 1994) to conclude that individual diffe rence s and situational variable s both in¯ uence coping, but that the re lative contributio ns of these two factors have ye t to be dete rmine d. Conclusio ns, Lim itatio ns, an d Cave ats In conclusion , the PMPI is a re liable , practical, cost-e ffective me asuring instrument that has good pote ntial applie d utility for research on stre ss and coping, psychopatho logy, and psychothe rapy. A s a measure of three pe rceived inform ation processing style s in stre ssful life situations, the PMPI may he lp to (1) improve pre dictions of adaptive coping and positive well-be ing, (2) unde rstand the role of information -processing style s in psychopatho logy, (3) ide ntify appropriate targe ts for treatme nt (e .g., de ® cits in rational proce ssing, ove rgene ralize d use of automatic processing) , ( 4) match clie nts to appropriate tre atme nts (e .g., cognitive therapy vs. e xpe rie ntial or inte rpersonal the rapy) , and (5) evaluate the outcome of spe ci® c tre atments (e .g., incre ase in rational proce ssing following cognitive the rapy, incre ase in e motional processing following e xpe rie ntial the rapy) . The pre se nt study has se veral limitations, howe ver, that sugge st that some caution is nee de d whe n inte rpre ting the ® ndings for the PMPI at this stage in its de ve lopme nt. The se limitations are re late d to: (1) the se lf-re port format of the PMPI, (2) the exclusive use of colle ge stude nt sample s, and (3) the e xclusive use of self-report crite rion measure s to e valuate the validity of the PMPI. A s a self-report instrume nt, the PMPI has the same pote ntial thre ats to validity as any othe r se lf-re port me asure (e .g., faking, e xpe ctancy e ffects, forge tting, re call biase s, compre he nsion proble ms). With speci® c re fere nce to the asse ssment of information proce ssing, a particularly important limitation of se lf-re port me thods is that they cannot be use d to dire ctly assess cognitive processes that are outside of aware ne ss. The se me thods are limite d to the prope rtie s and products of cognitive processes that are accessible to aware ness and, he nce , to self-report (Ingram & Kendall, 1986) . B e cause the ope rations in information processing are not e ntire ly accessible to se lf-re port, e spe cially in autom atic processing, the PMPI should not be vie we d as a measure of informatio n processing per se. Instead, the PMPI is pre se nted he re as a measure of pe rceived inform ation-proce ssing style s. The items asse ss an individual’ s pe rception of particular prope rtie s of thre e diffe re nt information-proce ssing modalitie s that are acce ssible to aware ne ss. The scale score s are assum ed to re ¯ e ct the exte nt to which individuals re ly on the diffe re nt mode s of processing whe n coping with stress, and nothing more. Hence , the PMPI cannot be used to test mode ls of information proce ssing, nor to study how diffe rent forms of information proce ssing work. Howe ver, this instrum e nt appe ars to have good pote ntial utility for the purpose s de scribe d pre viously. Neve rthe le ss, as part of the continuing evaluation of the construct validity of the PMPI, re search is nee de d that e xam ine s the re lations be twe en the PMPI scale s and more dire ct me asure s of information proce ssing unde r stre ss. The rigorous e mpirical me thods deve lope d in cognitive psychology and social cognition for studying the parame te rs of information

368

B u rns and D’Zurilla

processing might be use ful for this re se arch (se e B argh, 1994; Ingram & Kendall, 1986; Smith, 1994) . B e cause of the exclusive use of colle ge stude nt sample s in this study, caution is ne ede d whe n gene ralizing from the ® ndings to othe r populations until additional studie s are done with diffe rent subje cts from norm al as well as clinical populations. With re gard to the e xclusive use of se lf-re port measure s to e xam ine crite rion validity, we attempted to minimize the common thre ats to validity associate d with se lfre port by including proce dure s that prote cted the subje cts’ anonymity and ensure d the con® de ntiality of the ir te st re sponse s. In addition, we also use d the Coping Strate gie s Inve ntory (Tobin e t al., 1989) to e xamine the re lations betwe e n the PMPI scales and situational coping strate gie s. This instrum ent has a short-te rm re call format in which subje cts are aske d to de scribe a re cent stre ssful situation and then re port what speci® c coping strate gie s the y actually use d in that situation. Compare d to more traditional re trospe ctive formats, the proble ms of forge tting and recall biase s are like ly to be reduce d with this approach. Neve rthe le ss, howe ve r, additional studie s are now ne ede d that use non-se lf-report crite rion measure s to e xam ine e xte rnal validity.

A PPE ND IX A Ite ms of the Pe rceive d Mode s of Proce ssing Inve ntory (PMPI) . Ratio nal Pro cessing (RP) Scale (12 item s) 2. I often think about my stressful situations and the n try to ® nd ne w ways to re solve the m. 4. I usually try to cope with a stre ssful situation by breaking it down into smalle r parts and de aling with the m one at a time . 8. I usually think of as many alte rnative ways of coping as possible be fore I de cide what I am going to do. 11. B e fore trying to cope , I usually decide on a spe ci® c goal so that I know e xactly what I should try to do. 15. I usually try to ge t all the facts that I can be fore deciding how to cope . 16. I usually set aside e nough time to think things through care fully and ® gure out what is the be st thing to do. 18. Instead of acting on the ® rst idea that come s to mind, I care fully conside r all my options. 19. B e fore I atte mpt to cope , I think of all my options and care fully conside r the pros and cons of e ach one . 24. Whe n I am atte mpting to cope , one of the ® rst things I do is gathe r as many facts about the situation as possible so that I will be able to unde rstand what it is all about. 26. I usually put a lot of mental e ffort into ® guring out what is the be st thing to do. 28. I usually stick to the ``facts’ ’ and try to use a logical approach to cope . 31. For me, de ciding how to cope take s a lot of time and mental effort.

Pe rce ived Proce ssing

369

E m otio nal Processing (EP) Scale (10 item s) 1. To cope , I usually go with my instincts rathe r than trying to re ason things out. 3. My fee lings usually de te rmine how I will cope . 5. Whe n I am trying to de cide how to cope , I usually go with my ``gut’ ’ fe e ling. 6. Whe n I am atte mpting to cope , I de pe nd a gre at de al on my fee lings to he lp me ® nd the be st way to cope . 12. ``Gut’ ’ fee lings are more important to me than logic and evide nce whe n I have to cope . 14. Rathe r than spe nd my time trying to think of how to cope , I pre fe r to use my e motional hunche s. 20. E motions are usually more use ful than thoughts for coping. 22. I usually do what fe e ls right. 27. I trust my emotions to guide how I should cope . 30. Whe n I am atte mpting to cope I can usually trust my ``gut’ ’ fe e lings to te ll me what to do. A utom atic Processing (A P) Scale (10 item s) 7. I am ofte n aware of how to cope with a stre ssful situation e ven be fore I review all its aspe cts. 9. If an approach works I use it again and again so I don’ t have to come up with a ne w one for e ach stre ssful situation I face . 10. I’ ve had e nough e xpe rie nce to just know what I ne ed to do to cope most of the time without trying to ® gure it out e ve ry time . 13. The right way to cope usually come s to mind almost immediate ly. 17. I typically ® gure out the way to cope swiftly. 21. I quickly do the right thing whe n coping be cause I’ ve ofte n face d almost the same thing be fore . 23. Most of the time , I use the same method to cope . 25. I re ly mostly on my past expe rie nce to ® nd a way to cope . 29. I rare ly ne e d to mull things ove r; how to cope usually be come s quickly appare nt. 32. Whe n a stre ssful situation occurs I know right away what I ne e d to do to cope with it.

R E FE R E NCES A ldwin, C. M. (1994) . Stress, coping, and de velo pm en t: An in tegrati ve app ro ach . Ne w York: Guilford. B argh, J. A . (1989) . Conditional automaticity: Varie ties of automatic in¯ uence in social perce ption and cognition. In J. S. Uleman & J. A . B argh (E ds.), Un inten ded tho ught (pp. 3-51) . Ne w York: Guilford. B argh, J. A . (1994) . The four hourse me n of automaticity: A ware ne ss, intention, e f® cie ncy, and control in social cognition. In R. S. Wye r, Jr. & T. K. Srull (Eds.), Hand bo ok o f social co gn ition (2nd. e d.) (pp. 1-41) . Hillsdale, NJ: Lawrence E rlbaum. B e ck, A . T., W ard, C. H., Me nde lson, M., Mock, L., & E rbaugh, J. (1961) . A n inve ntory for me asuring depression. Archi ves of G eneral Psych iatry, 4, 561-571. B e ntler, P. M. (1990) . Comparative ® t indexes in structural models. Psych olo gical B u lletin, 107, 238-246. B ollen, K. A ., & Long, J. S. (E ds.). (1993) . Testing stru ctural equ ation m odels. Ne wbury Park, CA : Sage.

370

B u rns and D’Zurilla

B rewin, C. R. (1989) . Cognitive change processe s in psychothe rapy. Psycholo gical Re view , 96, 379-394. B rowne, M. W., & Cudeck, R. (1989) . Single sample cross-validation indices for covariance structure s. Multi variate B eha vio ral Research , 24, 445-455. B rowne, M. W., & Cudeck, R. (1993) . A lternative ways of assessing model ® t. In K. A . B ollen & J. S. Long, (E ds.), Testin g structural equ ation m odels (pp. 136-162) . Ne wbury Park, CA : Sage . B uss, A. H., & Plomin, R. (1975) . A tem p eram en t th eo ry of p erso nal de velo pm ent. New York: Wiley. Cacioppo, J. T., & Petty, R. E . (1982) . The ne ed for cognition. Jo urnal o f Perso nality an d Social Psycho logy, 42, 116-131. Carve r, C. S., Sche ier, M. F., & We intraub, J. (1989) . Asse ssing coping strategie s: A theoretically based approach. Jou rnal o f Personality and Social Psych olo gy, 56, 267-283. Costa, P. T., Jr., & McCrae, R. R. (1985) . Neo Five Factor-In ven tory: Form S. O de ssa, FL: Psychological A ssessmen t Re source s. Costa, P. T., Jr., & McCrae , R. R. (1989) . Neo PI /FFI Manu al Sup plem en t. O de ssa, FL: Psychological A ssessmen t Re source s. Diener, E ., E mmons, R. A ., Larsen, R. J., & Grif® n, S. (1985) . The Satisfaction With Life Scale. Jo urnal o f Personali ty Assessm ent, 49, 71-75. Dodrill, C. B . (1983) . Long-te rm reliability of the Wonde rlic Pe rsonne l Test. Jou rn al o f Co nsulting an d Clinical Psych olo gy, 51, 316-317. D’ Z urilla, T. J. (1986) . Pro blem -solving therap y: A social com petence app roach to clinical inter ven tion. Ne w York: Springer. D’ Z urilla, T. J., & Chang, E . C. (1995) . The re lations be twee n social problem solving and coping. Co gn itive Therap y and Research, 19, 547-562. D’ Z urilla, T. J., & Goldfried, M. R. (1971) . Proble m solving and be havior modi® cation. Jo urn al of A b norm al Psych olo gy, 78, 104-126. D’ Z urilla, T. J., & Mayde u-O livare s, A. (1995) . Conceptual and me thodological issues in social problemsolving asse ssme nt. B eh avior Therap y, 26, 409-432. D’ Z urilla, T. J., & Ne zu, A . M. (1990) . Developme nt and pre liminary evaluation of the Social Proble mSolving Inve ntory. Psycho logical A ssessm en t: A Jou rn al o f Con sulting an d Clinical Psych olo gy, 2, 156-163. D’ Z urilla, T. J., Nezu, A. M., & Mayde u-Olivares, A. (in pre ss). Manu al for th e So cial Pro blem -So lvin g In vento ry-Re vised. North Tonawanda, NY: Multi-Health Systems. E pstein, S. (1990) . Cognitive ± e xperie ntial se lf-theory. In L. Pervin (E ds.), Hand bo ok of perso nality theory an d research (pp. 165-192) . Ne w York: Guilford Press. E pstein, S. (1994) . Integration of the cognitive and the psychodynamic unconscious. A merican Psych olo gist, 49, 709-724. E pstein, S., Lipson, A., Holstein, C., & Huh, E . (1992) . Irrational reactions to ne gative outcome s: E vide nce for two conceptual syste ms. Jou rn al of Perso nality and So cial Psycho logy, 62, 328-339. E pstein, S., & Meier, P. (1989) . Constructive thinking: A broad coping variable with spe ci® c compone nts. Jou rnal o f Personality and Social Psych olo gy, 57, 332-350. E pstein, S., Pacini, R., Denes-R aj, V ., & He ier, H. (1996) . Individual differences in intuitive ± expe rie ntial and analytical ± rational thinking style s. Jou rnal o f Personality and Social Psych olo gy, 2, 390-405. Folkman, S., & Lazarus, R. S. (1980) . A n analysis of coping in a middle-aged community sample . Jo urnal o f Health and Social B eh avior, 21, 219-239. Hawkins, K. A ., Faraone , S. V ., Pepple , J. R., & Seidman, L. J. (1990) . WA IS-R validation of the Wonderlic Pe rsonnel Te st as a brief intelligence me asure in a psychiatric sample. Psych olo gical A ssessm ent, 2, 198-201. Ingram, R. E . (E d.) (1986) . In form ation p ro cessin g app roach es to clinical p sycho logy. New York: Academic Pre ss. Ingram, R. E ., & Ke ndall, P. C. (1986) . Cognitive clinical psychology: Implications of an information processing perspective. In R. E . Ingram (E d.), In form ation pro cessin g app roach es to clinical psych olo gy (pp. 3-21) . Ne w York: A cademic Pre ss. JoÈ re skog, K. G., & SoÈ rbom, D. (1985) . L ISRE L VI: A n alysis of linear structu ral relatio nsh ips by m axim um likelihoo d, instru mental variab les, an d least sq uares. Uppsala: University of Uppsala. JoÈ re skog, K. G., & SoÈrbom, D. (1993) . L ISRE L 8 User’ s reference gu ide. Chicago, IL: Scie nti® c Software. Larse n, R. J., & Die ne r, E . (1987) . A ffect intensity as an individual difference characte ristic: A re vie w. Jou rnal o f Research in Personality, 21, 1-39. Lazarus, R. S., & Folkman, S. (1984) . Stress, ap praisal, an d co pin g. New York: Springe r. Loehlin, J. C. (1987) . L atent variable m od els: A n in trodu ctio n to facto r, path, and stru ctural an alysis. Hillsdale, NJ: E rlbaum. Logan, G. D. (1988) . Toward an instance theory of automatization. Psycho logical Re view , 95, 492-527.

Pe rce ived Proce ssing

371

Logan, G. D. (1989) . A utomaticity and cognitive control. In J. A . B argh & J. S. Ule man (E ds.), Uninten ded thou gh t (pp. 52-74) . Ne w York: Guilford. Logan, G. D., & Cowan, W . B . (1984). O n the ability to inhibit thought and action: A theory of an act of control. Psycho lo gical Re view , 91, 295-327. Maydeu-O livare s, A ., & D’ Z urilla, T. J. (1996) . A factor analysis of the Social Proble m-Solving Inventory: A n integration of theory and data. Cogniti ve Th erapy and Research , 20, 115-133. McDonald, R. P. & Marsh, H. W . (1990) . Choosing a multivariate model: Nonce ntrality and goodness of ® t. Psych olo gical B u lletin, 107, 247-255. Pe arlin, L. I., & Schooler, E . L. (1978) . The structure of coping. Jou rnal o f Health an d So cial B eha vior, 19, 2-21. Rose nberg, M. (1965) . So ciety an d th e ado lescen t self-image. Prince ton, NJ: Princeton Unive rsity Pre ss. Shiffrin, R. M., & Schneide r, W. (1977) . Controlled and automatic human information proce ssing: II. Perceptual learning, automatic atte nding, and a general theory. Psycho logical Review , 84, 127-190. Smith, E . R. (1994). Procedural knowledge and proce ssing strate gies in social cognition. In R. S. Wye r, Jr. & T. K. Srull (E ds.), Han dbo ok of social cognition (2nd e d.) (pp. 99-151). Hillsdale, NJ: Lawre nce E rlbaum. Spielberge r, C. D. (1983). Manu al for th e State-Trait A nxiety In vento ry (re v. ed.), Palo A lto, CA: Consulting Psychologists Pre ss. Steiger, J. H. (1990) . Structural model e valuation and modi® cation: A n interval e stimation approach. Multi variate B eha vio ral Research , 25, 173-180. Sternberg, R. J., & Wagne r, R. K. (1986) . Practical in telligence: Nature and o rigins of com peten ce in the e veryd ay w o rld. Cambridge , E ngland: Cambridge University Pre ss. Tanaka, J. S. (1993) . Multifaceted conce ptions of ® t in structural e quation models. In K. A. B ollen & J. S. Long (E ds.), Testin g stru ctural equ ation mo dels (pp. 10± 39) . Ne wbury Park, CA: Sage. Tobin, L. D., Holroyd, K. A ., Re ynolds, R. V ., & Wigal, J. K. (1989) . The hierarchial factor structure of the coping strategie s inve ntory. Co gn itive Therap y an d Research, 13, 343-361. Watson, D., Clark, L. A ., & Te llege n, A. (1988) . De ve lopme nt and validation of brief me asure s of positive and negative affe ct: The PA NA S scale s. Jo urn al o f Perso nality an d So cial Psycho logy, 54, 1063-1070. Wonde rlic, E . F. (1992) . Won derlic Perso nn el Test: Form B . Libertyville, IL: Wonde rlic Personne l Te st, Inc.