The Utility of the Revised Life Orientation Test to ...

INTERNATIONAL JOURNAL OF PSYCHOLOGY, 1998, 33 (1), 45± 56

The Utility of the Revised Life Orientation Test to Measure Optim ism am ong Hong Kong Chinese Julia n C. L. Lai & Him Cheung City U niversity of H ong K ong, H ong K ong Wai-m an Lee & Helen Yu H ong K ong Polytechnic U niversity, Hong Kong

To exam ine the utility of the rev ised Life Orientation Test (LOT-R ) to measure optimism among Hong Kong Chinese, the psychometric properties of the revised and the original versions of the Life Orientation Test w ere compared. A total of 248 Hong Kong Chinese undergraduates w ere studied in the fall and 165 of these participants were tested agai n 5 months late r. Results indicated that the LOT-R is a reliable and valid measure of dispositional optimism among Hong Kong Chinese. Con® rmatory fac tor analysis showed that the LOT-R represents a one-fac tor model of optimism better than does the original ve rsion. Despite its brevity, the LOT-R is psychometrically sounder than the original sclae. These ® ndings point to the feasibility of replacing the original w ith the revised scale in future research among Hong Kong Chinese. However, the utility of the revised test in cross-c ultural comparisons may still be limited by the absence of emic components. Further research on optimism in the Chinese people w ith the LOT-R should pay more atte ntion to the identi® cation of e mic dimensions. L’ utiliteÂ du Life Orientation Test reÂ viseÂ (LOT-R) pour mesurer l’ optimisme chez des chinois de Hong Kong est eÂ val ueÂ e en comparant les proprieÂteÂ s psychomeÂ triques des ve rsions originale et reÂ viseÂ e. Un eÂ c hantillon de 248 eÂ tudiants preÂ gradueÂ s a eÂ teÂ eÂ tudieÂ aÁ l’ automne et 165 de ces participants ont eÂ teÂ retesteÂ s 5 mois plus tard. Les reÂ sultats indiquent que le LOT-R est une mesure ® able et valide de la tendance aÁ l’ optimisme chez les chinois de Hong Kong. Une analyse fac torielle con® rmatoire montre que le LOT-R repreÂ sente mieux que la version originale le modeÁ le unifac toriel de l’ optimisme. Bien qu’ il soit bref, le LOT-R est une meilleure mesure que l’ eÂ preuve originale. Ces reÂ sultats deÂ montrent qu’ il e st possible maintenant de remplacer l’ eÂ preuve originale par l’ eÂ preuve reÂ viseÂ e dans les futures recherches sur les chinois de Hong Kong. Cependant, l’ utiliteÂ de l’ eÂ preuve reÂ viseÂ e dans les comparaisons interculÁ l’ ave nir, les turelles peut encore demeurer limiteÂ e aÁ cause de l’ absence de composantes eÂ miques. A travaux sur l’ optimisme du peuple chinois qui utiliseront le LOT-R devront porter plus atte ntion aÁ l’ identi® cation des dimensions eÂ miques.

INTRODUCTION The Life Orie ntation Test (LOT) was ® rst developed by Scheier and Carve r (19 85) to assess the construc t of dispos itional optim ism, w hich w as conceptualize d as positive outcome expe ctanc ies having important he alth implic ations . This scale has been the most w idely used instrume nt for the assessment of optimism in psycholog ical

research. It consists of four pos itive ly w orded items (e.g. I’ m alw ays optimistic about my future), four negative ly w orded items (e.g. I hardly ever expe ct things to go my w ay), and four ® ller items (e.g. I enjoy my friends a lot). During the past decade, research using the LOT has yielded conve rgent evide nce that points to the bene® cial health effects associated w ith the cons truct of optim ism (review ed by Scheier &

Requests for reprints should be addressed to Julian Lai, Department of Applied Social Studies, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong (E-mail: ssjulwin@ cityu.edu.hk). Support for this research was provided by City University Grant No. 9030367. Thanks are due to two anonymous reviewers for their comments on an earlier version of this paper.

Ó

199 8 International Union of Psychological Science

46

LAI ET AL.

Carver, 1992, 1993). Highe r scores on the test hav e been show n to be assoc iated w ith better health outcomes during times of dif® culty in different populations suc h as coronary patie nts (D esharnais,G od in ,Job in ,V alois, & R oss, 1990), breast cancer patie nts (C arve r et al., 1993), gay men at risk for AIDS (Taylor et al., 1992), and colle ge students unde r acade mic stress (Aspinw all & Tay lor, 1992). Scheier, Weintraub, and Carver (198 6) attributed the health bene® ts associated w ith optim ism to adaptive coping. Optimists, in comparison w ith their less optim istic peers, are more like ly to use adaptive proble m-focuse d strategies and less like ly to engage in cognitive or behav ioural avoidanc e during stressful encounte rs. Studie s in w hich path analysis w as used have con® rmed the signi® cant mediating effect of coping (e.g. Carver et al., 1993 ; Scheier et al., 1989 ). More recently, ® ndings from a study of Sw edish tw ins indic ate a substantial genetic effect on optim ism (Plomin, Scheier, B ergeman, & Pederson, 19 92). This further accentuate s the importance of this construct in research on personality and health. Although the cons truct of optimism has been validate d successfully across diffe rent populations, the question of w hether it can ade quately be operationaliz ed by the LOT has lately stirred up controv ersies. Positions am ong researchers are divide d with regard to the dimensionality of the test. Scheier and Carver (1985, 1987) have suggested that the LOT should be st be treated as tapping a bipolar optim ism-pessimism dimension on the ground of the high loadings on the ® rst factor as w ell as the high correlation betw een the tw o factors (r = .6 4). The ® rst factor w as de ® ned by the pos itive ly phras ed items and the second by the negative ly phrase d ite ms. These tw o researchers arg ued that although the difference in ® t betw een a one -factor and a tw o-fac tor solution is ne gligibl e, the one-factor interpretation is more pars im onious and thus should be fav oured. On the other hand, other researchers have questione d this unidim ensional standpoint in view of evide nce show ing that the pos itive and the negative items of the LOT are loade d respectiv ely onto tw o different factors of optim ism versus pessimism (e.g. Lai, 199 4; Marshall, Wortman, Kusulas, Hervig, & V ickers, 199 2; Mook, Kleijn, & van der Ploe g, 1992 ). Using con® rmatory factor analysis, Marshall and Lang (1990) have found that their data on the LOT ® tted only a tw o-factor solution. Furthermore, the optimism and the pessimism subscales hav e been show n in other stu-

dies to correlate differentially w ith reports of physic al symptom s (Lai, 1994 ) as well as other external criteria (e.g. Marshall et al., 199 2). These ® ndings tend to sugg est that the LOT is multidimensional and should not be used as a single scale for assessing optim ism. In response to these intensifying controve rsies, Scheier, Carve r, and Bridge s (1994) have revised the LOT by deleting two positive items that, according to these inv estigators, assess generalized coping more than outcomes expe ctanc ies. To balanc e the num ber of positiv e versus negativ e items for separate scoring, a negativ e item was deleted and a new pos itive item w as adde d. As a result, the revised ve rsion (LOT-R ) cons ists of three positive , three negative , and four ® ller items. Exploratory factor analysis of the LOT-R produc ed a single factor that has a mean factor loading of .69 and explains 48 .1% of varianc e (Scheier et al., 1994 ). Con® rmatory factor analy sis further indic ated that the single -factor solution was superior to a tw o-fac tor one. In addition, the revised test w as also show n to have acceptable reliability as w ell as validity (Scheier et al., 1994). Despite these desirable psychometric properties (e.g . unidim ensionality and high content validity) , the utility of the LOT-R in future rese arch is still unclear for tw o reasons. First, data concerning the psy chometric propertie s of the LOT-R are still scanty. More evide nce is needed to con® rm further w hat Scheier e t al. (1994) have found, especially the unidime nsionality of the LOT. Second, the bulk of evidence concerning the predictive validity of the LOT comes from studie s using the original version, and it is unclear whether the revised test can predic t health outcomes as reliably as the original version did in previous researc h. R esearch on optimism using the LOT has been prolife rating during the past decade, but most of these studie s w ere carried out with English- spe aking samples in the West. As a consequence, the applic ability of the concept of optim ism, as assessed by the LOT, to non-W estern culture s remains to be inve stigate d. For instanc e, although Chine se folk wisdom has long held that be ing optim istic is bene® cial to health (Koo, 1987), scienti® c data regarding the optimism -health relation among Chines e are extremely scarce. Only a handful of studie s using the original Eng lish version of the LOT have been conduc ted recently among Chine se unde rgraduate s in Hong Kong (Lai, 1994, 199 5; Lai & Wan, 1996). The resulting data are consiste nt with previous ® ndings

LIFE ORIENTATION TEST

among Western samples. The tw o-fac tor structure of the test has been replic ated (Lai, 1994 ) and scores of the LOT have also been found to moderate the relation between hassle s and somatic complaints (Lai, 199 5). In anothe r study (Lai & Wan, 199 6), optimistic students were found to use more adaptive strategies than their less optim istic peers to cope with academic examinations. These ® ndings point to the applic ability of the LOT to research on optimism among Hong Kong Chines e. Althoug h encouraging, the data mentione d here must be interpreted w ith caution in view of a major lim itation concerning the use of English tests among responde nts w ho are not native English speakers. Scheier and Carver (1985 ) have pointe d out that negativ ely w orded items are semantic ally more comple x than pos itive ly worded ones, and this may form the basis of difference in responding between the pos itive and negative items. The sugge stion that direction of item wording can form the basis of the tw ofactor split of the LOT is supporte d by more recent evidence . Rephrasing either the pos itive or the negative items of the LOT into the oppos ite direction to form tw o scales with consiste ntly positive or negative items, Chang and McBrideChang (1996 ) found that only a one-factor solution emerges w hen items of the LOT have the same connotations. The semantic effects produc ed by differences in item w ording may be further accentuate d in Hong Kong Chine se taking the original English version of LOT. The tw o-fac tor solution and differential c orrelations reported by Lai (19 94) could have been caus ed by this semantic artefact. To assess dispositional optim ism among Hong Kong Chine se more acc urately, a Chines e version of the LOT must be used. Lai (1997) has looke d at the validity of a Chinese version of the LOT among different samples of Hong Kong C hine se. It w as found that the adapte d test exhibits acceptable reliability. In addition, the observed pattern of correlations between the test and symptom and affect reports is similar to that reported in prior studie s with Western samples. Contrary to ® ndings from previous studie s usi ng the original English version of LOT among Hong Kong Chine se (Lai, 1994 , 1995 ), the positive half and the negativ e half of the Chine se scale hav e comparable degrees of association w ith symptom and affect reports . Neverthe less, a tw o-fac tor split w as still observed. Unidim ensionality cannot be e stablishe d solely by translating the LOT into Chines e. In the absence

47

of a unidime nsional measure of optim ism, further research on the applic ability of the construc t in Hong Kong Chines e is hinde red. To resolve this proble m, an adaptation of the revised Life Orientation Test (LOT-R) seems to be one of the most straightforw ard solution. Although the LOT-R has been show n to be psychometrically sound (Scheier et al., 19 94), it has not been tested among Chine se and, as sugg ested earlie r, data concerning the psy chometric prope rties of the test are scanty. Therefore, a prope r evaluation of the applic ability of the concept of optim ism among Hong Kong Chine se using the Life Orientation Test should involv e a comparison of the original with the revise d version of the test. The present study was desig ned to compare the psy chometric prope rties of the original and the revised version of the LOT. If the LOT-R is going to replac e the original version in future research among Hong Kong Chine se , it must ® rst be at least as reliable and valid as the original scale and, second, unidime nsional. As in prior studie s done w ith Hong Kong Chine se samples (e.g. Lai, 19 94, 1997 ), exploratory factor analy ses were use d to examine the structure of C hine se versions of the LOT and the LOT-R. These w ould be supple mented by con® rmatory analyse s that are able to provide m ore conclusive evide nce regarding dimensionality. Construc t validity of the tw o versions of the Life Orientation Test w as studie d by looking at the produc t-moment c orrelations between the tw o versions of the test and reports of physic al symptoms as w ell as positive and negative affects. These health-re lated variable s w ere chosen becaus e health bene® t is one of the most central compone nts of the construc t of optimism (Scheier et al., 1994). To assess whether the tw o versions m easure the same construc t of optim ism or not, their respective correlations with a partic ular health-re lated measure w ere compared by usi ng Hotelling’ s T test (R osenthal & Rosnow , 1991, p. 50 7). If the tw o versions assess the same construct, the differences between these c orrelations should be negligibl e. Convergent an d discriminant validit y were examined by correlating the adap tedLO T-R w ith C hin ese version s of the Hopelessn ess Scale (CH O PE)(S hek, 1995) an d the Health Locus of Control Scale (HLC) (W allsto n, W allsto n, Kaplan ,& M aides, 1976). These two scales are measures of negative ou tcom eexpectancies an d speci® c exp ectan cies regarding health -related lo cus of contro l, respectively. If the LO T-R is a valid measu re of

48

LAI ET AL.

positive outco me expectan cies, it sho uld correlate highly with the LO T,m od eratelyw ith the C-H O PE, w hich is also a measu re of gen eralize d expectan cies, an d weak ly with the HLC, w hich assesses exp ectancies speci® c to a single dom ain.

were require d to indic ate on a 5-point scale the degree to w hich they agre e or disagre e w ith each of the 13 item s (0 = strongly disagre e; 4 = strongly agre e). Items pertaining to the LOT and the LOT-R w ere scored separate ly .

METHOD

C hine se Affect Scale (C AS). This affect checklist was base d on affective terms indige nous to the Hong Kong C hine se culture . It consists of 10 positive and 10 negative affect indic ators in Chine se (Hamid & Cheng, 1995 ). The scale has been found to be a reliable as w ell as a valid measure of both state and trait affect among Hong Kong Chine se (Hamid & Cheng, 1995). According to the two researchers, princ ipal compone nt analysis yielded a 2-fac tor solution w ith the 10 neg ative items (sad, helple ss, frighte ned, disappointe d, bitter, tense, insecure, exhauste d, annoye d, depressed) loade d onto the ® rst factor and the 10 positive items (contented, exuberant, excited, ag reeable , happy, meaningful, joyful, comfortable , relaxe d, peaceful) loade d onto the second. To complete the scale, partic ipants rated each item on a 5-point scale (0 = very slightly or not at all; 4 = extremely) to indic ate how they were feeling in the past month. A pos itive affect (PA) and a negative affect (NA) score w ere computed by adding the rating s of the positiv e and the negative items respectively. The correlation betw een PA and NA scores w as found to be low in prior studie s (e.g . r = 2 .30 , Hamid & Cheng, 1995; r = 2 .2 8, Lai, Hamid, & Chow , 19 96), and was also low for the present sample, r = 2 .28 .

Participants The partic ipants of the present study w ere 170 w om en and 78 men taking a c ourse in introduc tory psy chology at a unive rsity in Hong Kong. They were majoring either in social w ork or nursing. The mean age of w omen w as 28.3 yrs (SD = 2.1) w hereas that of men w as 30.8 yrs (SD = 1.8). The present study involve d tw o test sessions. The ® rst took plac e in the fall and inv olve d the complete sample of 170 women and 78 men. The second test w as done 5 months later in the spring and involve d 100 w omen and 65 m en of the original sample . The mean ag e for the female and male partic ipants of the second session w as 30 .1 yrs (SD = 3.2 ) and 31.0 yrs (SD = 2.1), respectively.

M easures The Life Orientation Test (LOT & LOTR). The LOT (Scheier & Carve r, 19 85) is comprised of four pos itive ly w orded, four negativ ely w orded, and four ® ller items. On the other hand, the LOT-R c onsists of three negativ e and tw o positive items from the LOT, one new posi tive item (`Òverall, I expe ct more good things to happe n to me than bad’ ’ ), and the four ® ller items of the LOT. The two versions of the scale have been used as a measure of the extent to w hich individuals have pos itive expe ctancies of life outcomes (Scheier et al., 199 4). For reasons mentione d earlie r, Chine se versions rather than Englis h versions of the LOT and LOT-R w ere used. Items of the tw o versions w ere mixed and administered to partic ipants as a sing le scale containing a total of 13 items. Sinc e the 12 items of the original LOT had been translated and used previously by the ® rst author (Lai, 19 97), only the new positiv e item w as translate d for the prese nt study . This item was ® rst translated into Chine se by the ® rst author. It w as then back-trans lated into English by one of the other authors , w ith the item being modi® ed until the exac t meaning w as back-translated. Partic ipants

Physica l Symptoms Checklist (PSC). This scale has a total of 30 commonly experienced physic al symptom s in Chine se (e.g. headac he, nause a, cough, upset stomach, running nose ). Previous studie s have show n that the scale is sensitiv e to stress in different Hong Kong Chine se populations (Lai, 199 5; Lai et al., 1996 ). Extracted from the Cohen-Hobe rman Inventory of Physic al Symptoms (C ohen & Hoberman, 1983) and the Pennebake r Inventory of Limbic Languidne ss (Pennebaker, 19 82), symptom s on the checklist were not appare ntly psyc holog ical in nature . To comple te the scale, partic ipants w ere asked to rate on a 4-point scale how often they had experienced each of the 30 symptom s in the las t month (0 = not at all; 4 = very often). A total symptom sc ore was generated by adding ratings on the 30 items. The Chine se Hopelessness Scale (C-HOPE). Originally develope d by Beck, Weissman, Lester,


and Trexler (1974) to measure negative expe ctancies about oneself and one’ s future life, this 20item test w as trans lated by Shek (199 5) and validate d among a sample of 50 0 Chine se unive rsity students in Hong Kong. The sc ale consis ts of 11 items phrase d in a hope less direction and 9 positive ly phrase d items. To complete the test, partic ipants rated each item on a 6-point scale (1 = strongly disag ree; 6 = strong ly agre e) to indic ate how w ell the items w ould apply to them. The sc ale w as scored in the hope less direction. A total score w as compute d by adding ratings on the 11 hope le ss items and reversed ratings on the 9 positiv e items. This test w as administe red only at the second test session. The Health Locus of Con trol Scale (HLC). The original Englis h version of this test (Wallston et al., 197 6) w as translate d by Lai and Yue (199 7) into Chine se. Every effort w as made to ensure the rete ntion of the meaning s of the orig inal items in the trans lated version. Comprising of 11 items, of w hich 6 are externally worded and 5 are internally worded, the test is a measure of ``generaliz ed expe ctanc ies regarding locus of control related to health’ ’ (Wallston et al., 1976 ). `Ìnternals’ ’ , indiv iduals who believe that their health is controlle d by internal more than external factors, have been reported to be more motivate d to expos e themselves to health information than `èxternals ’’ . To com plete the test, responde nts were asked to rate on a 6-point scale (1 = strongly disag ree; 6 = strong ly agre e) to indic ate the extent to w hich they agre e or disagre e with each of the 11 items. The test w as scored in an exte rnal direction. A total score was computed by adding rating s on the six `èxternal’ ’ items and

that on the ® ve reversed `ìnternal’ ’ items. Similar to the C-HOPE, the HLC w as adm iniste red only during the se cond test session.

Procedure During the ® rst test, w hich took plac e in the fall, questionnaire s comprise d of Chine se versions of the LOT and the LOT-R, the phy sical symptom s checklist (PS C) (Lai, 1997 ; Lai et al., 1996), and the Chine se Affect Scale (CAS) (Hamid & Cheng, 19 95) w ere adminis tered to the partic ipants (N = 24 8) at the end of a lecture for the ful® lment of part of the require ments of a course in introduc tory psychology . After the test session, partic ipants w ere debriefed about the aim of the study, and w ere assured that the inform ation they had provide d w ould be kept strictly con® dential. The second test took plac e about 5 months later in the spring. Since only some of the 24 8 stude nts taking part in the ® rst te st would meet in a sing le class again in the spring, the second test could only inv olve 165 of the original complete sample . The test procedure was similar to that of the ® rst session except that the LOT, the LOT-R , the CHOPE, and the HLC w ere administe red. The obje ctiv e of the second test w as to examine test± retest reliability as well as conve rgent and discriminant validity of the LOT-R.

RESULTS Descriptive Statistic s The means, SDs and C ronbac h alphas of the scales exam ined in the ® rst and the second test session w ere listed in Table 1. Scores on the LOT

TABLE 1 Means, Standard Deviations, and Cronbach Alphas of Scales Examined During the Two Test Sessions

First Test Session (N = 248) 1. LOT 2. LOT-R 3. Positiv e affect 4. Negative affect 5. PSC Second Test Session (N = 165) 1. LOT 2. LOT-R 3. HLC 4. HOPE a

49

Cronbach a

Mean

SD

18.6 13.9 19.0 11.2 36.5

3.4 3.3 5.6 6.6 14.1

.71 .70 .87 .89 .88

19.2 14.1 34.7 45.5

3.8 3.0 5.5 11.3

.68 .65 .65 .84

a

Data from a subgroup of the complete sample of 248 stude nts.

50

LAI ET AL.

and the LOT-R were studie d in both test sessions. Positive affect (PA), negative affect (NA), and scores of the phys ical symptoms checklis t (PS C) w ere exam ined in the ® rst session, w hile the CHOPE and the HLC w ere studie d in a subgroup of 16 5 students of the comple te sample during the second se ssion. All the scales show reasonable internal consistency, w ith Cronbac h a s rang ing from .65 to .89. Although the LOT-R is shorter than the original version, it remains internally consiste nt. The means of LOT scores of the tw o samples w ere 18 .6 (SD = 3.4) and 19.2 (SD = 3.8 ), respectively, w hich are comparable to that obtaine d in a prior study from a group of Hong Kong Chine se unde rgraduate s using the orig inal Englis h version (Lai, 19 94: M = 17.9 , SD = 4.3). The means of LOT-R scores w ere 13.9 (SD = 3.3 ) for the ® rst test and 14 .1 (SD = 3.0 ) for the second. These are close to thos e reported previous ly among different Western sample s (e.g. Scheier et al., 1994: colle ge students , M = 14 .3, SD = 4.3; bypass patie nts, M = 15 .2, SD = 4.1). The difference in the mean scores of the tw o versions of Life Orientation Test betw een men and w ome n of the two sample s is negligible . For the sample of ® rst test session, the means of LOT for men and women w ere 18.9 (SD = 3.0) and 18 .4 (SD = 3.5), respectively. These w ere 19.3 (SD = 3.9) and 19.0 (SD = 4.0) for the sam ple of the second test session. With respect to the LOT-R scores for the ® rst test session, the means w ere 14.1 (SD = 3.4) for men and 13.7 (SD = 3.6) for women. For the second test se ssion, the mean score for men w as 14 .3 (SD = 3.4 ) and that for w omen w as 13.8 (SD = 3.1). Test± retest reliability coef® cients were computed for the two verions of Life Orientation Test for the sample (N = 165) partic ipating in both test sessions. These w ere .68 for the LOT and .66 for the LOT-R. The latter score is close to that reported in prior studie s with similar test± retest intervals (e.g. LOT-R, .68, Sc heier et al., 19 94). How ever, the coef® cient of the LOT for the present sam ple is reasonably low er in comparison to that assoc iated w ith a 4-we ek interval reported previously (e.g. .79, Scheier & Carve r, 198 5).

Explorator y Factor Analysis of the Tw o Versio ns of the Life Orientation Test Principle compone nt analy sis w as performed on the separate sets of items comprising the LOT and the LOT-R (w ith the negative ly phrase d items

reversed). Analyse s w ere restricted to data of the complete sample colle cted during the ® rst test session. Tw o factors having eigenv alue s greater than 1 w ere generated for both of the tw o scales using varimax rotation. The factor loadings of the tw o scales are listed in Table 2. The 4 positive items of the LOT w ere loade d onto the ® rst factor, w hich accounte d for 34 .4% of total varianc e (eigenvalue = 2.75 ), w hereas the 4 negative item s w ere loade d onto the second factor and explaine d an additional varianc e of 17.8% (e igenvalue = 1.42). A similar tw o-factor solution was generated for LOT-R. The ® rst factor was loaded w ith the 3 positive items and explaine d 33.3% of total varianc e (eig envalue = 2.01). The negative items w ere loade d on the second factor and accounte d for 22 .8% of additional varianc e (e igenvalue = 1.3 7). The factor solution of the LOT yielded in the present sample is comparable to those reported in prev ious studie s w ith Western sample s (e.g. Scheier & Carver, 1985; Scheier et al., 1994). It is also similar to that produc ed previousl y in a group of Hong Kong unde rgraduate s using the orig inal Eng lish version (Lai, 1994). How ever, the one-factor solution for the LOT-R reported by Scheier et al. (1994 ) w as not replicated in the present sample. The reasons for this are not immediate ly appare nt but several factors may account for the observed dis crepanc y. First, TABLE 2 Factor Pattern Loadings of the Original LOT and the Revised LOT for 248 Chinese University Students

LOT Item Item Item Item Item Item Item Item

Factor 1

Factor 2

.78

.17 .57 .14 .19 .70 .69 2 .16 .66

a

1 3 4 5 8 9 11 12

2 .03 .80 .70 2 .04 .21 .71 .29

b

LOT-R Item 1 Item 3 Item 4 Item 7 Item 9 Item 10 a

.79 .16 .81 .17 .34 .74

2 .02 .68

2 .02 .75 .57 .15

O rder of i te m s the same as the orig inal LOT in Sc he ie r and Carver (1985). b Orde r of i te ms the s am e as the Revised LOT in Scheie r et al. (1994).


the factor solution reported by Scheier et al. (1994 ) w as based upon a sample of 2055 unde rgraduate s, w hich is about 10 times large r than that of the present study. It should not be surprising that the one-factor solution has not been replicated in the present sample . Second, the LOT-R was adminis tered as a separate scale in the study conduc ted by Scheier et al. (1994) , w hereas the items of both the LOT and the LOT-R w ere combine d and administe red as a single scale in the present study. Response s to item s of the LOT-R may be inadve rtently affected by the tendency to respond differentially to the pos itive and the negativ e items comprising the LOT.

Factor Com pariso ns To examine more closely the similarity betw een the factor structures of the original and the revised LOT reported previous ly among Western sample s and those for the present sample of Hong Kong Chine se, Tucker’ s coef® cients of congruenc e (Harman, 197 6, p. 334) w ere computed. The factors of the LOT reported by Scheier and Carver (1985) w ith a group of 624 unde rgraduates and those obtaine d in the present sample of 248 Chine se students w ere compare d. The coef® cient betw een the ® rst factor in the Western sample and its c ounte rpart in the present sample is .82, whereas that betw een the tw o second factors is .69 . With respect to the coef® cient betw een the ® rst factor in the Western sample and the second factor of the prese nt sample , a much low er value of .21 is obtaine d. The coef® cient betw een the second factor of the Western sample and the ® rst one in the prese nt sample is equal to .14 . The large difference between these tw o sets of coef® cients sugge sts a reasonable match of the factor solutions obtaine d in tw o diff erent sample s. Similar comparisons w ere carried out for the single factor of the LOT-R reported by Scheie r et al. (1994) and the tw o factors yielded in the present sample. A value of .6 8 w as obtaine d for the coef® cient associated w ith the ® rst factor (positive items), and that for the second factor (negative items) was .5 2. Althoug h a closer match w as obtaine d for the ® rst fac tor of positive items with the single factor reported by Scheier et al. (1994 ), a moderate deg ree of congrue nce w as also observed in the second factor. This means that both factors produc ed in the present sample have cons iderable ove rlap with the single factor reported by Scheier et al. (19 94), w hich lends

51

preliminary support to a unidim ensional interpretation of the Chine se LOT-R.

Con® rm ator y Factor Analysis C on® rmatory factor analys es were performed on both the LOT and the LOT-R data to contrast the applic ability of one- and two-fac tor models. For the LOT, a tw o-factor model ® ts the data better 2 than a one-factor model: [ one-factor: c (36) = 31 6.39, P < .001 ; Bentler-Bonett normed ® t 2 inde x = 0.6 1; tw o-fac tor c (35) = 261.93 , P = < .001; Be ntler-B onett normed ® t inde x = 0.67] . This replic ates prev ious ® ndings w ith the LOT in English-speaking samples (Marshall & Lang, 19 90). The LOT-R, on the other hand, better supports a one-factor than a tw o-fac tor model: [ one-factor: c 2 (38) = 486.48 , P < .00 1; B entler-Bonett 2 normed ® t inde x = 0.69 ; tw o-fac tor: c (37) = 52 6.24, P = < .001; B entler-B onett normed ® t inde x = 0.34] . The LOT-R is therefore more like ly to represent a unidim ensional than a multidim ensional construc t, as sugge sted by data reported previously in an Eng lish-speaking sample (Scheier et al., 1994 ). Table 3 summarizes the factor loadings for the one- factor solution of the LOT-R yielded by princ ipal compone nt analysis . The coef® cient of cong ruence betw een this factor and that reported by Scheier et al. (1994 ) w as .80, w hich implie s a close match. The test is like ly to measure the same construc t in at least tw o different culture s. Although the split betw een the pos itive versus the neg ative items of the LOT is supporte d by the analyse s just reported, the test was treated as a single scale for measuring optim ism in subsequent

TABLE 3 Factor Loadings for the Onefactor Solution of the Revised Life Orientation Test Item Item Item Item Item Item

1 3 4 7 9 10

.60 .52 .44 .65 .69 .70

Order of items is the same as the LOT-R in Sche ie r et al. (1994) (N = 248).

52

LAI ET AL.

analyse s for the follow ing reasons . First, as indicated, the difference in ® t between one- and tw ofactor models is much smalle r for the LOT than the LOT-R . Although the above data point clearly to a one-fac tor interpretation for the LOT-R, the evide nce for a tw o-fac tor model of the LOT is much less conclusi ve. Second, as demonstrate d by Lai (1997 ), the tw o-fac tor split of the LOT is not like ly to be substantive since the positive and the negative halv es m easure the same construc t of optim ism, although the former does this better. For reasons of pars imony, Lai (19 97) sugge sted that the test can still be treated as a sing le measure of optimism . However, this does not imply that the revised version is no different from the orig inal test. The LOT-R represents a one-factor model more conv incingly than does the original test and thus is more compatible w ith the concept of dispositional optimism as originally formulate d by Scheier and Carver (19 85, 19 87). In other words, the revis ed test is a more valid measure.

Constr uct Validity of the LOT and the LOT-R Correlations am ong the ® ve scale scores for the complete sample are listed in Table 4. The LOT w as signi® cantly correlated w ith the three healthrelated measures (PA, NA, and PS C), and so w as the LOT-R. In addition, similar patte rns of correlations betw een predictors and criteria w ere observed for the two versions of the Life Orientation Test. Higher scores on the LOT or the LOT-R w ere associated with low er scores on the NA and the PSC, but highe r PA scores. The correlations betw een the LOT and the three health measures replic ated prior data for tw o diffe rent sample s of Hong Kong Chine se (Lai, 199 7). This patte rn of assoc iation is consiste nt w ith that reported in prior Western rese arch (e.g. symptom reports, Scheier & Carver, 1985; pos itive and negative

affect, Marshall et al., 19 92). The correlation betw een the tw o Chine se versions of the test was high, r = .89 , which implie s that both are like ly to assess the same construct of dispositional optim ism. To look at the diffe rences in predictive pow er betw een the original and the revis ed version, c orrelation betw een LOT score s and scores on each of the thre e health m easures was compared to that betw een LOT-R scores and the same health measure using Hotelling’ s T tests (Rosenthal & R osnow , 1991, p. 507 ). The Hote lling’ s Ts associated with the PA, the NA, and the PSC w ere .52, .40, and .2 7, respectively. None of these Ts was statistically signi® cant at P = .05. This indic ates that the LOT and the LOT-R are correlated to a similar extent with each of the three health measures, which lends additional support to the contention that the se tw o versions to a very large extent measure the same construct of dis positional optimism. In other words, the revised and the original test have comparable predic tive pow e r as assessed by the three health-relate d criteria used in the prese nt study.

Convergent and Discrim inant Validity As mentioned earlie r, conve rgent and discriminant validity of the LOT-R w ere examine d using the data collec ted from the partic ipants (N = 165 ) of the second test session. Given that dispositional optim ism has originally be en conceptualiz ed at the level of belief or value processes (Scheier & Carver, 198 5), measures used for validating any version of the Life Orientation Test should also pertain to the same level of conceptualization. How ever, measures that have been used most commonly in research w ith Western sample s to validate the Life Orientation Test are those that assess neurotic ism or affectivity (e.g. Marshall et al., 199 2; Smith, Pope , Rhode w alt, & Poulton,

TABLE 4 Product-moment Correlations of the Scores of the Original LOT, the Revised LOT, Positive Affect, Negative Affect, and the Physical Symptoms Checklist

1. 2. 3. 4. 5.

LOT LOT-R Positive affect Negative affect PSC

1

2

3

4

5

±

.89** ±

.35** .32** ±

2 .23**

2 .17*

2 .28**

2 .25**

Findings are based on 248 Chinese univ ersity students. *P < .05; **P < .01 (2-tailed).

2 .27** ±

2 .17*

.37** ±


1989 ). According to Chiu, Hong, and Dweck (1994 ), these concepts are more related to motivational than belief processes. In other w ords , results from validation studie s using tests that tap motivational or affectiv e construc ts should be interpreted cautiously. In the prese nt study, the original LOT, the Chine se Hopelessness Scale (C-HOPE) and the Chine se version of the Health Locus of Control Scale (HLC) w ere used to demonstrate the c onve rgent and discriminant validity of the LOT-R. As discussed earlie r, the C-HOPE and the HLC w ere selected because the construc ts tappe d by these tw o scales are closely related to outcome expe ctanc ies and thus pertain to the level of belief or value processes. The produc t-moment correlations between the LOT-R and the LOT, the C-HOPE, and the HLC w ere r = .86, r = 2 .45, and r = 2 .16 , respectiv ely. These results are consiste nt w ith our predictions stated earlie r. The LOT-R is highly correlated w ith the LOT, a measure of the same construct,moderatelycorrelatedw ith the C-HO PE, which assesses pessimism, and weakly correlate d with the HLC, a test of expe ctanc ies speci® c to locus of control of health. Conve rgent and discriminant validity are demons trated clearly for the LOT-R.

DISCUSSION In summary, the present data show that as a test of dispositional optimism am ong Hong Kong Chinese stude nts, the revise d Life Orientation Test appe ars to be more adequate than the original version for the follow ing reasons. First, the revised test is more valid in terms of its content since the two items that measure coping rather than outcome expe ctanc ies hav e been deleted. Second, the LOT-R correlates with health measures to an equal extent as the LOT, and demonstrates conve rgent and discriminant validity. Third, evide nce from con® rmatory factor anlysis favours a unidim ensional interpretation of the test. Take n toge ther, these imply that the LOT-R is psychom etrically sounde r than the original version and should replace the latter in future research among Hong Kong Chine se. In addition to the psychometric prope rties just mentione d, ``metric equivale nce’ ’ (Berry, 198 0) has also been demonstrate d for the LOT-R w hen the one-factor solution is used for factor comparison. This means that similar structure of the LOT-R w as observed across at le ast tw o different culture s and this m akes cross-cultural compari-

53

sons bas ed on the test meaningful. Althoug h functional and conceptual equivale nce should also be demonstrated in addition to metric equiv alence (B erry & Dasen, 1974 ) before cross-cultural comparability can fully be asserted, metric equivalence is w ithout question one of the very few empiric al indic es of culture -fairne ss. To our know ledge , no such evide nce has been reported for anothe r measure of optim ism. On the other hand, it has been pointe d out that a six-ite m scale may suffer certain psy chometric disadv antag es compared to a scale that contains m ore items (DeVellis , 1991). Fortunate ly, the present results show that the LOT-R does not suffer m uch despite its brevity. The coef® cient of internal c onsiste ncy of the LOT-R is decent as that of the long er original version (.70 versus .71) and the tw o versions have comparable test± retest reliability (.6 8 for the LOT versus .66 for the LOT-R). In addition, the predictive pow er of the revised test is not signi® cantly different from that of the original test. Unaffected by these potential dis advantage s, the LOT-R actually has certain advantage s that the LOT does not. For instance , brevity can facilitate use of the test in research, especially for proje cts containing long protoc ols in w hich subject burde n is an important conside ration. Seve ral other aspects of the results reported here deserve further discussion. One of these is related to the fundam ental issue concerning the conv ergent and discriminant validity of the LOT-R and the other is concerned w ith the potential differences in the conceptualiz ation of optim ism betw een Western and Chine se culture s. Data regarding the conv ergent and discriminant validity of the LOT-R is still extremely scarce, since the original test has been revised only very recently and the revision has not been used m uch. Scheier et al. (1994) hav e show n that the LOT-R is only modestly c orrelated w ith several other conceptually related scales (e.g. trait anx iety, neurotic ism and self-esteem ), but highly correlated w ith the LOT. Validating the LOT-R, a test of optim ism, with scales tapping construc ts suc h as neurotic ism or self-esteem is proble matic, since the conceptual linkage betw een optim ism and the latter concepts is questionable . According to Scheier and Carver (1985, 19 87), dispositional optim ism is charac terized by a generaliz ed tendency to expe ct positive outcome s, or the belie f that ``good, as oppose d to bad things w ill generally occur in one’ s life’ ’ (Scheier & C arv er, 1993, p. 26). Apparently, disposition al optimism is c onceptualiz ed at the level of belief

54

LAI ET AL.

or value processes. On the other hand, self-esteem is the ov erall affective evaluation of one’ s ow n w orth (Blascovic h & Tomaka, 199 1) w hereas neurotic ism is closely related to emotional stability (Hjelle & Ziegler, 1992). As pointe d out earlier, these two concepts are more related to affective than belie f processes. Past research has ov erlooke d this important difference and the continual use of neurotic ism and other related constructs to validate LOT or LOT-R has not been questioned. To circumvent this proble m, a measure of generaliz ed expectancies (the Chine se Hopelessness Scale, C-HOPE) and an inde x of expec tancies speci® c to a sing le domain (the Health Locus of Control Scale, HLC), w ere use d in the present study . As reported above , the LOT-R w as more correlated w ith the C-HOPE than w ith the HLC. Since these tw o scales are conceptualized at the same level as the LOT-R , the present results provide stronge r support for conve rgent and discriminant validity of the LOT-R. Although studie s on optim ism using the Life Orientation Test have prolife rated during the pas t decade, most of these have been carried out w ith English- speaking samples in the West. The limited number of studie s done among Hong Kong Chine se (e.g. Lai, 1994, 1995, 1997 ; Lai & Wan, 19 96) w ere also based on this m easure, w hich has been develope d in a Western culture . As a consequence, the important differences in the conception of optimism between Chine se and Western culture is playe d dow n. In a review of Chine se folk w isdom of behav ioural health, Koo (1987) pointe d out that for the Chine se, being optimistic means to be able to accept one’ s current life conditions posi tiv ely, in addition to expe ct good things occurring to one’ s life. How ever, the conceptualization of optim ism by Scheier and Carver (19 85, 1987) w as rooted in the long tradition of expec tancy-v alue theories of psychology and bas ed primarily on pos itive outcome expectancies. In other w ords, optimism as operationaliz ed by different versions of the Life Orientation Test may not be compatible w ith that construe d by the Chine se people. Most relevant to the present discussion is the recent ® nding that optim ism as construe d by Chine se unde rgraduate s as a combination of the Western dimensions of extrove rsion and emotional stability, and cannot be matched to any single fundam ental dimension originating in Western psychology (Yang & Bond, 1990 ). This implie s that certain im portant emic compone nts are inevitably omitted in etic adaptations. Although the LOT-R has been show n in the pre-

sent study to measure positive outcome expectancies reliably and validly, how central outcome expectanc ies are to the Chine se concept of optimism remains uncertain. This cons titutes one of the m ajor limitations of using the scale in future rese arch among Hong Kong Chine se. Nevertheless, the evide nce and argum ents presented here do not unde rmine the psy chometric status of the adapte d LOT-R. The test does m easure positive outcome expe ctanc ies among the prese nt sam ple of Hong Kong Chine se and scores on the scale hav e signi® cant correlations with health-re lated measures. What has been pointe d out is the need of more emic inputs to the operationaliz ation of optim ism. Such inputs become more important w hen the personality construc t in question orig inate s from the philosophie s or folklore of a speci® c culture (Church & Katig bak, 1988), as in the present case of optim ism among Hong Kong Chines e. Moreove r, it has been reported that scales in w hich univ ersal etic elements are combine d w ith appropriately de rived emic elements are more capable of identifying cultural differences than are etic tests (Triandis & Marin, 19 83). In view of these advantage s, future attempts of revision of the LOT-R in Chine se contex ts should giv e increased attention to the identi® cation of emic dimensions. Manuscript received January 1997 Rev ised manuscript accepted Septem ber 1997

REFERENCES Aspinwall, L.G., & Taylor, S .E. (1992) . Modeling cognitive adaptation: A longitudinal investigation of the impact of individual differences and coping on college adjustment and performance. Jour nal of Persona lity and Social Psychol ogy, 63, 989 ± 1003. Beck, A.T., Weissman, A., Lester, D., & Trexler, L. (19 74). The measurement of pessimism: The Hopelessness Scale. Jo urnal of Consulting and Clinic al Psychology, 42, 861± 865. Berry, J.W. (1 980). Introduction to methodology. In H.C. Triandis & J.W. Be rry (Eds.), Handb ook of cross-cultural psychol ogy: Vol. 2. M eth odolog y (pp. 1± 28). Boston, MA: Allyn & B acon. Berry, J.W., & Dasen, P. (Eds.) (1 974). Introduc tion to culture and cognition. London: Methuen. Blascovich, J., & Tomaka, J. (19 91). Measures of self-e steem. In J.P. Robinson, P.R. Shaver, & L.S. Wrightsman (Eds.), Measures of personali ty and social psychology atti tude s: Vol. 1 (pp. 115± 16 0). S an Diego, CA: Academic Press. Carver, C.S., Pozo, C., Harris, S.D., Noriega, V ., S cheier, M.F., Robinson, D.S., Ketcham, A.S.,


Mofat, F.L., & Clark, K.C. (1993 ). How coping mediates the effect of optimism on distress: A study of w omen w ith early stage breast cancer. Journ al of Person ality and Social Psycho logy, 6 5, 375± 390. Chang, L., & McB ride-Chang, C. (1996) . The fac tor structure of the Life Orientation Test. Educational and Psychologi cal Measu rement, 56, 325± 329. Chiu, C.Y., Hong, Y.Y., & Dw eck, C.S. (1994) . Toward an integrative model of personality and intelligence. A general fram ew ork and some preliminary steps. In R.J. Sternberg & P. Ruzgis (Eds.), Personality and intelligence (pp. 10 4± 134). Cambridge: Cam bridge University Press. Church, A.T., & Katigbak, M.S. (1988) . The emic strategy in the identi® cation and assessment of personality dimensions in a non-Western culture: Rationale, steps, and a Philippine illustration. Jo urnal of Cross-Cultural Psychol ogy, 19, 140± 163. Cohen, S., & Hoberman, H. (1983) . Positive events and social supports as buffers of life change stress. Journ al of Applied Social Psycholo gy, 13, 99± 125. Desharnais, B .S., Godin, G., Jobin, J., V alois, P., & Ross, A. (1990) . Optimism and health-related cognitions afte r a myocardial infarction. Psychol ogical Repor ts, 67, 1131± 113 5. DeVellis, R.F. (1991) . Scale develop men t: Theory and appl ication. New bury Park, CA: Sage . Hamid, P.N., & Cheng, S.T. (19 95). The development and validation of an ind ex of emotiona l dispo sition and mood state: The Chinese Affect Scale. Manuscript submitted for publication. Harman, H.H. (1976) . Mode rn fac tor analys is (3rd edn.). Chicago, IL: The University of Chicago. Hjelle, L.A., & Ziegler, D.J. (1992) . Person ality the ories (3rd edn.). New York: Mc Graw-Hill. Koo, L.C.L. (198 7). Nour ishment o f life: Health in Chinese society. Hong Kong: The Commercial Press. Lai, J.C.L. (1994) . Differential predictive power of the positively versus the negati vely w orded items of the Life Orientation Test. Psychologi cal Reports, 75, 1507 ± 1515. Lai, J.C.L. (1995) . The moderating effect of optimism on the relation between hassles and somatic complaints. Psychologi cal Repor ts, 76, 883± 8 94. Lai, J.C.L. (199 7). R elative predictive power of the optimism versus the pessimism index of a Chinese version of the Life Orientation Test. Psychol ogical Record, 47, 399± 410 . Lai, J.C.L., Hamid, N.P., & Chow, P. (19 96). Gender difference in hassles and sym ptom reporting am ong Hong Kong adolescents. Jour nal o f Social Behaviour and Person ality, 11, 149± 164. Lai, J.C.L., & Wan, W. (1996) . Dispositional optimism and coping with academic exami nations. Perceptual an d Motor Skills, 83, 23± 27. Lai, J.C.L., & Yue, X.D. (1997) . M easuring dispos itional opti mism in Chinese unde rgra duate s. Unpublished raw data. Marshall, G.N., & Lang, E.L. (19 90). Optimism, selfmastery, and symptoms of depression in women professionals. Jour nal of Persona lity and Social Psychol ogy, 59, 132± 139 .

55

Marshall, G.N., Wortman, C.B., Kusulas, J.W., Hervig, L.K., & Vickers, R.R. (1992 ). Distinguishing optimism from pessimism: Relations to fundamental dimensions of mood & personality. Jour nal of Person ality a nd Social Psycholog y, 62, 1067± 1074. Mook, J., Kleijn, W., & van der Ploeg, H.M. (1992) . Positively and negative ly w orded items in a self-report measure of dispositional optimism. Pscyho logical Reports, 71, 275± 278. Pennebaker, J.W. (1982) . The psychol ogy of phys ical symptoms. New York: S pringer-V erlag. Plomin, R., Scheier, M.F., B ergeman, C.S., & Pederson, N.L. (19 92). Optimism, pessimism, and me ntal health: A twin/adoption study. Personal ity and Individual Differenc es, 13, 921 ± 930. R osenthal, R., & Rosnow, R.L. (1 991). Essentials o f behav iora l research: Me thods and data a nalys is. New York: McGraw -Hill. Scheier, M.F., & Carver, C.S. (1985) . Optimism, coping and health: Asse ssment and implications of ge neralized outcome expectanc ies. Health Psychol ogy, 4, 219± 247. Scheier, M.F., & Carver, C.S. (1 987). Dispositional optimism and physical w ell-being: The in¯ uence of generalized outcome expectancies on health. Jour nal o f Persona lity, 55 , 169± 21 0. Scheier, M.F., & Carver, C.S. (1 992). Effects of optimism on psychological and physical w ell-being: Theoretical overview and e mpirical update. Cogni tive Therapy and Researc h, 16, 201± 228. Scheier, M.F., & Carver, C.S. (1993) . On the powe r of positive thinking: The bene® ts of being optimistic. Current Directions in Psychologi cal Scienc e, 2, 26± 30. Scheier, M.F., Carver, C.S., & Bridges, M.W . (1994) . Distinguishing optimism from neuroticism (and trait anx iety, self-mastery, and self-esteem): A re-eval uation of the Life Orientation Test. Jour nal of Personal ity and Soc ial Psychol ogy, 67, 10 63± 1078. Scheier, M.F., Matthe w s, K.A., Ow ens, J.F., Magove rn, G.J., Lefebvre, R.C., Abbott, R.A., & Carver, C.S. (19 89). Dispositional optimism and recovery from c oronary bypass surgery: The bene® c ial effects on physical and psychological w ell-being. Journ al of Personali ty and Social Psychol ogy, 57, 1024± 10 40. Scheier, M.F., Weintraub, J.K., & Carver, C.S. (1986) . Coping with stress: Divergent strateg ies of optimists and pessimists. Jour nal of Personal ity and Social Psycholo gy, 51, 1257± 1264. Shek, D.T.L. (1995) . Measurement of pessimism in Chinese adolescents: The Chinese Hopelessness Scale. Social Behav ior and Personal ity, 21, 107± 12 0. Smith, T.W., Pope, M.K., Rhodew alt, F., & Poulton, J.L. (1989) . Optimism, neuroticism, coping, and symptom reports: An alternative interpretation of the Life Orientation Test. Jour nal of Person ality and Social Psychol ogy, 56 , 640± 6 48. Taylor, S.E., Ke meny, M.E., Aspinwall, L.G., Schneider, S .G., Rodriguez, R., & Herbert, M. (1992) . Optimism, c oping, psychological distress, and highrisk sex ual behavior among men at risk for

56

LAI ET AL.

Acquired Immunode® ciency Syndrome (AIDS). Jour nal of Personal ity and Social Psychol ogy, 63, 46 0± 473. Triandis, H.C., & Marin, G. (1 983). Etic plus e mic versus pseudoetic: A te st of a basic assumption of contemporary cross-cultural psychology. Jour nal of Cross-Cultur al Psychology, 14, 489± 500. Wallston, B .S., Wallston, K.A., Kaplan, G.D., &

Maides, S .A. (1976) . Development and val idation of the Health Locus of Control (HLC) S cale. Jour nal of Consulting and Clinica l Psychol ogy, 44 , 580± 585 . Yang, K.S., & B ond, M. (1 990). Exploring implicit personality theories w ith indigenious or imported constructs. Jour nal of Personal ity an d Social Psyc holog y, 58, 1087± 1095.

Copyright of International Journal of Psychology is the property of John Wiley & Sons, Inc. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.