Scribd
Upload
1K views14 pages

Gibbons and Buunk

Gibbons and Buunk

Uploaded by

Danijel Vasilj
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
1K views14 pages

Gibbons and Buunk

Gibbons and Buunk

Uploaded by

Danijel Vasilj
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 14
opie 18 yh ee Pohl eet Individual Differences in Social Comparison: Development of a Scale of Social Comparison Orientation Frederick X. Gibbons Towa State University Bram P. Buunk University of Groningen Development and validation of a measure of individual difereaces in social comparison orientation (the Towa-Netherlands Comparison Orientation Measure (INCOM) are described. Assuming thatthe ten: dency toward socal comparison is universal, the scale was constructed so as to be appropriate to and comparable in 2 cultures: American and Dutch. It was the administered to several thousand people in each country. Analyses of these data are presented indicating thatthe scale has good psychometric properties In addition, a laboratory study and several Field studies are describes that demonstrated the INCOM's ability to protie comparison behavior effectively, Possible uses of the scale in basic and applied setings ae discussed ‘Few statements in social psychology have generated as much research interest as Hypothesis 1 of Festinger’s (1954) social comparison theory: “There exists, in the human organism, a drive to evaluate his opinions and abilities” (p. 117). Although some right quibble with his use of the term drive, most psychologists ‘would probably agree that the desire to lear about the self through comparison with others is universal. Interestingly, that does not appear to be the perception held by many people outside the 4,300), with about one third of the total completing it twee. AC the same time, it was administered to 12 samples in the Netherlands (W > 3,200), 1 of which completed it twice. The validation analyses discussed subsequently were based on these 22 samples Descriptive Statistics Means, standard deviations, and sample sizes are presented in Table | for both countries. The distribution of scores was normal ‘The mean item response was close to the midpoint of the scale (G.10 in the Netherlands and 3,60 in the United States), and the item standard deviations were fairly small (Ms = 0.68 and 0.58, respectively), Among 2,500 American college students (combined sample, described later), kurtosis was 43 and skewness was = 42. Factor Structure Exploratory analyses. Exploratory principal-components, analyses were conducted on the {item scale in samples from both countries.” The initial American sample consisted of 403 older adolescents (M age = 17 years). Two factors were extracted in this analysis, one with an eigenvalue of 4.17 and the other with an eigenvalue of 1.08; these factors explained 38% and 10% ofthe variance, respectively. A varimax rotation was then performed. ‘The first factor, comprising 6 items (5 of which concerned perfor: mance; see Appendix), was labeled “ability.” The second factor, labeled “opinions,” inchuded the remaining 5 items, As can be seen in the Appendix, none of this latter group included the word compare of comparison; 4 of te items concerned others’ thoughts or opinions. This analysis was replicated with @ later U.S. sample consisting of 847 college students that produced virtually identical results (eigenvalues: 3.77 and 1.21; explained variance: 34% and 11%). The same was true of the first Dutch college sample (N 170; eigenvalues: 4.07 and 1.50; variance: 37% and 14%). The ‘only exception in this later case was that the 11th item loaded on the first factor. Additional analyses with the other samples (includ- ing the combined college sample described subsequently) pro- duced the same basic two-factor structure. 2 Four risk behaviors were asessed in this study. The Protosype % Social Comparison interaction was significant for one of them (ceckless Arving), and the Prototype % Social Comparison X Gender interactions ‘were significant for two others (Smoking and drinking). the predicted pattem being stronger forthe mate paticpants. Finally, forthe fourth ik behavior (nefective contraception), both she two-way and the three-way interactions were significant, the patter again being stonger forthe male participants. * Analyses conducted op the eae so of 34 tems revealed the same two factors wih essentially the sare composition, long with ations factors ‘comprising the upward and downward comparison its (see Gibbons & ‘Buunk, 1998), 132 GIBBONS AND BUNK Table 1 Sample and Scale Descripives: lowa-Netheriands Comparison Orientation Measure Sample y MD aw ‘The Netherlands 1. Stadens 1m 38176498 2 Stadens 1238596378 5 Students 12-1076 5 Students oO 3701759 ks 5. Studer 780 63480 6 Therapiss Woo 3355653 BE 7 agate Leis 327078 BS 8 Adults 390 a6 9 Adults Da 368726 10, Adal TR Rat Bas 8s 41, Adults (Some depressed)? 1223587918 12. Cancer patinss yor 3078 90s 8s Unied States 1, Students (high sebood) 403 4019 68883 2. Stadente or 412757578 5. Stodents 6 4095637 4: Adolescents 220 © 400563878 5. Stodente 445404362780 6. Stodents 847 3908637180 7. Seadents Sis 3925 gai 78 Students m2 ©3980 «6702 9, Adals (female) 22 3695-67085 10, Adults (male) 1853533635 vote, All students in samples were college students, except forthe fist US. sample. * Scales administored via interactive tlevsion (Le, Telepanel: se Buunk {& Van den Eijnden, 1997, fr further deseipson).” "Incioced a subtan- al minority of depressed participant, Confirmatory analyses. Although the analyses revealed two: factors, there was also evidence that singe facto structure was Viable: Before rotation, the subscales associated with the {wo factors Gc, means ofthe sx items and the five items) comelated 61, and all items loaded highly (>.46) on the primary factor. ‘Consequently, an additional, coatirmatory factor analysis (asing LISREL 8; Joreskog & Sorbom, 1993) was performed; this anal Ys tested a single-factor and then a dua-factor solution. A com bined sample of Dutch and American college students (maximum age: 23 yeas) was created (W = 3,115), and then the analysis was conducted onthe two groups together. Resuls indicated thatthe ‘two-factor solution tthe data very well (goodness-of-fit index [GF] and adjusted goodness-of-fit index [AGFI] both >.95). It so fit the data beter than the one-factor solution, as indicated by 2 significant change in chi-square value, which dropped from 791.6 in the single-factor solation to 520.2 in the two-factor solution (with one degree of freedom, this drop was significant at P< 001), Once again, however, there was evidence of the Viability of the single factor. Specifically, the correlation between the ability and opinion factors inthe two-factor solution was very high (.79), the single-factor solution also fit the data well (GF and AGFT both > 92), and al items loaded highly (>.31) on the single factor. In shor, the confirmatory factor analysis indicated that the INCOM comprises two distinguishable factors that are very highly related. Reliability Internal consistency. Cronbach's alpha in the original sample ‘was .83. All 1 items produced corrected item-total correlations greater than .36;climination of any one of them reduced the alpha. ‘The alpha was very consistent accoss the other samples (see Table 1), ranging from .78 to .85 in 10 American samples and from .78 to 84 in 12 Dutch samples, levels that are considered good (Nonnaly, 1978). Temporal stability. Temporal stability was assessed on six different occasioas inthe American samples and once in the Dutch samples. These analyses produced correlations ranging from 71 (for 3-4 weeks) up to .60 (for 1 year) in the United States and a correlation of 72 (for 7.5 months) in the Netherlands. This evel of Stability is reasonable but not as high as that for some measures, Which is t0 be expected given that the construct is sensitive t0 situational factors (described later) and therefore would be ex- pected to change somewhat over time (see Kelly & McGrath, 1988) Scale Validation Construct Validity A version ofthe known-groups validation technique was used in assessing construct validity. On the basis of previous research, we expected a difference between the two countries in terms of absolute level of CO. For example, in a recent study of American and Danish youths from rural areas (M age = 14 years), Gibbons, Helweg-Larsen, and Gerrard (1995) found that the American ad- olescens reported engaging in more social comparison than did the Danish adolescents (p << .001), as indicated by a two-item version of the three-item ad hoc scale used by Gibbons and Gerrard (1995). It was suggested that this finding was consistent with Hofstede’s (1980) observation thatthe United States is more achievement oriented than are Norther European countries such as Denmark and the Netherlands. ‘To examine this hypothesis, we compared the American and Dutch samples within the age group that was the largest and most ‘comparable: college students. A 2 2 (Country x Gender) analysis of variance (ANOVA) was conducted on the combined college-aged samples (mentioned earlier). As anticipated, the mean level of CO (see Table 2) was higher in the American than in the Dutch samples, F(1, 3055) = 39.44, p < 001. In addition, ‘women reported a level of CO that was modestly but significantly Table 2 Mean Iowa-Netherlands Comparison Orientation Measure Scores by Country and Sex: College Students Country Mea Women Combined ‘The Netherlands aM 36.78 3867 38.05 sD 669) 688, 579 * 196 403 39 nied States ow 30.16 sos 305 SD. 622 637 639 n 1.063 1397 2460 ‘COMPARISON ORIENTATION 133 higher than that of the men (M 21.60, p< 001 39.84 vs. 38.80), FUL, 3085) Trait Measures Social orientation. Although there are curently no CO scales available, several scales exist that assess a relted construct best labeled a8 social or other orientation (ie, the extent t0 which inditdvals pay stention to and base ther own behavior on the way others behave). Each ofthese scales was administered along with the INCOM to atleast oe sample, with the assumption that the coreelations would be significant. As expected (see Table 3) amoderately strong relations were found with interpersonal orien tation (Swap & Rubin, 1983; 7 = 45) and public self consciousness (Fenigstein, Scheer, & Buss, 1975; rs = 38 0 49) (The weighted average ofthese corelatons, using 0-2 transfor ration, was 43, The strongest correlations were with Lennox and Wolfe's (1984) Atention to Social Comparison Information {ATSCH Scale (rs = 47 and 65). There were also somewhat weaker conelations with Clark, Ouclete, Powell, and Milbera's (1987) Communal Orientation Scale (r= 31) and Snyder's (1974) Self: Monitoring Seale (r = ~.23). Negative afectvity. Yn addition, the dispositional factors that, have been suggested as correlates of CO, most of them related to negative affetvity, were modesty but in most cass significamly conated with the INCOM. For depression, the comeations zanged from .13 forthe Beck Depression Inventory (Beck, 1967) 10.25 (M = .19) forthe Center for Epidemiologic Studies Depres- sion Scale (Radloff, 1977). The corelaion for social anxiety (Fenigstcn eta, 1975) was 31, andthe correlation fr state-tait anxiety Spielberg Gorsuch, & Lushene, 1970) was 22. Sef extezm was comeated (nepatively) withthe INCOM in 8 ofthe 10 “American samples (weighted average = 18) and in all 5 of the Dutch samples (weighted average = —32) in which it was in- cluded, Finally, the INCOM was negatively related to optimism (Goheier & Carver, 1985) in two Dutch samples (22 and [see Footnote 4] ~.36) but only weakly related in the American sam- pls (weighted average = ~09) Newroticism. Although the relations between CO and bath depression and low se-esteem were consistent with expectations and previous hypotheses, proponents ofthe Big Five personality structure (Costa & MeCrae, 1985) have claimed that these two constructs are actually subsumed by a higher order factor: Neu- roticism. Infact, Neuroticism, asessed using the NEO Personality Inventory and its Dutch version, the Five-Factor Personality Ine ventory (Hendriks, 1997, as well as the Netherlands Personality Questionnaire (Luteyn, Stearen, & Vanbijk, 1985), correlated faily highly with te INCOM (rs ranged from .28 to 37 [ee Footnote 4: weighied average = 31). To assess the relative relations betwzen the three negative afectivity factors (Neurot- cism, depression, and low selfesteem) and CO, we conducted commonality analyses (Pedhazur, 1982) ona subse of one of the “American student samples that received al ofthese measures (n= 366). These analyses indicated the amount of variance in the INCOM explained by each factor by itself and in conjunction with cach of the other factors. Results (ee Table 4) indicated that although all thee factors were comelated with CO, only Neuroti- cism hada significant unigue relation, explaining 4.3% of the variance by itself. It would appear, then, thatthe oft-fund (or presumed) relations between social comparison and both depres- sion and self-esteem may be subsumed by their respective relations with neuroticism, State Measures We also assessed several state measures that we assumed would relate to an increased tendency in the need to know about the self, Significant but modest relations (see Table 3) were found with perceived stress (Cohen, Kamarck, & Mermelstein, 1983) and with negative affect (Watson, Clark, & Tellegen, 1988). There was only one instance (out of seven) in which the correlation with positive affect was significant Finally, one of the Dutch samples, consist ing of eancer patients awaiting or undergoing radiotherapy (Van der Zee, Bunk, & Sanderman, 1998; see Footnote 4), completed measures of health status and disease-ireatment stress (de Ha ‘Van Knippenberg, & Neijt, 1990), both of which were related t0 the INCOM, as expected. Discriminant Validity ‘The INCOM was also administered in the various samples with ‘2 number of other scales tha, theoretically, should not correlate With CO (e.., social support and need for cognition). As can be seen in Table 3, a few of these correlations were significant, but, ‘generally speaking, there was litle evidence of relations. One set fof correlations worth noting involves life satisfaction (Diener, Emmons, Larsen, & Griffin, 1985), which was typically not related to CO, Thus, those who indicated that they compared frequently ‘with others were no more or less satisfied with their life situations than were those who did not compare often (et. Diener & Fujita, 1997), Comparison of Correlations for the Two Factors To determine whether the two factors (ability and opinion) related differentially to the other scales, we ealeulated correlations separately between the two factors and all of the other scales in ‘each country (more than 200 correlations). A review ofthese pairs ‘of correlations indicated that the two subscales correlated very similarly with all of the scales with one exception: The ability subscale correlations were consistently about .20 or .25 higher (absolute value) than the opinion subscale correlations for the various negative affectivity scales (ie., self-esteem, depression, and neuroticism). This was true in both countries. In other words, individuals who were low in self-esteem, high in depression, or high in neuroticism were more likely to engage in ability-based comparison but no more or less likely to engage in opinion-based ‘comparison than those who were at the opposite levels of these dimensions. One possible explanation for this is that individuals ‘who are high in neuroticism, who tend to have unstable self concepts, are primarily interested in self-relevant information that is at least potentially verifiable. Comparison on abilities is more likely than comparison on opinions to provide that type of infor- “The later corelation involved the Dutch cancer patent sample (Van der Zee, Buunk, & Sanderman, 1998). The particular situation of this ‘simple, consisting of eancer patients awating or undergoing radiotherapy, ‘was unique relative wo that of participants inthe other samples, 134 GIBBONS AND BUUNK Table 3 Correlations of the lowa-Netherlands Comparison Orientation Measure With Other Scales: Dutch and American Samples Seale Conelaton(s) Weighted average Sample! atch samples Seit-etem (Rosenberg) 16" 10 510+ -2 235.812 Social desirability (Marlowe-Crowne) <4 2 [Neuroteism (Luteyn eta, 1985) sree 0 Big Five (ends, 1997) Neuroiism aoe, spe 36 Burversion 7,09 1 Agreebleness 14-24 % Canscintiousness 17-25" -a Openness ate, = 198 38 Depression (Beck Depression Taventory) 3 Depression (CES-D) Daves, 256 2 Positive affect (PANAS) 07,15 10 14h, ape 2B [Negative affect (PANAS) ‘Anxiety (Stato-Trat Anxiety Inventory) Iteral-exteral contol Rete, 1966) ‘Mastery (Poatlin & Schooler, 1978) Public selconsciousness (SCS) Private self-consciousness (SCS) ‘ATSC Exchange orientation (Clank etal, 1987) Communal oenttion Well-being (Diver eta, 1985) Optimism (LOT) 28 [Need to evaluate (lars & Pety, 1996) Coping active: Carver et al, 1989) 2 “Health (ad hoc scales) ‘Health uncertainty (de Haes eta, 1980) ‘Ape (adults) Education (adults) Setfesteem (Rosenberg) =38 19 Social desirability (Mariowe-Crowne) 6 Social desirability (Eysenck Lie) 3 [eurotcism (Eysenek) 5 Big Five (Corts & McCrae, 1992) ‘Neuroticism gees 1 Bucaversion 2 7 Agreableness 08 1 Canscientousess o 7 Openness o 7 Depression (Derogatis, 1983) meee 2 Depression (CES-D) ie 7 Positive affect (PANAS) = 21" © -01 -o1 1.2.4,.9,10 Negative affect (PANAS) 15* 0.29%" 2 1249, 10 Social anxiety (SCS) aise 6 Public self.consciousness (SCS) BEF to apes a8 1246 Pvate setfconsciousness (SCS) 22¢ 0 430" 3 1246 ATSCI apes 7 Interpersonal orientation (Swap & Rubin, 1983) ase 3 Sele monitoring (Snyder, 1974) 250 6 Well-being (Diner eta. 1985) = 13+ to 08 071 1,24 Optinism (LOT) ~Do—12% -09 12,67 [Need for cognition (Casioppo & Pety, 1982) =08" to — 156+ 12 3.5.6 Perceived stess (Cohen et al, 1983) 23 9,10 Social suppor (Curona & Ress, 1987) t Ipalsvity CWils et al, 1998) 1 -Erotphobia (Whit ta, 1977) 35 [Negative life events (Sarason, Srason,Poter, & Antoni, 1985) 1.2,9,10 Health (ad oe scales) 06 1,2,9,10 Religion (ad hoe scales) 03 12,9,10 ‘Ecaton (als) 00 310 Age (adults) 8 9.10 0 Financial status (ad hoc seals) Note. CES-D = Centr for Epidemiologic Smdies Depression Scale; PANAS ‘ATSCI = Aneatioa to Social Comparison Information Scale; LOT = Life Orientation Test. See Table | for sample mumbers and descriptions. spe. Mpc ol. ep < 001 Positive and Negative Affect Seales; SCS = Self- Consciousness Seal; COMPARISON ORIENTATION 135 ‘Table 4 Commonality Analysis for the lowa-Netherlands Comparison Orientation Measure With Neuroticism, Self-Esteem, and Depression Factor Variance Explained (%) Depression (unique) 00 Self-esteem (unigue) 00 ‘Neuroiism (unique) 45 Depression and self-esteem (shared) 00 Depression and Neurovicsm (shared) 08 SetFestem and Neurotcism (sired) 19 AN shared) to Tol 83, ‘Note, The analysis involved American Sample 7 (see Table 1; among all Participants who completed the four scale). ‘mation. Although this patter i of interest, we did not anticipate it, land e consider it to be beyond the scope of this article. Future ‘studies might further examine this difference, however. Social Desirability ‘As noted earlier, previous researchers have suggested the exis- tence of normative sanctions against acknowledging or admitting social comparison (Hemphill & Lehman, 1991; Wood, 1996). ‘Thus, there was some reason to expect negative relation between the INCOM and measures of social desirability. In fact, these correlations were also small. The correlation with the Marlowe Crowne scale (Crowne & Marlowe, 1960) was nonsignificant in the Netherlands (—.14) and weak inthe United States. (rs = =.12, p< .05). The correlation with the Eysenck Lie scale (Eysenck & Eysenck, 1975) was .08. Thus, concems that responses on the INCOM might be strongly influenced by social desirability mo- tives appeared unfounded. Criterion-Related Validity Four studies have been conducted that have assessed the citerion-relaed validity of the INCOM. The first was conducted specifically for that purpose and is described in full detail here. ‘The other three are described in more detail elsewhere and so are discussed only briefly. Study 1: Social Comparison After Test Performance ‘The first study was an experimental laboratory investigation of, social comparison behavior after test performance. A variation of 4 procedure scen in several previous laboratory studies (c.g. Pyszezynski, Greenberg, & LaPrelle, 1985) was used in which participants received bogus feedback on atest they had taken and then were given an opportunity to see the test scores of previous test takers. The prediction was that those with high INCOM scores would be more interested in seeing how others performed than would those with low scores, Method: Participants and procedure. Participants were students en rolled in tower level psychology courses at fowa State University who took atin mass pretesting sessions atthe Begining ofthe semester in which the INCOM was included. Fifty students, 25 each fom the top and botom thins ofthe INCOM distribution Scores of less han 36 or reter than 43), wore called and asked @ come to the laboratory. From this group, 2 individuals were eliminated because they had trouble with the compute, and 10 others were eliminated as a result of suspicion of the feedback $ ‘This eft atral of 19 low CO students (10 men, Mf INCOM score = 32.2, and 9 women, M score = 323) and an equal aumber of high CO students (© men, Mf score = 468, and 10 women, M score = 475) ‘The cover story suggested thatthe siody was intended to assess the relation between two consiuts. Each construct was to be measured sepaatly the fist by atest calle the Wilder Proxima Parts Tes ad the second by computer. The Wilder was a bogus sea created forthe study that contained five sections (e.. vocabulary and “emotional maturity”) with tot of 38 items. Paricipant ook par in groups of 4. After having he 4 people inroduce themselves, the experimenter escorted them o separate cubicles where they received instrctions for the rest ofthe study hough an intercom. They ‘were given 13 mi to work on the Wilder and then waited for 3 min while the test was supposedly scored, Each participant received a score of 37 and was told individually that she oF he had done “prety well” somewhat shove average. Patcipants were then told that they would pair up forthe second part ofthe study, which would ocur in sequence, with one person working oo the computer fist while the other waited, after which the two roles would switch The third part supposedly involved the ewo of them having 2 discussion together abou sever relevant topics. Esch prticipant was then informed that his r er partners been selected to do the next at Fist. This fet our 8 min during whic he first wo ofthe three social ‘ompison measures were clleted. ‘These fist two social comparison measures were assessed on the com> puter using the Micro Experimental Laboratory (MEL) program (Scho ‘der, 1988). Participants were tod that while they waited for their partner fo finish, they could, if they waned, look at information on the computer out how others had done of the Wilder text. Tete were 14 screns fall ‘of soch (bogus information. The frst 7 contained information on mean Scores for diferent groups of participants el college students or male stadents from fowa Stat), The second group of 7 sreens contained the ‘aicipan’s score es well a information onthe individual scores ofthe Tas 10 groups of participants, listed by identification numbes. After dem ‘onstatng use of the computer, the experimenter departed, leaving the participant alone. The MEL program then measured time spent viewing. each ofthe sreens ‘The last sereen reminded the patsipant thatthe third part ofthe study involved a disession with his or her partner on several issues relevant 10 ‘the two constructs and then iniested that participants would be able to choose the discussion ropics. To do tit, hey were to open the foer ia front of them and answer the questions contained therein. Each question pertained 1a possible discussion topic, followed bya scale ranging fom | (not at all interested) 0 7 very interested. The fs item, which com ‘rise the third dependent measure, was a face-valid indicator of social ‘comparison interest the topic was “performance on tests.” The nex ree were filler items (es. pois. ‘Results and discussion. — As a result oftheir open-ended nature, the exposure time data were skewed and, therefore, subjected to a ‘natural Jog transformation, Then a 2 (sex) x 2 (high-low social ‘comparison group) multivariate analysis of variance was con- «ducted on the three primary dependent measures, which were the ‘mean (transformed) exposure times for the two sets of normative information on the Wilder test and the response on the first discussion topic question (see Table 5). This analysis revealed a ‘main effect of social comparison group, F(3, 28) = 5.55, p <.008, * Most of the suspicious participants had been involved in a study with similar (bogus feedback) deception in the same laboratory within a few ays before their participation inthis study. 136 GIBBONS AND BUUNK ‘Table 5 ‘Mean Exposure Time to Previous Performance Information and Interest in Discussing Test Performance With Parmer as a Function of Gender and Comparison Orientation (CO) Level: Study 1 C0 tevel Low High Messure Male Combined Mate Female Combined Exposure time Section 1” uw 352 313 333 407 401 404 sD 083 oo 078 033 076 oss ” 10 ° 9 9 10 9 Exposure sme: Section Mw 190 1.86 188 201 307 302 sD 1B L12 Ls 036 103 os " 0 9 19 9 10 » Performance discusson® uw 48 as 431 450 3.60 sul sD 069) 142 Lis 076 108 108 . 7 5 16 8 10 8 "Mean exposure time (in seconds) to the first st of screen with information about previous groups west peformancs. ‘et of seeens with information about previous individuals” test performances in the anticipated direction. There was no main effect of sex, and ‘there was no interaction (ps > .13). Univariate ANOVAS indicated that the high CO group mean was higher than the low CO group ‘mean on each of the three measures, Fs(1, 34) = 10.26 and 12.44, PS <.003, and F(1,30) = 4.90, p < 04, respectively. Finally, the correlation with INCOM score was significent for the two expo- sure times, 15(36) = 33 and 48, ps < .05 (wovtiled), and ‘marginal forthe topic preference item (r = .27, p = 12). Thus, the desire to look at previous participants’ test performance and the desire t0 discuss current performance with the partner were asso- ciated with scores on the INCOM, as expected, Study 2: Social Comparison and Reactions to Downward Comparison In this study, three experiments were conducted that examined the impact of encouraged downward comparison on couples’ sat isfaction with their relationships (Oldersma, Buunk, & De Dreu, 1997), Two ofthe experiments included the INCOM. Inthe first of the two, members of couples (M age = 41 years) in the control condition were asked to generate as many reasons as they could ‘why their partner was a good (relaionship) partner. Those in the experimental condition were encouraged to compare their partner with other partners. Specifically, they were told to generate re sons why their partner was better than most partners, Before doing this, they answered questions on relationship distress. Afterward, participants were asked to indicate how satisfied they were with their relationship. As expected, CO tendency and relationship stress together moderated the impact of social comparison on relationship satisfaction. That is, downward comparison had a more positive impact on those members who were (a) experiencing discontent with their relationship (Le., were threatened cf. Aspin= wall & Taylor, 1993; Gibbons & Boney-MeCoy, 1991) and (b) high in CO tendencies. These participants reported the greatest * Moan exposure to the second Totes in discussing test performance with panes (scale = 7). Increase in relationship satisfaction after the comparison opportu- nity, The second experiment involved college students (M age 22 years) and avery similar procedure, and it produced essentially te same results. Once again, CO interacted with comparison opportunity and relationship discontent in predicting alleviation of this discontent. Thus, eltionship-isressed individuals who were high in CO reported the largest increase intlaionship satisfaction ater the comparison, Study 3: Social Comparison and Responses to Vivid Comparison Information In pwo studies, health care workers (ie. sociotherapists treating criminal offenders, along with murses) were confronted with Vivid description of either an upward comparison or a downward, comparison target (Buunk et al, 1998). In the first study, con- acted with nurses at an academic hospital, level of professional ‘bumout (e.g. perceptions of personal accomplishment at work) was first assessed, and then participants were presented with a (bogus) description ofa colleague followed by a mood assessment, ‘The description, which was based on actual in-depth interviews with members ofthe target population, indicated thatthe colleague was pesforming either very well or poorly at her job. As predicted, CO, occupational burnout, and comparison direction together moderated the impact of comparison on mood. That is, burnout and CO did not have main effects on responses to upward com- parison, nor was there an interaction effect (ef. Buunk, Collins, Taylor, Van Yperen, & Dakof, 1990; Van der Zee, Buunk, & Sanderman, 1998). However, confrontation with a downward com- parison target generated the most negative affect among those who ‘were (a) high in professional burnout (Le. ad a low professional selfconcept) and (b) high in CO. The second study, among sociotherapists, produced essentially the same results, Again, CO had no main or interaction effects on COMPARISON ORIENTATION 137 responses to upward comparison, whereas affective responses to downward comparison varied as a function of CO. Specifically, among individuals who were higher in burnout, downward com- parison generated more negative affect if they were also high in CO; burnout was more or less unrelated to negative affect among, those low in CO, ‘Similar results were obtained in a recent study that also involved ‘computer-based social comparison information (as in Study 1 of this article), this time among a sample of cancer patients (Van der Zee, Oldersma, Buunk, & Bos, 1998). The information was inthe form of transcripts of interviews with other cancer patients in which they discussed how well they were coping with their illness. As expected, those patients with high INCOM scores chose to read ‘more compatison information and showed more affective reaction to that information than did those with lower INCOM scores. ‘Study 4: Social Comparison as a Moderator of the Influence of Risk Images on Behavior ‘The fourth study (Gerrard, Gibbons, Reis-Bergan, Buunk, & Blanton, 1998) involved a conceptual replication and extension of 4 study mentioned earlier (Gibbons & Gerrard, 1995) in which three-item social comparison seale was shown to moderate the impact of risk prototypes on changes in risk behaviors. This time the INCOM was used, along with « messure of alcohol prototypes, to predict changes in alcohol consumption among a sample of ‘more than 400 adolescents (M age at Time 1 = 16 years). In both cases, prototypes interacted with CO as expected; specifically, risk prototypes predicted increases in risk behavior only for those participants who were high in CO. Discussion Using measures that previous studies have shown to be indica- tive of comparison activities, such as exposure to (performance) comparison information and changes in affect after comparison ‘opportunities, these four studies provided evidence that individuals who were high in CO were engaging in more social comparison than were chose wo were lower on the scale. AS a group, then, the studies attest to the exterion-elated validity of the INCOM, General Discussion ‘The structure of the INCOM was generally consistent with ‘previous discussions of social comparison processes dating back to Festinger (1954). The frst scale factor reflected an interest in performance or ability-related comparison, whereas the second factor reflected interest in comparison based more on opinions. Although the two factors were discriminable, we would caution, researchers against using them independently in the future for wo reasons. First, although the two-factor structure emerged across ‘numerous samples, a single-factor structure also fit the data fairly ‘well. Second, the two factors did correlate highly with one another, and so we believe they are measuring the same basic underlying process, Infact, we view the two types of comparison—abilities ‘and opinions-~as two sides of the same coin: seeking information from others to increase self-understanding. For these reasons, we ‘would recommend that researchers use all 11 items whenever possible, If space isa problem, however, the first factor items will do a reasonable job as a proxy for the entire scale (the mean ‘correlation of the 6 items with the total was .92 across the 22 samples, all rs > 90). This would be especially ue in situations in which the basis of comparison is strictly or primarily abilities or performance. Scale Characteristics and Relations With Other Measures Other orientation. ‘The scale correlated weakly or not at all with a variety of individual-differences measures that, theoreti- cally, should have been distinct (e.g, impulsivity, religiosity, and life satisfaction). More informative isthe evidence of convergent validity provided by the correlations with other theoretically rel- evant measures. Fist, there is an element of conformity and other orientation in social comparison, and that orientation is reflected in the INCOM’s relations with scales such as the Public Self- Consciousness Scale and the ATSCI (which appears to be a ‘measure of other orientation and conformity; see Bearden & Rose, 1990). At the same time, there is more to social comparison than, simply other orientation; i is a perspective on others vis-2-vis the self, In fact, none of these other scales assess this primary social comparison motive of self-evaluation per se, which is probably why the corelations with them were moderate Negative affecvity. A second consistent relation was that between the INCOM and various measures of negative affectvity, both state and trait. The former relations suggest that the disposi- tion itself is sensitive to acute situational factors (as Festinger, 1954, suggested), probably more so than a number of other more stable traits, such as extraversion. Further evidence of this isthe INCOM’S test-retest reliability, which, although acceptable, was modest. The desire to socially compare should increase during periods of heightened uncertainty, and there are many situational factors that could increase self-uncertainty. As a result, even though the tait may be fairly stable, it will reflect temporary contextual influences. Neuroticism. The negative affectivty trait that had the stron- gest relation with the INCOM was Neuroticism (cf. P. Gilbert & Allan, 1994; Van der Zee, Buunk, & Sanderman, 1998; Van der Zee, Buunk, Sanderman, Botke, & Van der Bergh, in press). In fact, the commonality analysis indicated that the (positive) rela- ‘tions between social comparison and the other negative affectivity traits (ow self-esteem, depression, and anxiety) were attributable to their relations with Neuroticism (others have suggested a similar hierarchical structure for these three traits relative to Neuroticismm; eg., Watson, Clark, & Harkness, 1994). Given that a primary component of Neurodcism is uncertainty about the self (Costa & ‘McCrae, 1992), its relation with social comparison is theoretically consistent Social desirability. Generally speaking, we did not detect much of the reticence 10 acknowledge social compaison behavior that others have noted or inferred (Hemphill & Lehman, 1991; Taylor etal, 1983; Wood, 1996), Fist, the correlation between the INCOM and the social desirability scales was low. Second, most participants appeared quite willing to acknowledge their compar- ison tendencies. The mean response on the academic comparison frequency item (“How often do you compare your test scores with cothers?”) was well above the midpoint. Moreover, the mean re- sponse on the S-point scales for the 11 items was typically be- tween 3.0 and 3.5, which, although not hearty endorsement, is not 138 GIBBONS AND BUNK 4enial either (keeping in mind thatthe scale is worded somewhat in the extreme, with terms such as never and always). One reason for this may have to do with the preface tothe scale, in which we attempted to sanction and normalize the behavior (as others have suggested; cf. Wood, 1996). The most notable exception to this tendency toward acknovledgment was in the Dutch cancer patient sample, which had the lowest mean. It may not be coincidental that most of the previous studies detecting social comparison reticence involved people with serious health problems as well (Helgeson & Taylor, 1993; Hemphill & Lehman, 1991; Schulz & Decker, 1985; Wood et al., 1985) Instead, it may be that social comparison among those with serious illness is thought to be inappropriate, is actually reduced, or both. This issue will most likely attract future empirical atention in studies using the INCOM. Downward comparison. Interestingly, the mean item response on the downward comparison scale (Gibbons & Buunk, 1998) was quite a bit lower (typically at 2.7) than that oa the INCOM, which is consistent with a belief that people ate less willing to admit this particular kind of comparison (Hemphill & Lehman, 1991; Wills, 1981). Of course. it may also be the case that downward compar- ison is simply Jess common than social comparison in general (Gibbons et al, 1997) or upward comparison (the item mean on the upward comparison scale was 3.1). Once again, the INCOM and the two directional-comparison scales should facilitate future at tempts to address these types of questions. Cross-cultural differences. As suggested in a previous study (Gibbons et al, 1995), the U.S. sample was more comparison criented than was # comparable Northem European sample. Other ‘than this overall mean difference, however, there were relatively few distinctions between the two countries in terms of scale characteristics. This speaks to the universality of the trait, as suggested earlier. and to the utility of the measure. As one would hhope with any scale, the INCOM should prove to be a good barometer of cultural and ethnic differences in social comparison ‘behavior rather than just a consequence of such differences Summary. The prototypical image of a high comparer sug- ‘gested by the various relations of the INCOM with other scales is of an individual (in either culture) who (a) is interpersonal more than introspectively oriented, being sensitive to the behavior of others, and (b) has a degree of uncertainty about the self, along ‘with an interest in reducing this self-uncertainty and, in so doing, improving. Of course, typicality is not universality, and these related attributes are just that; together, they do not explain more than 25% of the variance in the INCOM. Thus, the scale does appear to be assessing a unique and distinct trait that is manifested in a number of predictable and observable bebaviors. Scale Uses We see three primary research areas in which the scale can be used: basi, applied, and intervention. Basic research. Social comparison is not an easy construct 10 ‘measure in laboratory settings. The automaticity ofthe process (ct. D. T. Gilbert et al, 1995; Wood, 1996) makes control or manip- tulation of comparison in the laboratory difficult. Moreover, the reticence to admit social comparison, mentioned earlier, tends to undermine the credibility of self-repors of comparison activity or, actually, lack of activity (ie., denial). As a result, many social ‘comparison researchers have relied on indirect assessments, ether bby varying comparison opportunites (e.g. target availability or affiliation preferences) or by inferring comparison activity from other observable reactions, such as mood change or changes in selfassessments. A reliable CO scale provides another tool— albeit also indirect—for assessing social comparison processes. If individuals known 10 be high in social comparison tendencies report more reaction (e.g., mood change) in the presence of a comparison opportunity but notin its absence, then a researcher's confidence that comparison has, infact, ocurred would be signf- ‘cantly increased. Such was the casein the studies described in the Criterion-Related Validity section ofthis article. More generally, being able to discriminate among those who do and those who do rot engage in a particular process should facilitate efforts to determine what itis they are doing (in other words, to understand the process itself). Applied research. ‘There are @ myriad of applied empirical questions involving comparison processes as well. Many of them, although certainly not all involve issues in health psychology (cf. Buunk & Gibbons, 1997). For example, social comparison is, thought to be an integral part of the coping process for such stressors as pain, surgery, and illness (e.., Kulik & Mabler, 1997), and it is thought to be involved in pain assessment and illness imerpretation (Leventhal ct a, 1997). The same is true regarding decisions involving medical care (Suls et al., 1997). Social com- parison is also believed tobe an important element in the etilogy ‘of depression (Ahrens & Alloy, 1997) and other emotional prob- lems, such as professional burmout (Buunk & Ybema, 1997; Buunk etal, 1998). In each instance, the existence of a valid measure of social comparison tendencies should serve to further understand- ing ofthe various behaviors in question. In addition, to the extent that comparison processes influence academic performance and Teaming, a5 appears to be the case (Blanton, Buunk, Gibbons, & Kuyper, in press; Gibbons etal, in pres), one would assume that educators and education researchers would find a measure of social comparison useful Interventions. Another type of applied usage involves the de- sign of more effective interventions. Research on health-risk be- haviors (Gibbons & Gerrard, 1997), for example, has shown that risk images or procotypes have a stronger impact on the associated health behaviors of young people who are high in social compar- ison tendencies (as was shown again with alcohol consumption in Gerrard et al,, 1998, discussed earlier). Comparison with risk images is also related to adults’ success in smoking cessation (Gibbons & Eggleston, 1996). In adition, these influential images hhave been shown to be malleable (cf. Eggleston, 1997), which ‘means thatthe opportunity is there to change the related behaviors tuough image alteration. Knowing ahead of time which adoles- cents or adults are most likely to respond to such efforts should prove useful. Similarly, research by Klein and his colleagues (Klein, 1997; Klein & Weinstein, 1997) has shown that both perceptions of risk or danger and the subsequent willingness to lake these risks are influenced by self-other comparisons (cf. Misovich, Fisher, & Fisher's, 1997, discussion of perceptions of AIDS risk, in particular). Effors to alter these perceptions or to disabuse young people of potentially dangerous cognitive misper: ceptions that are socially based (e.g. optimistic bias; Perloff & Feizer, 1986; Weinstein, 1980) should also be facilitated by the ability to ascertain ahead of time who compares extensively and who does not. COMPARISON ORIENTATION 139 ally, for 2 number of health behaviors (eg. recovery from ‘cardiac surgery; Berkhuysen, 1994), as well as Jeaming new cop- ing skils, social comparisons may actually interfere in that indi viduals focus too much on how they are doing vis--vis others rather than on their own improvements. Identifying individuals with high comparison tendencies may be useful in developing specific interventions. In short, given the fact that CO is not a ‘completely fixed characteristic, it might even be possible to de- velop interventions that reduce, in some situations, the tendency to compare oneself with others. Given the ubiquity of the construct being measured, we believe the utility ofthe scale that measures it should be significant as wel. References Aleck, G., & Tennen, H. (1991). Social comparison and coping with rmsjor medical disorders. In J. Suls & T. A. Wills (Eds), Socal com parison: Contemporary theory and research (pp. 369-395), Hillsdale, NJ Elba, Alnens, A. H., & Alloy, L. B. (1997). Social comparison proceses in depression. in BP. Bunk & FX. Gibbons (Eds), Health coping, and wellbeing: Perspectives from solal comparison theory (pp. 389-410), Mahwah, NJ: Esbaum, Aspinwall, LG. (1997), Furure-oviented aspects of socal comparisons. In BB. P. Buunk & F. X. Gibbons (Fs, Health, coping. and well-being Perspectives from social comparison theory (pp. 125-166). Mahwah, NE Enbau, Aspinwall, L.G., & Taylor, S. B (1993). Effects of social comparison ‘iecton, thes, and self-esteem on affect, self-evaluation, and expected succes. Jounal of Personality and Social Psychology, 64, 708-72, ‘Bandura, A. Jourdan FJ. (1991). Selfregulatory mechanisms govern- ing the impact of social comparison on complex decision making. Journal of Personality and Social Psychology, 60, 941-951 Bearden, W. 0., & Rose, RL. (1990). Attention to social comparison informstion: An individu diference factor affecting consumer confor- rity. Joumal of Consumer Research, 16, 461-471 Beck, A. T. (1967), Depression: Clinical, experimental, and theoretical ‘aspects, New York: Harper & Row. ‘Berkhuysen, M. (1994, March). Social comparison among cardiac patients involved i rehabilitation. Paper presented atthe Kurt Lewin Taste ‘Workshop on Social Comparison and Relative Deprivation, Schiermon= rikoog, the Netertands Blanton. H., Buunk, B.., Gibbons, F.X., & Kuype, H. (in press). When bettertham-others compares upward: The independent effects of ‘compison-ehoice and comparson-evaluston on academe. perfor: ‘mance, Journal of Personality and Social Psychology. Brickman, P, & Bulan, R. J. (1977). Pleasure and pin in social com: parisoa” In J. Suls & RL. Miler (Eds), Social comparison processes: Theoretical and empirical perspectives (pp. 149-186). Washington, DC Hemisphere ‘Buunk, BP. (1998), Social comparison processes under stess: Towards an integration of classic and recent perspectives. In W. Suoebe & M. Hewstone (Eds), Ewopean review of socal psychology (Vol. 5. pp 211-241), Chichester, England: Wiley. Buunk B.P, Collins, RL, Taylor, S. E, Van Yperen, N. W., & Dako G. A. (1990, The affective consequences of socal comparison: Either rection has its ups and downs. Journal of Personality and Social Paychology, $9, 1238-1249, Buunk. B. P.. & Gibbons, F. X. (Eds). (1997). Health, coping. and ‘welding: Perspectives from socal comparison dors. Mahwah, Ni Erlbaum, Buutk, B. P, Schaufeli, W, B, & Ybema, J, F. (1994), Burnout, uncer: tainty and the desire for socal comparison among nurses, Journal of Applied Social Paychoogy, 25, 1499-1317. Buunk, B.P, & Van den Eijnden,R.J.J.M. (1997), Perceived prevalence, recived superiority, and relationship satisfaction: Most relationships tue good, but ous is the best. Personality and Social Psychology Bulletin. 23, 219-228, ‘Buvnk, B.P. & Ybema, JF (1997). Socal comparisons and occupational stress: The identifestioncontast model. In B. P. Buuok & F. X Gibbons (Eds), Health coping. and well-being: Perspectives from so cial comparison theory (pp. 359-388). Mahwah, NJ: Erlbaum ‘Buunk, BP. Ybem, JF, Van der Zee, K., Ipenburg, M. L, Schaufeli W. B., & Gibbons, F.X. (1998). Affect generated by social comparisons a related to professional burnout and individual differences in social ‘comparison orientation. Manuscript submited for publication. (Cacioppo J.T, ety, R. F.(1982). The need for cognition. Joumal of Personality and Socal Psychology, 42, 16-131 Camptell, J D. (1990), Sefesteem an the clarity of the self-concept. Journal of Personality and Social Psychology, 59, 538-849, Carve, C.S, Scheier, MF, & Weintraub, JK (1989). Assessing coping strategies: A theoretically based approach. Jounal of Personality and Social Psychology, 56, 267-283 (Clark, M.S, Ouellet, R, Powell, MC. & Milberg, S. (1987). Resipi- ‘ents mood, relationship type, and helping, Journal of Personality and Social Psychology, 53, 94-103, Coben, 6. Kamarck, . & Mermelsei, R. (1983) A global measure of perceived sess. Jounal of Health and Social Behavior, 24, 85-396. CComrey, A. L. (1988) Faetor-analytic methods of sale development in personality and clinial psychology. Journal of Consulting and Clinical Psychology, 36, 784-161 Conta, P.T, Jr, & MoCrae, RR, (1985), Hypochondriass, neuroticism, and aging: When are soma complaints unfounded? American Psychol gis, 40, 19-28 Costa, P. TJ, & McCrae, RR, (1992). Revised NEO Personality Inventory (NEO-PLR) and NEO Five Factor Inventory (NEO-FFD) professional manual. Odesss, FL: Psychological Assessment Resources Crowne, DP, & Maziowe, D, (1960). Anew scale of social desirability independent of psychopathology. Journal of Consulting Pycholagy, 24, 349-354 Cutona, C- E., & Russell, D. W. (1987). The provisions of social rela: ‘ioaships and adaptation to stress. In W. H, Jones & D. Perman (Eds), Advances in personal relationships (Vo. 1, p. 37-68). Greenwich, CT IAT Press. de Haes, J.C. J. M., Van Knippenberg, F. C.B., & Nei. P. (1990) ‘Measuring psychological and physical distress in cancer patients Struc ‘ure and application ofthe Rocerdam Symptom Checklist British Jour al of Cancer, 62, 1034-1038 Derogatis, LR. (1983). SCL-90-R: Adminisraton, scoring. and proce: ‘dures. Towson, MD: Clinical Psychologists Research. DeVellis, RF. (1991), Scale development: Theory and applications ‘Newbury Pat CA: Sage Diener, E, Emmons, R. A., Larsen, R. J. & Griffin, $, (1985). The Satisfaction with Life Scale. Joumal af Personality Assessment 49, 1-45 Dieoer, E., & Fujita, F (1997). Social comparisons and subjective well being. In B, P. Buuk & FX, Gibbons (Eds), Health, coping, and wellbeing: Perspectives from socal comparison theory (pp. 329-358). Mahwah, NJ: Felbanm Bepleston, T. J (1997), Sefestem, risk behvior and reactance: The “impact of prototype modification on women's wilingress to engage in unprotected sex. Unpublished doctoral dissertation, Iowa State Unive sity, Ames. Eysenck, H. 1. & Eysenck, S.B.G. (1975), Eysenck Personality Inventory “Manual, London: Hodder & Stoughton Fenigstein, A, Scheer, M. F., & Buss, A. H. (1975). Public and private selfconsciousness: Assesiment and theory. Journal of Consulting and (Clisical Peychology, 43, 322-527 140 GIBBONS AND BUUNK Festinger, L (1988). A theory of socal comparison processes. Human Relations, 7, 117-140, Faji, F. (1995, October). There are nwo kinds of people inthe world Those who engage im social comparison dnd those who do not. Paper presented at & department colloquium of the University of Minos, ‘Urbana-Champaign, Gerard, M, Gibbons, FX, ReisBetgan, M., Buuok, B. P, & Blanton, H.C. (1998). Comparison oriemation moderates the impact of drinker Images on drinking behavior. Manuscript in preparation Gibbons, FX, Benbow, C. P., & Gerard, M, (1994), From top dog 10 ‘bottom half: Social comparison strategies in response to poor perfor- ‘mance. Joumal of Personality and Socal Psychology, 67, 638-652. Gibbons, F. X, Blaston, H.C, Gerard, M, Buuak, BP. & Eggleton, "TI. (in pres) Does socal comparison make adiference? Optimism as 1 moderator ofthe impact of social comparison on perfomance. Per- sonality and Social Psychology Buln. Gibbons, F. X., & Boney-MeCoy, S. (1991), Selfestem, similarity, and reactions to active verus passive downward comparison. Journal of Personality and Social Psychology, 60, 814~824 Gibbons, FX. & Buunk, B.P. (1998). Constructing adirecional measure ‘of comparison preference. Manuscript in preparation. Gibbons, F. X., & Eggleston, T. J. (1996), Smoker networks and the “ypical smoker": A prospective analysis of smoking cessation. Healt Paychology, 15, 469-477. Gibbons, F.X. & Gerad, M. (1995), Predicting young adults” health ik behavior. Journal of Personality and Socal Psychology, 69, 508-517. Gibbons, F. X, & Gerrard, M, (1997) Health images and their effects on heath behavior. In B,P, Buunk & F. X. Gibbons (Eds), Health, coping, land well-being: Perspectives from social comparison theory (pp. 63 94). Mahwah, NJ Ebaum, Ginbons, F. X, Helweg-Larsen, M, & Gerrard, M. (1998), Prevalence ‘estimates and adolescent risk behavior Crose-cultual differences in Social influence. Journal of Applied Psychology, 80, 107121, Gitber,D.T, Giesler, RB. & Mons, K. A. (1995). When comparisons arise Jounal of Personality and Socal Peychology, 69, 227-236. Gilbert, P., & Allan, S. (1994), Assertiveness, submissive behavior and Social comparison. British Journal of Cincal Psychology, 33, 29-306. Gitber, Price, 1, & Allan, S. (1995). Social comparison, social atrac- tiveness and evolution: How might they be related? New Ideas in Psychology, 13, 149-165. Helgeson, V. .. & Taylor, 8.E. (1993) Social comparisons and adjust ‘ment among card patients. Journal of Applied Socal Psychology 23. 1I7-1195, Hemphill, K.1., & Lehman, D. R. (1991). Socal comparisons and their affective consequences: The imporance of comparison dimension and individual difference variables. oumal of Social and Clinical Pychol- cy, 10, 372-394, Hendriks A A.J. (1997). The constuction ofthe Five-Factor Personality “Inventory (FFPI). Uapublished doctoral dsseration, University of Gro- ingen, Groningen, the Netherlands Hotede, G, (1980). Cuture’s consequences: Intemational differences it ‘work-related values. Bevesly Hills, CA: Sage Jarvis, W. B. G, & Petty, RE, (1996) The need to evaluate. Joumal of Personality and Social Psychology, 70, 172-198. Joreskos, K.G., & Sérbom, D. (1998). LISREL 8 user's reference guide Chicago: Scientific Sofware Internation Kelly, J. Ry, & MeGrath, JE, (1988). On me and method. Beverly Hills, CA: Sage. Kiein, W. M. (1997). Objective standards are not enough: AMfectve, self-evalative, and behavioral responses to social comparison informa- tion. Joural of Personality and Social Psychology, 72, 783-774 Klein, W. Mc, & Weinstein, N. D. (1997. Social comparison and unreal istic optimism about personal risk. In B. P. Buuak & F. X. Gibbons (Gis), Health coping, and wellbeing: Perspetives from social com: prison theory (pp. 25~62). Mahwah, NI: Erlbaum Kulik. A, de Mablr, H. IM. (1997. Social comparison, affiliation, and coping with acute medical threats. In B. P. Buunk & F. X. Gibbons (Gs), Health, coping, and wellbeing: Perspectives from social com- prison theory (pp. 227-262). Mahia, NI: Erbau. Lennox, RD. & Wolfe, R. N. (1988). Revision ofthe Self Monitoring Seale. Joumal of Personality and Socal Prychology, 4, 1349-1369. Leventhal H, Hudson, S., & Robitaille, C. (1997. Social comparison and health: A prooess model. In B.P. Bunk & FX. Gibbons (Eds), Health, coping. and well-being: Perspectives rom socal compariton theory (PB. 411-432), Mahwa, NJ: Eebaum. Liteya, J, Staren, H., & VanDijk, H. (1985). Hercone handeiding ‘Nederlandse Persoonitheldsoragenlisr (Revised manual ofthe Dutch Personality Questionnaire). Liste, the Netherlands: Swets & Zeitnger Marsh, K.L., & Webb, W. M. (1996). Mood uncersinty and socal ‘comparison: Implications for mood management. Journal of Social Behavior and Personality, 1, 1-26, Misovih, S.J, Fisher, JD. & Fisher, W. A. (1997). Social comparison processes and AIDS risk and AIDS preventive behaver. In B. P. Bunk EX. Gibbons (Eds), Health, coping, and wellbeing: Perspectives rom socal comparison theory (pp. 95-124). Mahwah, NJ: Erlbaum, Molleman, E, Pruya 1. & Van Knippenberg, A. (1986). Social compar- {son proesiesamoag cancer patients. British Journal o Social Psycho. 083,25, 1-13, Nunnay, J. €. (1998). Psychometric theory Cade). New York: ‘MeGraw- Hil ‘Oldersma, FL, Bounk, BP. de De Dreu, C. K. W. (1997). Downward ‘comparison as a relationshp-enhancing mechanism: The moderating oles of relational discontent and individual diferencs in social com parison orientation. Manuscript submited for publication. PeaslinL. L, & Sehooles, C. (1978) The stueture of coping. Jounal of Health and Social Behavior, 19, 2, Pedhazur, E. (1982). Multiple regression in behavioral research, New ‘York: Holt, Rinehart & Winston, Perloff, LS, Fetzer, B. K.(1986). Sef other judgments and perceived valaerabilty to victimization. Joumal of Personality and Social Psy- chology, $0, 502-510. ysaczynski, T, Greenbors, J, & LaPrlle, J (1985). Social comparison after sucess ‘nd fall: Biased search for information consistent with 4 self-serving conclusion, Journal of Experimental Social Psychol. ogy, 21, 198-211. Raloff, L. (1977), The CES-D Scale: A selfseport depression scale for esearch in the general population. Applied Psychological Measure- ment, 1, 385-401. Rover, J. B. (1966). Generalized expectancies for internal vs. extemal contol of reinforcement. Pychological Monographs, 80(L, Whole No. 609). Ruble, D.N., & Frey, K. $. (1991). Changing pattems of comparative ‘behavior assis are acquired: A funcional model of self evaluation. ileal, NJ: Exbaum. Sarason, IG, Sarason, B. R., Potter, EH, & Anton, M. H. (1985). Life events, socal support, and Hness. Psychosamatic Medicine, 47, 156— 163. Schachter, $ (1959), The psychology of afin. Stnfors, CA: Stanford Univesity Press. Scheies, MF. & Carver, CS. (1985). Optimism, coping, and heal: ‘Assessment and implications of generalized outcome expectancies. Health Psychology, 4, 219-287 Schaoidr, W. (1988). Micro Experimental Laboratory: An integrated system for IBM PC compatibles. Behavior Research Methods, Instr iments, and Computers, 20, 206-217. ‘Schoeneman, T. J. (1981), Repors of the sources of self-knowledge. “Journal of Personality, 49, 284-24,

You might also like