Journal of Experimental Social Psychology 59 (2015) 60–66

Contents lists available at ScienceDirect

Journal of Experimental Social Psychology

j ourna l homepage: www.e lsev ie r .com/ locate / j esp

Self-evaluative effects of dimensional and social comparison

Jason E. Strickhouser, Ethan Zell ⁎University of North Carolina at Greensboro, Greensboro, NC, USA

H I G H L I G H T S

• Participants learned that they performed better or worse than other people (social comparison).• Participants learned that they performed better or worse in one domain than another domain (dimensional comparison).• Both comparison types significantly influenced self-evaluations and affective reactions.• The influence of social comparison was significantly stronger than dimensional comparison.

⁎ Corresponding author at: Department of Psychology,Greensboro, PO Box 26170, Greensboro, NC 27402, USA.

E-mail address: e_zell@uncg.edu (E. Zell).

http://dx.doi.org/10.1016/j.jesp.2015.03.0010022-1031/© 2015 Elsevier Inc. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 2 December 2014Revised 11 March 2015Available online 23 March 2015

Keywords:Social comparisonDimensional comparisonSelf-evaluationSelf-concept

During self-evaluation, people compare their performance in one domain to their performance in other domains(dimensional comparison). Additionally, people compare their own performance to the performance of relevantpeers (social comparison).Most experimental research on self-evaluation has investigated the effects of either di-mensional comparison or social comparison, despite the fact that people often evaluate themselves in the contextof both standards. To address this gap, the current research examined the interplay of dimensional and socialcomparison during self-evaluation. Participants received manipulated feedback indicating that they performedbetter or worse in one domain than another domain, and better or worse than other people. Both comparisontypes significantly influenced self-evaluations and affective reactions; however, the effect of social comparisonwas significantly stronger than dimensional comparison. These findings support prior theories on the importantroles of dimensional and social comparisons in self-evaluation, but also suggest that social comparison is moreimpactful.

© 2015 Elsevier Inc. All rights reserved.

Julie receives her GRE scores and is decidingwhether or not to pursuegraduate study in a math-related field. Her math and verbal scores aresignificantly above average; however, her verbal score is significantlybetter than her math score. How should Julie evaluate her math perfor-mance? Should she be pleased that she ranked above average in compar-ison to her peers? Or should she be disappointed that she performedworse in math than in verbal? This example illustrates two fundamentalprocesses that occur when people receive feedback about their perfor-mance, namely, social comparison (Festinger, 1954; Mussweiler, 2003)and dimensional comparison (Marsh, 1986; Möller & Marsh, 2013).According to social comparison research, people compare their own per-formance to the performance of relevant peers; people who outperformtheir peers (downward comparison) typically feel better about them-selves than people who underperform their peers (upward comparison;Alicke, Zell, & Guenther, 2013). According to dimensional comparison re-search, people compare their performance in a given domain (e.g., math)to their performance in other domains (e.g., verbal); people typically feel

University of North Carolina at

better about their performance in the target domain when it rankssuperior (downward comparison) as opposed to inferior to their perfor-mance in a standard domain (upward comparison; Pohlmann & Möller,2009).

Educational theories, such as the Internal/External Frame ofReferenceModel, have long argued that dimensional and social compar-isons are fundamental determinants of self-evaluations (Marsh, 1986;Möller, Retelsdorf, Köller, & Marsh, 2011). In support of this position,numerous studies conducted by educational psychologists demonstratethe effects of dimensional and social comparison information onself-evaluations (see Chiu, 2012; Möller, Pohlmann, Köller, & Marsh,2009), but only a few of these studies are experimental in nature. Con-versely, social psychologists have conducted numerous experimentsdemonstrating the causal effects of social comparison informationon self-evaluations and affective reactions following performance(see Fiske, 2011; Suls & Wheeler, 2000), but as of present, few socialpsychologists have recognized the possibility of dimensional compari-son effects. As illustrated in our example above, people often haveaccess to both standards, and daily diary studies support the notionthat dimensional and social comparisons both occur regularly in every-day life (Möller & Husemann, 2006; Summerville & Roese, 2008). Thus,

61J.E. Strickhouser, E. Zell / Journal of Experimental Social Psychology 59 (2015) 60–66

it is critical to understand the causal effects of both dimensional andsocial comparison on self-evaluations.

In the present report, we introduce dimensional comparison theory(Möller & Marsh, 2013) to social psychology in effort to address twoprimary questions regarding self-evaluation. First, we examinewhetherdimensional and social comparisons can exert simultaneous effects inthe same experimental context. In other words, if people receive feed-back that provides both dimensional and social comparison informa-tion, do both standards exert significant effects? Research addressingthis question would contribute to a growing body of scholarship onhow people evaluate themselves in the context of multiple standardsandwhether social comparison effects continue to occur in the presenceof other standards (Wood & Wilson, 2003). Along these lines, previousresearch indicates that social comparisons influence self-evaluationsevenwhen people also have objective information (Klein, 1997) or tem-poral comparison information (Zell & Alicke, 2009). The current studycontributes to this literature by examining whether social comparisonsinfluence self-evaluations even when people also have dimensionalcomparison information.

Second, we examine the relative size of dimensional and socialcomparison effects. Although both standards have long been regardedas important determinants of self-evaluations, our novel approachallowed for the first test of whether one standard has a more potent in-fluence on self-evaluations and affective reactions than the other. Suchknowledge would be of important theoretical and practical value,given that self-evaluations of competence in specific domains predictimportant life choices such as the selection of academic majors andcareers (e.g., Ehrlinger & Dunning, 2003). Emerging research indicatesthat social comparisons are more strongly associated with self-evaluations than dimensional comparisons (see Möller et al., 2009 fora meta-analysis), but it remains unclear whether social comparisonshave a stronger causal influence on self-evaluations than dimensionalcomparisons.

Previous research

To our knowledge, only two published experiments have examinedhowpeople evaluate themselves in the context of both dimensional andsocial comparison. In one study (Pohlmann & Möller, 2009, Study 3),participants received social comparison information indicating thatthey ranked above average, average, or below average on a wordanalogies test (downward, lateral, and upward social comparison),and average on a figure analogies test. The result of this feedback wasthat their performance on the figure analogies test ranked worse,about the same, or better than their performance on theword analogiestest (upward, lateral, and downward dimensional comparison). Partici-pants in the downward dimensional comparison condition evaluatedthemselves more favorably and felt better about their performancethan participants in the lateral and upward dimensional comparisonconditions, despite the presence of social comparison information.However, the potential simultaneous effect of social comparisoninformation was not assessed in this experiment.

A related study provided initial evidence suggesting that dimension-al and social comparison simultaneously influence self-evaluations andaffective reactions (Möller & Köller, 2001, Study 3). Participants re-ceived social comparison information indicating that they performedabove average, average, or below average on a figure analogies test(upward, lateral, and downward social comparison), and dimensionalcomparison information indicating that their figure analogies perfor-mance ranked better or worse than their performance on a word analo-gies test (downward and upward dimensional comparison). Participantsin the downward social comparison conditions evaluated themselvesmore favorably and felt better about their performance than participantsin the upward social comparison conditions. Similarly, participants in thedownward dimensional comparison conditions evaluated themselvesmore favorably and felt better about their performance than participants

in the upward dimensional comparison conditions. However, this studydid not examine whether social comparison had a stronger influencethan dimensional comparison. Thus, additional research is needed toexplore the simultaneous and relative effects of dimensional and socialcomparison in self-evaluation.

The current research

We conducted two experiments to address these critical gaps in theself-evaluation literature. In both experiments, participants completedbrief tests of quantitative and verbal reasoning skills. Then participantsreceived manipulated dimensional and social comparison feedback. So-cial comparison in both studies was manipulated by informing someparticipants that they ranked 30 percentile points better or worsethan average in the target domain. Similarly, dimensional comparisonwas manipulated by telling some participants that they ranked 30percentile points better or worse in one domain (verbal reasoning)than another domain (quantitative reasoning). Thus, the absolute sizeof social and dimensional comparison manipulations in the presentresearch was largely equivalent (i.e., 30 percentile points). On thebasis of prior theories specifying that dimensional and social compari-son yield simultaneous, independent effects (Marsh, 1986; Mölleret al., 2011) as well as preliminary experimental evidence (Möller &Köller, 2001, Study 3), we anticipated that both standards would exertsignificant effects on self-evaluations and affective reactions.

Furthermore, we proposed that social comparisons would have amore potent influence on self-evaluations and affective reactionsthan dimensional comparisons. Success in a variety of contexts(e.g., athletics, school, and work) is typically defined by one's social sta-tus, that is, howwell one is performing in relation to others (Fiske, 2011;Frank, 1985). Additionally, obtaining high social status confersnumerous benefits, including better physical (Akinola & Mendes,2014; Kraus, Adler, & Chen, 2013) and mental health (Boyce, Brown, &Moore, 2010) as well as increased respect, admiration, and influencein important social groups (Anderson, Kraus, Galinsky, & Keltner,2012). Conversely, performing better in one domain than another do-main may not necessarily reap social and material rewards, especiallyif one's performance in both domains is below average. Thus, althoughsocial and dimensional comparison may be useful sources of informa-tion that people consider during self-evaluation, the influence ofsocial comparison should be substantially stronger than dimensionalcomparison.

Study 1

Our initial test of dimensional versus social comparison adopted sev-eral elements from a related study (Pohlmann &Möller, 2009, Study 3).Specifically, participants completed two tests and receivedmanipulatedfeedback about their performance on the first test, while feedback onthe second test was held constant. Social comparison information indi-cated howwell participants performed on thefirst test in comparison torelevant peers. Next, the dimensional comparison implications of thefeedback were highlighted, by stating howwell participants performedon the second test in comparison to the first test. As in the prior study,dependent measures were administered following the second set offeedback to assess self-evaluative and affective reactions to dimensionalcomparison. However, unlike the prior study, dependentmeasures werealso administered following the first set of feedback to assess self-evaluative and affective reactions to social comparison. Thus, our designnot only served as a conceptual replication of a key study conductedpreviously on dimensional comparison (Pohlmann & Möller, 2009,Study 3) but also included additional measures that allowed for anovel test of whether social comparison is more influential thandimensional comparison.

Table 1Between-Subject Feedback Conditions in Studies 1 and 2.

Feedback conditions Quantitative test Verbal test Social comparison direction Dimensional comparison direction

Study 1Condition 1 Below average Average Upward DownwardCondition 2 Average Average Lateral LateralCondition 3 Above average Average Downward Upward

Study 2Condition 1 Below average Below average Upward LateralCondition 2 Below average Average Lateral DownwardCondition 3 Above average Average Lateral UpwardCondition 4 Above average Above average Downward Lateral

Note: Dimensional comparison direction is determined from the perspective of the second (verbal) test in relation to thefirst (quantitative) test. Social comparison direction is determinedby feedback on the first test in Study 1 and by feedback on the second test in Study 2.

62 J.E. Strickhouser, E. Zell / Journal of Experimental Social Psychology 59 (2015) 60–66

Method

Participants and procedure

Participants were 122 introductory psychology students who partic-ipated in exchange for course credit at a university in the SoutheasternUnited States (89 females, Mage = 19.37, 57 Caucasian, 36 AfricanAmerican).1 Participants first completed a 15-item quantitative reason-ing test and a 20-item verbal reasoning test. Pilot testing found that thequantitative and verbal reasoning tests were of comparable difficultyand that performance on both tests was sufficiently ambiguous topromote the believability of false performance feedback.

After completing both tests, participants received manipulatedfeedback about their performance that was determined using randomassignment. First, participants received social comparison feedback,indicating that their quantitative performance ranked better than 81%,51%, or 21% of 259 test takers at their school and therefore rankedabove average, average, or below average. Then, participants evaluatedtheir quantitative performance, ability, and task-specific ability on 1(very poorly/bad) to 9 (very well/good) scales (α = .85). Participantsalso indicated how happy, sad, proud, distressed, satisfied, anddisappointed they felt about their quantitative performance on 1 (not atall) to 9 (extremely) scales (α= .91).

Next, all participants received feedback indicating that their verbalperformance was better than 53% of 259 test takers at their school andtherefore ranked average. Additionally, to highlight dimensionalcomparison, participants were told that their verbal performanceranked better, about the same, or worse than their performance onthe quantitative test, as appropriate for their condition. Participantsthen completed items assessing verbal self-evaluations (α = .82) andaffective reactions (α = .88) that were similar to those provided afterthe quantitative test. A summary of the feedback conditions is providedin Table 1 and a sample feedback protocol is provided in Table 2.

Table 2Sample feedback protocol.

Quantitativefeedback:

Quantitative Reasoning Test Score:259 students have completed the Quantitative Reasoning Test.Your performance ranked better than 81% of these students.Your performance ranked ABOVE AVERAGE.

Verbal feedback: Verbal Reasoning Test Score:259 students have completed the Verbal Reasoning Test.

Manipulation checks and suspicion probe

At the end of the study, participants completed manipulation checksasking them to recall whether their quantitative and verbalperformances were above average, average, or below average, andwhether their verbal performance was better, about the same, orworse than their quantitative performance. Participants were alsoprobed for suspicion about the feedback using a funnel debriefing. Datafrom 16 participants were excluded for failing one ormore of themanip-ulation checks and/or the suspicion probe. All reported results wereunchanged when these participants were retained (see SupplementalMaterials).

1 In both studies, datawere collected until at least 40 participantswere obtained in eachexperimental group.

Analytic strategy

In our repeated measures design, the influence of quantitative feed-back on quantitative self-evaluations and affective reactions reflects theinfluence of social comparison (see Table 1). Along these lines, feedbackindicating above average, average, and below average performance onthe quantitative test reflects downward, lateral, and upward socialcomparisons, respectively. Conversely, the influence of quantitative feed-back on verbal self-evaluations and affective reactions reflects the influ-ence of dimensional comparison. Along these lines, feedback indicatingabove average, average, and below average performance on the quantita-tive test reflects upward, lateral, and downward dimensional compari-sons, respectively. We used planned contrasts to examine whether theeffect of upward versus downward social comparison was significantlystronger than the effect of upward versus downward dimensionalcomparison on both self-evaluations and affective reactions.

Results

Self-evaluations

A 3 (quantitative feedback: above average, average, below average)by 2 (evaluation dimension: quantitative, verbal) ANOVAwas conduct-ed on self-evaluations with repeated measures on the second factor. Asanticipated, the quantitative feedback by evaluation dimension interac-tion was significant, F(2, 103) = 133.73, p b .001, ηp

2 = .72 (see Fig. 1,left panel). Follow-up analyses were conducted to decompose thisinteraction.

A significant effect of social comparison was obtained, F(2, 103) =57.70, p b .001, ηp

2 = .53, such that participants told that their quantita-tive performance ranked above average evaluated their quantitativeperformance and ability more favorably than participants told thattheir quantitative performance ranked average or below average (allgroups were significantly different, ps b .001, ds N .89). Further, a signif-icant effect of dimensional comparisonwas obtained, F(2, 103)=18.27,p b .001, ηp

2= .26, such that participants told that their quantitative per-formance ranked above average evaluated their verbal performance andability less favorably than participants told that their quantitative

Your performance ranked better than 53% of these students.Your performance ranked AVERAGE.You performed WORSE on the Verbal Test than theQuantitative Test.

2.0

3.0

4.0

5.0

6.0

7.0

8.0

Quantitative Verbal Quantitative Verbal

Self-Evaluations Affect

Below Average

Average

Above Average

Fig. 1. Mean self-evaluations and affective reactions as a function of quantitative perfor-mance feedback (Study 1). Quantitative evaluations reflect social comparison effects; ver-bal evaluations reflect dimensional comparison effects. Higher values indicate morefavorable evaluations. Error bars are ± 1 SEM. Numeric means and standard deviationsare provided in the Supplemental Materials.

63J.E. Strickhouser, E. Zell / Journal of Experimental Social Psychology 59 (2015) 60–66

performance ranked average or below average (all groups were signifi-cantly different, ps b .01, ds N 0.63). However, consistentwith our primaryprediction, the influence of quantitative feedback was significantlystronger on quantitative self-evaluations than on verbal self-evaluations,F(1, 103) = 9.59, p = .003, ηp2 = .09. This finding suggests that socialcomparison had a stronger influence on self-evaluations than dimension-al comparison.

Affective reactions

Parallel results were obtained on affective reactions (see Fig. 1, rightpanel). A 3 (quantitative feedback) by 2 (evaluation dimension) ANOVAwith repeatedmeasures on the second factor yielded a significant quan-titative feedback by evaluation dimension interaction, F(2, 103) =107.08, p b .001, ηp2 = .68. As before, the effect of social comparisonwas significant, F(2, 103) = 54.01, p b .001, ηp2 = .51. Participants toldthat their quantitative performance ranked above average felt betterabout their quantitative performance than participants told that theirquantitative performance ranked average or below average (all groupswere significantly different, ps b .005, ds N 0.66). The effect of dimen-sional comparison was also significant, F(2, 103) = 14.00, p b .001,ηp2 = .21. Participants told that their quantitative performance rankedabove average felt worse about their verbal performance than partici-pants told that their quantitative performance ranked average orbelow average (all groups were significantly different, ps b .02,ds N 0.61). More importantly, however, the influence of quantitativefeedback was significantly stronger on quantitative affective reactionsthan on verbal affective reactions, F(1, 103) = 9.01, p = .003, ηp2 =.08, which suggests that social comparison had a stronger influence onaffective reactions than dimensional comparison.2

Supplemental analyses

Recent research suggests that downward comparisons, regardless ofwhether they are dimensional or social, may have a stronger influenceon self-evaluations and affective reactions than upward comparisons,presumably because downward comparisons are more flattering(Möller & Pohlmann, 2010; Pohlmann & Möller, 2009). We conducteda weighted contrast to test this possibility in the current study usingthe following contrast codes (+1, −2, +1 for upward, lateral, anddownward comparisons, respectively; see Möller & Pohlmann, 2010for a similar analysis). Consistent with prior research, the effect ofdownward (versus lateral) social comparisonwas significantly stronger

2 In both studies, actual performance on the quantitative and verbal reasoning tests didnot significantly differ by feedback condition (ps N .30). All reported results were un-changed when controlling for actual performance (see Supplemental Materials).

on affective reactions than the effect of upward (versus lateral) socialcomparison, F(1, 103) = 5.59, p = .02, ηp2 = .05 (see Fig. 1). However,downward social comparison did not have a significantly stronger effecton self-evaluations than upward social comparison, F(1, 103) = 2.87,p = .09, ηp2 = .03. Further, downward dimensional comparison didnot have a significantly stronger effect on self-evaluations or affectivereactions than upward dimensional comparison, F(1, 103) = 0.11,p= .74, ηp2= .001, and F(1, 103)=0.02, p= .90, ηp2 b .001, respectively.Thus, our results provided inconsistent support for the argument thatdownward comparisons exert stronger effects than upward compari-sons; we elaborate on this point in the General Discussion.

Discussion

Study 1 results complement a previous study by suggesting that di-mensional comparison effects can be obtained even when people alsohave social comparison information (Pohlmann & Möller, 2009, Study3). Unique to the current report, Study 1 results also suggest that socialcomparisonmay be a stronger determinant of both self-evaluations andaffective reactions than dimensional comparison. That is, the effect ofupward versus downward social comparison was significantly strongerthan the effect of upward versus downward dimensional comparison.

Study 2

Our first experiment had two limitations that necessitated additionalresearch. First, we tested social comparison by examining the effect ofperformance feedback on quantitative evaluations and tested dimension-al comparison by examining the effect of performance feedback on verbalevaluations. One could argue that perhaps social comparison yielded astronger effect because quantitative evaluations are more susceptible tothe influence of performance feedback than verbal evaluations. Second,in Study 1, the effect of social comparison was tested before dimensionalcomparison. Perhaps the effect of dimensional comparison was dimin-ished as a function of being assessed after social comparison.

Study 2 utilized a similar methodology as Study 1 but included im-portant modifications to the feedback levels to address these concerns.That is, two feedback conditions examined the effect of upward versusdownward dimensional comparison with regards to verbal perfor-mance in the context of a lateral (i.e., neutral) social comparison. Addi-tionally, two feedback conditions examined the effect of upward versusdownward social comparisonwith regards to verbal performance in thecontext of a lateral (i.e., neutral) dimensional comparison. In sum, socialand dimensional comparison information in Study 2 were based on thesame dependent variables (verbal evaluations) thus eliminating anypossibility of an order effect or confounding of social/dimensionalcomparison with feedback domain.

Methods

Participants and procedure

Study 2 followed an identical procedure to that of Study 1 in all waysexcept those noted below. Participants were 174 introductory psychol-ogy students (139 females, Mage = 19.86, 78 Caucasian, 45 AfricanAmerican). After completing quantitative and verbal reasoning tests,participants received manipulated feedback indicating that their quan-titative performance ranked better than 81% or 21% of 259 test takersat their school and therefore ranked above average or below average.To reinforce the quantitative feedback, participants then completedquantitative self-evaluation and affective reaction items (αs N .86).However, these items were not relevant to our hypotheses and are notdiscussed further.3

3 The effect of quantitative feedback on quantitative evaluations in Study 2 was consis-tent with results obtained in Study 1 (see Supplemental Materials).

2.0

3.0

4.0

5.0

6.0

7.0

8.0

Social Dimensional Social Dimensional

Self-Evaluations Affect

Upward

Downward

Fig. 2. Mean verbal self-evaluations and affective reactions as a function of comparisontype and comparison direction (Study 2). Higher values indicate more favorable evalua-tions. Error bars are ± 1 SEM. Numeric means and standard deviations are provided inSupplemental Materials.

64 J.E. Strickhouser, E. Zell / Journal of Experimental Social Psychology 59 (2015) 60–66

Next, participants receivedmanipulated feedback about their verbalperformance. Participants in the above average quantitative feedbackconditions were told that their verbal performance ranked better than83% or 53% of 259 previous test takers at their school and thereforeranked above average or average. To highlight dimensional comparison,these participantswere told that their verbal performance ranked aboutthe same or worse than their quantitative performance, as appropriatefor their condition. Similarly, participants in the below average quanti-tative feedback conditions received information, indicating that verbalperformance ranked better than 53% or 23% of 259 previous test takersat their school and therefore ranked average or below average. Tohighlight dimensional comparison, these participants were told thattheir verbal performance ranked about the same or worse than theirquantitative performance, as appropriate for their condition. A summa-ry of the feedback conditions is provided in Table 1. After reviewing theverbal feedback, all participants completed items assessing verbalself-evaluations and affective reactions (αs N .84), which served as thekey dependent variables.

Manipulation checks and suspicion probe

Data from 26 participants were excluded for failing one or moreof the manipulation checks and/or the suspicion probe. All reportedresults were unchanged when these participants were retained (seeSupplemental Materials).

Analytic strategy

In two of the feedback conditions, we varied social comparison(upward, downward) while holding dimensional comparison constant(lateral comparison). Conversely, in the other two feedback conditions,we varied dimensional comparison (upward, downward)while holdingsocial comparison constant (lateral comparison; see Table 1). Thus, ourkey analysis examined whether the effect of upward versus downwardsocial comparison on verbal self-evaluations and affective reactionswassignificantly different from the effect of upward versus downwarddimensional comparison.

Results

Self-evaluationsA 2 (comparison type: social, dimensional) by 2 (comparison

direction: upward, downward) ANOVA was conducted on verbal self-evaluations. This analysis yielded a significantmain effect of comparisondirection, F(1, 144) = 171.64, p b .001, ηp2 = .54. That is, participants inthe downward comparison conditions evaluated themselves signifi-cantly more favorably than participants in the upward comparisonconditions. Follow-up analyses examined the effect of comparisondirection within each comparison type.

As expected, participants in the downward social comparisoncondition evaluated themselves significantly more favorably thanparticipants in the upward social comparison condition, F(1, 144) =139.61, p b .001, ηp2 = .49. Similarly, participants in the downward di-mensional comparison condition evaluated themselves significantlymore favorably than participants in the upward dimensional compari-son condition, F(1, 144)=44.61, p b .001, ηp2= .24. However, consistentwith our primary prediction, the effect of social comparison was signif-icantly stronger than dimensional comparison, as reflected by a signifi-cant comparison type by comparison direction interaction, F(1, 144) =13.81, p b .001, ηp2 = .09 (see Fig. 2, left panel).

Affective reactions

Parallel results were obtained on affective reactions. The effect ofcomparison direction (upward versus downward) on verbal affectivereactions was significant for both social and dimensional comparison,

F(1, 144) = 119.56, p b .001, ηp2 = .45 and F(1, 144) = 28.82, p b .001,ηp2 = .17, but the effect of social comparison was significantly strongerthan dimensional comparison as reflected by a significant two-way in-teraction, F(1, 144) = 16.07, p b .001, ηp2 = .10 (see Fig. 2, right panel).

Discussion

Study 2 replicated the key results of Study 1 while addressing itslimitations. That is, social comparison had amore pronounced influencethan dimensional comparison even when both standards were evaluat-ed using the same performance domain (verbal reasoning) and whenpotential order effects were neutralized.

General discussion

Consistent with prior theories arguing that internal and externalframes of reference simultaneously contribute to the self-concept(Marsh, 1986; Möller et al., 2011), our experiments obtained significanteffects of dimensional and social comparison, and both effects were rel-atively large in size. Furthermore, we directly pitted dimensional andsocial comparison standards against each other and found, consistentwith predictions, that social comparison had a stronger influence onself-evaluations and affective reactions than dimensional comparison.As we outline below, our data contribute to both dimensional and socialcomparison theories.

First, although numerous correlation-based studies have providedevidence consistent with dimensional comparison effects (see Mölleret al., 2009), our results are among only a few experimental studiesdemonstrating the causal nature of such effects. That is, our results inde-pendently support and extend prior experimental research on dimen-sional comparison effects, which had up until now been conductedby only a small group of researchers (e.g. Möller & Köller, 2001;Pohlmann&Möller, 2009). Further, our results bolster a previous exper-iment showing that dimensional comparison effects obtain even whenpeople also have unambiguous social comparison information(Pohlmann & Möller, 2009, Study 3). It is surprising that people withthe sameperformance rank (e.g., 53rd percentile) can have dramaticallydifferent self-evaluations and affective reactions as a function of dimen-sional comparison. One might expect people to focus exclusively ontheir social comparative rank during self-assessment, but our resultsshowed that people also consider the dimensional comparison implica-tions of their performance. In this light, our findings support emergingtheories arguing that dimensional comparison is a key determinant ofthe self-concept (Marsh et al., 2015; Möller & Marsh, 2013).

Second, our research contributes to a growing body of scholarshipon the fundamental importance of social comparison to self and identityprocesses (Alicke et al., 2013; Fiske, 2011; Suls & Wheeler, 2000).

65J.E. Strickhouser, E. Zell / Journal of Experimental Social Psychology 59 (2015) 60–66

Whereas past research has shown that social comparison effects contin-ue to occur when people also have objective information (Klein, 1997)or temporal comparison information (Zell & Alicke, 2009), the presentfindings show that social comparison effects continue to occur whenpeople also have dimensional comparison information. Further,complementing emerging research suggesting that social comparisonshave a stronger influence on self-evaluations than temporal compari-sonswith past selves (Van Yperen & Leander, 2014), our results suggestthat social comparisons have a stronger influence on self-evaluationsthan dimensional comparisons. We argue that the social and materialbenefits of social status (Anderson et al., 2012; Boyce et al., 2010;Kraus et al., 2013) make people more attuned to social comparisonsthan dimensional comparisons, but future research is needed to furtherexplore this possibility.

Although our studies provide initial evidence suggesting that socialcomparisons have a stronger influence on self-evaluations than dimen-sional comparisons, future research is needed to explore the generalityof this effect and whether there are situations in which dimensionalcomparison is stronger than social comparison. Social comparison ef-fects are reduced in large groups and when comparison referents aredissimilar to the self (see Garcia, Tor, & Schiff, 2013; Zell & Alicke,2010), and thus it is possible that these types of social comparison areweaker than some dimensional comparisons. Additionally, dimensionalcomparison effects are maximized when people think that there is astrong inverse relation between the two skills being considered, asthey do for math and verbal skills (Marsh et al., 2015; Möller & Marsh,2013). Nonetheless, it remains possible that dimensional comparisoneffects are even largerwhen examining other feedback domains. Finally,as an initial test of dimensional versus social comparison, we selectedfeedback conditions that were most relevant to our hypotheses. Forthe sake of completeness, however, future research should examineother combinations of dimensional and social comparison that werebeyond the scope of the present report (e.g., upward social and upwarddimensional, downward social and downward dimensional).

Another important task for future experiments will be to examinethe relative effects of downward comparisons versus upward compari-sons. Study 1 results were consistent with prior findings (Möller &Pohlmann, 2010; Pohlmann & Möller, 2009) in that downward socialcomparison had a stronger influence on affective reactions than upwardsocial comparison. However, downward social comparison did nothave a stronger influence than upward social comparison on self-evaluations, and downward dimensional comparison did not havea stronger influence on self-evaluations or affective reactions thanupward dimensional comparison. Future research should continueto examine whether andwhen downward comparisons are more im-pactful than upward comparisons. The tendency for people to em-phasize flattering downward comparisons may be more common incircumstances in which the performance domain is self-relevant(Tesser, 1988) or among people who are dispositionally high inself-enhancement (Sedikides & Gregg, 2008).

Conclusions

Scholars have long recognized that dimensional and social compari-sons are primary determinants of the self-concept. The current researchargues that although both standards are important, social comparisonmay have a stronger influence on self-evaluations of performance andability than dimensional comparison. Future study is needed to identifymoderator variables that influence the relative effects of dimensionalversus social comparison. Now that research has begun to resolve thefirst-order question of how comparison standards influence the self inisolation of each other, the time is ripe for researchers to shift tothe second-order question of how comparisons work in combinationto produce self-evaluations and affective reactions as they do ineveryday life.

Appendix A. Supplementary Materials

SupplementaryMaterials to this article can be found online at http://dx.doi.org/10.1016/j.jesp.2015.03.001.

References

Akinola, M., & Mendes, W. B. (2014). It's good to be the king: Neurobiological benefits ofhigher social standing. Social Psychological and Personality Science, 5, 43–51. http://dx.doi.org/10.1177/1948550613485604.

Alicke, M. D., Zell, E., & Guenther, C. L. (2013). Social self-analysis: Constructing,protecting, and enhancing the self. In J. M. Olson, & M. P. Zanna (Eds.),Advances in Experimental Social Psychology, Vol. 48. (pp. 173–234). San Diego,CA: Academic Press.

Anderson, C., Kraus, M. W., Galinsky, A. D., & Keltner, D. (2012). The local-ladder effect:Social status and subjective well-being. Psychological Science, 23, 764–771. http://dx.doi.org/10.1177/0956797611434537.

Boyce, C. J., Brown, G. A., & Moore, S. C. (2010). Money and happiness: Rank of income,not income, affects life satisfaction. Psychological Science, 21, 471–475. http://dx.doi.org/10.1177/0956797610362671.

Chiu, M. (2012). The internal/external frame of reference model, big-fish-little-pondeffect, and combined model for mathematics and science. Journal of EducationalPsychology, 104, 87–107. http://dx.doi.org/10.1037/a0025734.

Ehrlinger, J., & Dunning, D. (2003). How chronic self-views influence (and potentiallymislead) estimates of performance. Journal of Personality and Social Psychology, 84,5–17. http://dx.doi.org/10.1037/0022-3514.84.1.5.

Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7,117–140. http://dx.doi.org/10.1177/001872675400700202.

Fiske, S. T. (2011). Envy up, scorn down: How status divides us. New York, NY: Russell SageFoundation Publications.

Frank, R. H. (1985). Choosing the right pond: Human behavior and the quest for status. NewYork, NY: Oxford University Press.

Garcia, S. M., Tor, A., & Schiff, T. M. (2013). The psychology of competition: A socialcomparison perspective. Perspectives on Psychological Science, 8, 634–650. http://dx.doi.org/10.1177/1745691613504114.

Klein, W. M. (1997). Objective standards are not enough: Affective, self-evaluative, andbehavioral responses to social comparison information. Journal of Personality andSocial Psychology, 72, 763–774. http://dx.doi.org/10.1037/0022-3514.72.4.763.

Kraus, M. W., Adler, N., & Chen, T. (2013). Is the association of subjective SES andself-rated health confounded by negative mood? An experimental approach. HealthPsychology, 32, 138–145. http://dx.doi.org/10.1037/a0027343.

Marsh, H. W. (1986). Verbal and math self-concepts: An internal/external frame ofreference model. American Educational Research Journal, 23, 129–149. http://dx.doi.org/10.2307/1163048.

Marsh, H. W., Lüdtke, O., Nagengast, B., Trautwein, U., Abduljabbar, A. S., Abdelfattah, F.,et al. (2015). Dimensional comparison theory: Paradoxical relations betweenself-beliefs and achievements in multiple domains. Learning and Instruction, 35,16–32. http://dx.doi.org/10.1016/j.learninstruc.2014.08.005.

Möller, J., & Husemann, N. (2006). Internal comparisons in everyday life. Journal ofEducational Psychology, 98, 342–353. http://dx.doi.org/10.1037/0022-0663.98.2.342.

Möller, J., & Köller, O. (2001). Dimensional comparisons: An experimental approach to theinternal/external frame of reference model. Journal of Educational Psychology, 93,826–835. http://dx.doi.org/10.1037/0022-0663.93.4.826.

Möller, J., & Marsh, H. W. (2013). Dimensional comparison theory. Psychological Review,120, 544–560. http://dx.doi.org/10.1037/a0032459.

Möller, J., & Pohlmann, B. (2010). Achievement differences and self-concept differ-ences: Stronger associations for above or below average students? BritishJournal of Educational Psychology, 80, 435–450. http://dx.doi.org/10.1348/000709909X485234.

Möller, J., Pohlmann, B., Köller, O., & Marsh, H. W. (2009). A meta-analytic path analysis ofthe internal/external frame of reference model of academic achievement andacademic self-concept. Review of Educational Research, 79, 1129–1167. http://dx.doi.org/10.3102/0034654309337522.

Möller, J., Retelsdorf, J., Köller, O., & Marsh, H. W. (2011). The reciprocal I/E model: An in-tegration of models of relations between academic achievement and self-concept.American Educational Research Journal, 48, 1315–1346. http://dx.doi.org/10.3102/0002831211419649.

Mussweiler, T. (2003). Comparison processes in social judgment: Mechanisms and conse-quences. Psychological Review, 110, 472–489. http://dx.doi.org/10.1037/0033-295X.110.3.472.

Pohlmann, B., & Möller, J. (2009). On the benefits of dimensional comparisons. Journal ofEducational Psychology, 101, 248–258. http://dx.doi.org/10.1037/a0013151.

Sedikides, C., & Gregg, A. P. (2008). Self-enhancement: Food for thought. Perspectives onPsychological Science, 3, 102–116. http://dx.doi.org/10.1111/j.1745-6916.2008.00068.x.

Suls, J., & Wheeler, L. (2000). Handbook of social comparison: Theory and research.Dordrecht, Netherlands: Kluwer Academic Publishers.

Summerville, A., & Roese, N. J. (2008). Dare to compare: Fact-based versus simulation-based comparison in daily life. Journal of Experimental Social Psychology, 44,664–671. http://dx.doi.org/10.1016/j.jesp.2007.04.002.

Tesser, A. (1988). Toward a self-evaluation maintenance model of social behavior. In L.Berkowitz, & L. Berkowitz (Eds.), Advances in experimental social psychology, Vol. 21.(pp. 181–227). San Diego, CA, US: Academic Press.

66 J.E. Strickhouser, E. Zell / Journal of Experimental Social Psychology 59 (2015) 60–66

Van Yperen, N. W., & Leander, N. P. (2014). The overpowering effect of social comparisoninformation: On the misalignment between mastery-based goals and self-evaluationcriteria. Personality and Social Psychology Bulletin, 40, 676–688. http://dx.doi.org/10.1177/0146167214523475.

Wood, J. V., & Wilson, A. E. (2003). How important is social comparison? In M. R. Leary, & J.Tangney (Eds.),Handbookof self and identity (pp. 344–366).NewYork,NY:Guilford Press.

Zell, E., & Alicke, M. D. (2009). Self-evaluative effects of temporal and social comparison.Journal of Experimental Social Psychology, 45, 223–227. http://dx.doi.org/10.1016/j.jesp.2008.09.007.

Zell, E., & Alicke, M. D. (2010). The local dominance effect in self-evaluation: Evidence andexplanations. Personality and Social Psychology Review, 14, 368–384. http://dx.doi.org/10.1177/1088868310366144.