BRIEF REPORTS

A Bumpy Train Ride: A Field Experiment on Insult, Honor,and Emotional Reactions

Hans IJzerman and Wilco W. van DijkVU University Amsterdam

Marcello GallucciUniversity of Milan–Bicocca

The present research examined the relationship between adherence to honor norms and emotionalreactions after an insult. Participants were 42 Dutch male train travelers, half of whom were insulted bya confederate who bumped into the participant and made a degrading remark. Compared with insultedparticipants with a weak adherence to honor norms, insulted participants with a strong adherence to honornorms were (a) more angry, (b) less joyful, (c) less fearful, and (d) less resigned. Moreover, insultedparticipants with a strong adherence to honor norms perceived more anger in subsequent stimuli thannot-insulted participants with a strong adherence to these norms. The present findings support a directrelationship among insult, adherence to honor norms, and emotional reactions.

Keywords: emotions, honor, insult, field experiment

Several societies in the Mediterranean area of Europe and thesouthern United States have been described as “honor cultures”(Caro Baroja, 1965; D. H. Fischer, 1989; Nisbett & Cohen, 1996;Pitt-Rivers, 1965, 1977; Rodriguez Mosquera, Manstead, &Fischer, 2000). Honor refers to people’s value both in their owneyes and in the eyes of others, and a distinctive feature of honorcultures is the extent to which one’s personal worth is determinedinterpersonally (Miller, 1993; Nisbett & Cohen, 1996; RodriguezMosquera et al., 2000; Stewart, 1994). Whereas an individual’spersonal worth is important in both honor and individualisticcultures, the role of social esteem in determining one’s personalworth is more important in honor cultures than in individualisticcultures.

Research has shown that, compared with individuals from indi-vidualistic cultures, those from honor cultures attach more impor-tance to family-related values and social recognition (A. H.Fischer, Manstead, & Rodriguez Mosquera, 1999). Furthermore, ithas been shown that such values are more important in honorcultures than in other cultures in shaping the experience andexpression of emotions like pride, shame, and anger (Rodriguez

Mosquera et al., 2000). For example, in cultures where honor ismore salient, attacks on one’s honor, as in the case of insults,appear to be a common anger-eliciting event. Moreover, the elic-itation of anger in attacks on one’s honor usually leads to hostilityand retaliation against the perpetrator as a way of restoring one’shonor (Cohen & Nisbett, 1994, 1997; Cohen, Nisbett, Bowdle, &Schwarz, 1996; Nisbett & Cohen, 1996; Stewart, 1994). Thesouthern United States has been regarded as a prime example of anhonor culture in which affronts are met with violent retribution. Ithas been argued that adherence to honor norms might explain theobservation that the American south is more violent than theAmerican north (Cohen, 1996; Nisbett & Cohen, 1996). For in-stance, Southerners are more likely to agree that violence is ac-ceptable in defense of home and family and as a mechanism ofsocial control; consequently, they are more likely to endorse vio-lence as a response to an affront. In honor cultures, “even smalldisputes become contests for reputation and social status” (Cohenet al., 1996, p. 945); thus, individuals are expected to defend theirhonor or reputation, even if this means that they have to fight orkill for it (Nisbett & Cohen, 1996; Pitt-Rivers, 1965).

In an intriguing series of experiments, Cohen et al. (1996) foundsupport for their hypothesis that White males reared in the south-ern United States react differently to an affront than those reared inthe north. Compared with Northerners, who were relatively unaf-fected by an insult, Southerners were (a) more likely to think thattheir masculine reputation was threatened, (b) more upset, (c) morephysiologically and cognitively primed for aggression, and (d)more likely to engage in aggressive and dominant behavior. Cohenet al. suggest that participants’ strength of adherence to honornorms is the underlying mechanism for these effects. That is,Southerners are argued to have a stronger adherence to honornorms than Northerners and, consequently, react stronger andmore aggressively to an affront.

In their research, Cohen et al. (1996) compared the emotionalreactions toward an affront of Southerners with those of North-

Hans IJzerman and Wilco W. van Dijk, Department of Social Psychol-ogy, VU University Amsterdam, The Netherlands; Marcello Gallucci,Department of Psychology, University of Milan–Bicocca, Milan, Italy.

Hans IJzerman and Wilco W. van Dijk contributed equally to theresearch and publication of this study and share first authorship. Thespecific order of these authors’ names was determined in mutual agree-ment. We thank Piotr Winkielman and two anonymous reviewers forvaluable comments on an earlier version of this article. We thank WilliamDjoko, Nicole van Elteren, Bas Kockmann, and Irma Oldenburg for theircontributions to the present study.

Correspondence concerning this article should be addressed to Wilco W.van Dijk, Department of Social Psychology, Vrije Universiteit Amsterdam,van der Boechorststraat 1, 1081 BT, Amsterdam, The Netherlands. E-mail:[email protected]

Emotion Copyright 2007 by the American Psychological Association2007, Vol. 7, No. 4, 869 – 875 1528-3542/07/$12.00 DOI: 10.1037/1528-3542.7.4.869

869

erners in terms of a cultural mechanism based on inner represen-tations such as a script. The script in a cultural of honor, forexample, as described by Cohen and Gelfand (2006), prescribeshow an individual is to act when insulted. Indeed, Cohen et al.demonstrated that the cultural script concerning honor of South-erners and Northerners contains important differences. However,although scripts are essential to understanding cultural mecha-nisms, G. R. Smith and Semin (2004) suggested that individuals donot merely rely on scripts to perform action B when situation Aoccurs. To further comprehend such cultural mechanisms, oneshould examine an individual’s values as to how to cope withsituation A at the moment of occurrence.

The fact that norms in Cohen et al. (1996) were not directlymeasured precludes establishing a direct relation among an insult,adherence to honor norms, and emotional reactions. In the presentresearch, we aimed at providing empirical support for this directrelation. Using a field experiment, we intended to show that, afteran insult and compared with individuals with a weak adherence tohonor norms, individuals with a strong adherence to these normsreact differently. Moreover, these different emotional reactionsshould be moderated by differences in the strength of adherence tohonor norms. These anticipated findings would provide an impor-tant contribution to existing findings on the relation between honorand emotions.

In our research, we modified Cohen et al.’s (1996) experimentalset-up, in which a confederate bumps into an unsuspecting partic-ipant and insults him, in two important ways. First, we measuredparticipants’ strength of adherence to honor norms. Second, weused a field setting to stage the study in more ecologically naturalcircumstances. Moreover, we examined the effect of an insult onboth participants’ immediate emotional reactions and their percep-tion of hostility in subsequent stimuli. We assessed participants’emotional reactions by observing their facial expressions and bodylanguage after an insult and expected that participants with strongadherence to honor norms would be more likely to express emo-tional responses indicating aggression and hostility; that is, theywould appear (a) more angry, (b) less joyful, (c) less fearful, and(d) less resigned. Subsequently, participants’ perception of hostil-ity in subsequent stimuli was assessed via a face-rating task, inwhich they had to indicate which emotion was being expressed ina series of photographs of emotional and neutral faces and howintensely this emotional expression was. This procedure allowedus to examine whether, after an insult, participants with a strongadherence to honor norms will perceive more hostility (a) inneutral stimuli, (b) in all stimuli, or (c) only in stimuli that alsoinvolve an affront or challenge.

Method

Participants and Design

A total of 42 male train travelers (mean age � 30.90 years,SD � 15.80) participated in the study on a voluntary basis.Participants were assigned to the experimental (insult) condition orcontrol (no insult) condition according to a rule that was estab-lished before the experiment commenced: Of all male participantswho entered, every other (odd number) participant was insulted(n � 21), whereas the remaining participants were not insulted.1

Materials and Procedure

The experiment was conducted on a local train on theAmsterdam–Rotterdam track (� 70 km). Confederates served asobservers or insulters and were seated in different parts of onecompartment of the train. The two observers each faced opposingways and switched positions throughout the experiment. Bothobservers could hear everything participants said and could readtheir body language (although from different perspectives). Twomale confederates of similar height and weight served alternativelyas insulters. None of the confederates were fully aware of thepurpose of the study. Participants were train travelers who enteredthe local train at different stops between Amsterdam and Rotter-dam. One of the two observers signaled when a participant wasapproaching. Subsequently, a confederate stood up, bumped intothe participant, and added: “Hey, watch it!” (Hey, kijk eens uitjoh!).2 Immediately after the bumping incident, the observers ratedthe participants’ emotional reactions on a 7-point scale (1 � not atall, 7 � very much). Both observers rated how angry, irritated,joyful, happy, nervous, fearful, and resigned participants appeared.

After participants were seated, they were asked to completeseveral questionnaires. First, they answered several demographicalquestions. Second, participants’ adherence to honor norms wasassessed by asking them to what extent they agreed with nine items(1 � do not agree at all, 7 � totally agree) extracted fromRodriguez Mosquera, Manstead, and Fischer’s (2002) honor ques-tionnaire. These items included questions about family honor,respect, reputation, and humiliation in public (e.g., “My honor ismy reputation”).3 Finally, participants were given a face-ratingtask in which they were asked to indicate which emotion was beingexpressed in a series of four photographs. For each of the fourphotographs, participants were asked to choose one of six options(happiness, sadness, anger, fear, surprise, or disgust) and to indi-cate the intensity of the expression on a 5-point scale (1 � not atall strong, 5 � very strong). Stimuli were black-and-white photo-graphs of emotional and neutral faces taken from the KarolinskaDirected Emotional Faces set (KDEF; Lundqvist, Flykt, & Öhman,1998). These photographs were morphs of facial photographs of 37different males expressing the same emotion: anger (KDEF code:MANS), sadness (MSAS), fear (MAFS), or neutral (MNES).

1 From each condition, 1 additional participant declined to participate inthe experiment. Also, 3 participants indicated that their religious affiliationwas Islam, whereas the majority were Christians or nondenominational(similar to national averages). Combined with confederates’ observationsof participants, we were fully assured that there were not more than 3nonnative Dutch participants. We included these participants in our anal-ysis because the results were equivalent with or without them. None of theparticipants was aware of the relationship between the insult and theexperiment.

2 To prevent the situation from becoming too hostile, confederatesrefrained from using the offensive term “asshole” (klootzak) as used inCohen et al.’s (1996) study. Confederates were also instructed to continuewalking after the incident to minimize contact between him and theparticipant and not to challenge the participant any further.

3 We only chose nine items because we wanted to keep our questionnaireas short and concise as possible (in light of the short duration of partici-pants’ average train ride). Items 2, 7, 9, 11, 14, 17, 20, 24, and 26 wereextracted from Rodriguez Mosquera et al.’s (2002) 27-item honor ques-tionnaire on the basis of their relevance to the present research.

870 BRIEF REPORTS

After all questions were completed, participants were thor-oughly debriefed and the insulter apologized to the individuals heinsulted. Participants assigned to the control condition completedthe same procedure without being insulted (see Appendix A formore details of our experimental set-up).

Results

Adherence to Culture-of-Honor Norms

A principal-component analysis on the nine honor items ex-tracted one factor with an eigenvalue of 5.55 (all other eigenval-ues � 1), explaining 61.6% of the variance.4 Participants’ scoreson the nine items were averaged to obtain a score for theiradherence to honor norms (Cronbach’s � � .92). There was nosignificant difference between conditions on adherence to honornorms, F(1, 31) � 1.00, p � .82, suggesting that self-perception ofemotional reactions in relation to honor norms is unlikely toaccount for the obtained results.

Emotional Reactions

Observers’ ratings of emotional reactions5 were averaged to obtaina measure of emotional reactions.6 Ratings of anger and irritationwere averaged to obtain a measure of anger (� � .92), ratings of joyand happiness were averaged to obtain a measure of joy (� � .86),and ratings of fear and nervousness were averaged to obtain a measureof fear (� � .85). Resignation was assessed with a single rating.7

To estimate the relation between adherence to honor norms andemotional reactions, we first analyzed the entire set of reactionswith a general linear model with the four reactions as dependentvariables, type of emotion as repeated measure factor, adherence tohonor norms as a continuous independent variable, and partici-pants’ age as a covariate variable. Age was included becauseemotional reactions may vary across age, introducing uninterestingvariance. Results showed a significant main effect of type ofemotion, F(3, 36) � 6.50, p � .001, p�2 � .32, which simplyindicates that participants, on average, showed less joy (M � 2.53)and fear (M � 3.15) than anger (M � 4.67) and resignation (M �4.57). More interesting, there was an interaction between type ofemotion and adherence to honor norms, F(3, 36) � 10.53, p �.001, p�2 � .47, indicating that participants with different levelsof adherence to honor norms showed a different pattern of emo-tional reactions (Table 1). Univariate regressions showed thatparticipants with a strong adherence to honor norms appearedsignificantly more angry, t(13) � 3.61, p � .001, B � 0.91, with

Sr � .64; less fearful, t(13) � �2.24, p � .05, B � �0.45, withSr � .54; less joyful, t(13) � �2.84, p � .025, B � �0.62,with Sr � �.54, and less resigned, t(13) � �2.38, p � .01, B ��0.48, with Sr � �.45, after being insulted than participants witha low adherence to honor norms.8 Thus, after an insult and com-pared with participants with a weak adherence to honor norms,participants with a strong adherence to honor norms displayedmore facial expressions and body language indicating aggressionand hostility.

Perceived Hostility

In analyzing participants’ responses to emotional faces, we firstchecked whether accuracy of classification of the faces was dif-ferent for different emotions, across conditions (insult vs. noinsult), and across different levels of adherence to honor norms.Upon establishing no influence of the latter two variables on theaccuracy level, we tested whether the perceived intensity of theemotion displayed by the picture was influenced by condition andadherence to honor norms. For the sake of completeness, we alsotested perceived intensity on correctly identified faces. In theseanalyses, we did not include neutral faces because neither accuracy

4 Despite the small sample size, the analyses complied with all therequirements (only one factor extracted, high communalities; n � 25) that,according to Preacher and MacCallum (2002), guaranteed reliable factorstructures in small samples.

5 All analyses controlled for differences between the two insulters. Therewas no effect of insulter on angriness, t(15) � 1.55, p � .15, nor was therean effect of insulter on resignation, t(15) � 1.00, p � .92. There was aneffect, however, of insulter on fearfulness, t(15) � 4.20, p � .05, B � 0.85.

6 All emotional reactions were obtained after the bumping incident,because no comparable set opportunity existed that could be used tomeasure emotional reactions in the control condition.

7 Correlations for observers’ judgments were .71 for anger, .38 for fear,.38 for resignation (all ps � .05, one-tailed), and �.02 for joy (ns.). Thislack of agreement for joy may be explained by the situation (beinginsulted), in which joy was not appropriated and in fact, showed a lowmean (M � 2.50) and variance (SD � 1.24).

8 One might well be concerned with the physical condition of theparticipant in relation to the insult. However, there was no effect ofparticipant’s height on angriness, t(15) � 1.00, p � .67, resignation,t(15) � 1.00, p � .78, and fearfulness, t(15) � 1.00, p � .35. There was,however, an effect of height on joyfulness, t(15) � 2.26, p � .04, B � 0.12,Sr � .50. Our subsequent analyses controlled for this effect.

Table 1Estimated Means for Emotional Reactions for Low- and High-Honor Participants

Emotional reaction Low honor High honor tsimple tpartial

Anger 3.48 (3.62) 5.99 (5.98) 3.37*** 3.61***

Fear 3.79 (3.78) 2.61 (2.48) 1.89* 2.24**

Joy 3.52 (3.36) 1.84 (1.85) �2.61** �2.83**

Resignation 5.06 (5.27) 3.89 (3.91) �1.59 �2.37**

Note. Least square estimated means for 1 SD above (high honor) and 1 SD below (low honor) the honor scale sample mean. Values in parentheses arethe estimated means for the average age. tsimple � t test and statistical significance for the univariate regression with the corresponding emotional reactionas a dependent variable. tpartial � t test and significance of the same regression with participant age effect held constant.* p � .10. ** p � .05. *** p � .001.

871BRIEF REPORTS

of classification nor perceived intensity can be interpreted for facesshowing no emotion.

Emotional faces were generally identified correctly except forthe fearful face, which was classified as showing surprise by halfof the sample. Accuracy levels for insulted participants were 94%,50%, and 89% for the angry, fearful, and sad faces, respectively,compared with 94%, 35%, and 94%, respectively, for not-insultedparticipants. We analyzed the influence of condition (insult vs. noinsult) and adherence to honor norms on the accuracy (correct vs.incorrect) of classification for angry, fearful, and sad faces using ageneralized linear model (Firth, 1991; Nelder & Wedderburn,1972) and the generalized estimating equations (GEE) method(Liang & Zeger, 1986) to account for correlated responses. Essen-tially, we ran a logistic regression with accuracy as the dependentvariable; adherence to honor norms, condition, and type of emo-tional face as independent variables; and type of emotional face asa repeated measure factor.9 Results showed that accuracy changedacross type of emotional face. In particular, participants were lessaccurate in classifying fearful faces than angry (Z � 3.10, p �.002) and sad (Z � 2.11, p � .03) faces. The analysis yielded nosignificant effect of adherence to honor norms concerning condi-tion and its interaction with type of emotional face.

After establishing that the key variables of our study did notaffect the accuracy of classification of the faces into the correctemotion category, we analyzed the perceived intensity of theemotions displayed in the picture. To study the effects of adher-ence to honor norms on perceived intensity, we used a generallinear model with adherence to honor norms, condition (insult vs.no insult), and their interaction as independent variables; intensityscores as the dependent variable; and type of emotional face(angry, fearful, sad) as the within-participant factor. This modelallows testing the overall effects and possible differential effects ofthe independent variables across different emotional faces. Bothindependent variables were centered before the interaction wascomputed (Aiken & West, 1991). The analysis yielded no statis-tically significant difference between insulted participants andnot-insulted participants, F(1, 27) � 1, p � .33; no overall effectof adherence to honor norms, F(1, 27) � 1, p � .66; and no overallinteraction, F(1, 27) � 3.09, p � .09. However, more importantly,the analysis demonstrated a statistically significant three-way in-teraction among condition, adherence to honor norms, and type ofemotional face, F(2, 54) � 4.61, p � .01, p�2 � .15. This

interaction indicates that the interaction between adherence tohonor norms and condition is different for different emotionalfaces. To probe the three-way interaction, we performed a regres-sion analysis with adherence to honor norms, condition (insult vs.no insult), and their interaction as independent variables for eachtype of emotional face. In regard to the angry face, we found astatistically significant interaction of adherence to honor normsand condition, t(27) � 2.61, p � .014, B � 0.62, with Sr � .43,and no linear effects. Simple slope analysis (Aiken & West, 1991)showed that insulted participants with a strong adherence to honornorms (1 SD above the mean) perceived the angry face as angrierthan not-insulted participants with a strong adherence to honornorms, t(27) � 2.62, p � .021, B � 0.31, with Sr � .43. Nostatistically significant difference was found between insulted andnot-insulted participants with a weak adherence to culture-of-honor norms (1 SD below the mean; Figure 1), t(27) � 1.38, p �.43. Finally, as further support of our hypothesis, there was asignificant positive relation between adherence to honor norms andperception of anger in the angry face after an insult, t(27) � 2.06,p � .020, B � 0.35, with Sr � .57. No statistically significantresult was found in the no-insult condition, t(27) � �1.41, p �.18. In regard to responses to fearful faces, we found no lineareffects and a weak interaction between honor norms and condition,t(27) � �2.04, p � .050, with Sr � .36. Although the pattern ofestimated intensity of the fearful face is suggestive (Table 2), theoverall regression model was not significant, F(3, 27) � 1.58, p �.22, and a subsequent simple slope analysis showed a nonsignifi-cant effect of adherence to honor norm in both the insult conditionand the no-insult condition. Accordingly, we refrain from inter-preting this result. No effect was found of the independent vari-ables on responses to a sad face.

Finally, we repeated the previous analyses of perceived intensityfor each type of emotional face only for participants who perceived

9 We fit the logistic model with SAS PROC GENMOD, defining abinomial distribution for the dependent variable and a logistic link func-tion. The within-participant effect was emotional face (angry, sad, andfearful), which was modeled using a compound-symmetry covariancematrix (cf. Liang & Zeger, 1986). Because of the presence of many zerosin the complete cross-tabulation of the independent and dependent vari-ables, we could not fit the full model. We fit a submodel with all the maineffects and the interaction between emotional face and condition (insult vs.no insult).

Figure 1. Mean perceived intensity of an angry face for 1 SD above (highhonor) and 1 SD below (low honor) the honor scale sample mean at the twolevels of experimental manipulation (insult vs. no insult).

Table 2Perceived Intensity of Faces as Presented and CorrectlyClassified

Low honor High honor

Emotional face No insult Insult No insult Insult

Angry 3.41 (3.42) 3.03 (2.40) 2.41 (3.10) 3.80 (3.60)Sad 2.99 (2.94) 3.55 (3.34) 3.40 (3.20) 3.32 (3.42)Fearful 3.33 (4.12) 2.58 (1.40) 2.17 (2.55) 3.02 (2.38)

Note. Least square estimated means for 1 SD above (high honor) and 1SD below (low honor) the honor scale sample mean at the two levels ofexperimental manipulation (insult vs. no insult). Values in parentheses arethe estimated means only for participants who correctly classified the face.

872 BRIEF REPORTS

the face as displaying the correct emotion.10 In regard to the angryface, a regression with adherence of honor norms, condition (insultvs. no insult), and their interaction as independent variables wasconducted on the 30 participants who perceived the face as dis-playing anger. Coherent with previous analyses, the model yieldeda significant interaction between adherence to honor norms andcondition, t(26) � 2.09, p � .046, B � 0.52, with Sr � .34, andno linear effect. No significant effects were found for perceivedintensity of correctly classified sad faces and fearful faces.

Discussion

The present research investigated the relations between adher-ence to honor norms and emotions. Using a field experimentamong train travelers, we examined the effect of an insult andadherence to honor norms on people’s immediate emotional reac-tions and on their perception of hostility in subsequent stimuli.

Our research is the first to show a direct relation among insult,adherence to honor norms, and immediate emotional reactions,thereby extending previous research in important ways. First, thepresent results show that, after an insult, individuals with a strongadherence to honor norms differed greatly in their emotionalreactions from individuals with a weak adherence to honor norms.The former individuals were rated by observers as angrier but lessfearful, less joyful, and less resigned than the latter. In previousresearch, Cohen et al. (1996) obtained inconsistent findings con-cerning observed emotional reactions of insulted Southerners andNortherners, leading them to comment that “the results of Exper-iment 1 regarding anger and amusement must be treated withcaution until subsequent research replicates the findings of Exper-iment 1 in ecologically natural circumstances” (p. 951). Thepresent findings provide an elegant replication and extension ofCohen et al.’s findings in a real-world context.

On the basis of Cohen et al.’s (1996) theory, it is expected thatindividuals with a strong adherence to honor norms would be lessresigned and less fearful after an insult than those with a weakadherence to honor norms. Our present findings are the first to provideempirical support for this hypothesized relation between adherence tohonor norms and the experience of fear and resignation.

Previous research found significant differences in emotionalreactions after an affront between men reared in the southernversus the northern United States. Cohen et al. (1996) suggestedthat these findings are traceable to differences in the strength ofadherence to honor norms, because the northern and southernUnited States differ significantly in ascribing to these norms.However, their findings were not supported by a direct assessmentof the strength of adherence to honor norms, thereby precludingthe establishment of a direct relation among insult, adherence tohonor norms, and emotional reactions. By incorporating a directmeasure of strength of adherence to honor norms in our fieldexperiment, we were able to establish this direct relation anddemonstrate that stronger adherence to honor norms leads to moreaggressive and hostile responses after an affront.

Furthermore, our present findings are the first to show thatinsulted individuals with a strong adherence to honor norms per-ceive more hostility in subsequent stimuli than not-insulted indi-viduals with a strong adherence to honor norms. The former weremore likely to perceive angry faces as angrier than the latter. Thefinding that this difference in perceived hostility was only found

for angry faces suggests that individuals with a strong adherence tohonor norms do not perceive neutral stimuli as more hostile orperceive more hostility in all stimuli, but they only perceive morehostility in those stimuli that involve an affront or challenge. Thiscan be related to findings of Cohen et al. (1996), who found thatan insult did not create a generalized hostility but rather hostilitytoward specific situations that concerned issues of honor.

Why do people with a strong adherence to honor norms reactmore hostile after they have been insulted? The most promisinganswer to this question may be derived from appraisal theories(Frijda, 1986; Roseman, 1984; Scherer, 1984; C. A. Smith &Ellsworth, 1985). According to the universal contingencies hy-pothesis, people from different cultures will experience similaremotions if they appraise a situation in the same way, whereas theywill experience different emotions if they appraise a situationdifferently (Ellsworth, 1994; Scherer, 1997). This hypothesis im-plies that cultural differences in the emotional experiences may beexplained by differences in appraisals. Cohen et al.’s (1996) find-ings regarding the different emotional reactions of individualsfrom honor and nonhonor cultures may be attributed to differencesin their appraisals concerning the “bumping incident.” The presentfindings contribute to previous research and to the universal con-tingencies hypothesis by suggesting that individuals with a strongadherence to honor norms may perceive the situation as moreunexpected, more unpleasant, more obstructing their goals, ormore unfair. Evaluating the situation more strongly in terms ofthese appraisals is likely to amplify anger and hostility (see, e.g.,Roseman, Wiest, & Swartz, 1994; Scherer, 1997).

Our present findings concerning both immediate emotional re-actions and subsequent perception of hostility may also be ex-plained by an argument made for cultural salience; that is, sometype of events or stimuli may be seen as especially relevant insome cultures but not others (cf. Mesquita & Ellsworth, 2001). Forinstance, certain kinds of events or stimuli may be more salient tomembers of some cultures than others and so are more likely to benoticed, appraised, and reacted to. Frijda and Mesquita (1994)suggested that in cultures in which honor is a strong value peopleare extremely sensitive to events that enhance or diminish theirhonor. They refer to these situations as focal events, or events that“never remain unnoticed” in a culture, and that when they occur“the individual can hardly escape being emotionally affected” (p.71). Our research thus further contributes to the existing literature,because these events “never remain unnoticed” not merely toindividuals from honor cultures but also to those adhering highlyto honor values. In a country where honor is not the salient culturalmechanism, differences we obtained in emotional reactions mayresult from individuals with a strong adherence to honor normsbeing more likely to notice events and stimuli that may signaldisapproval.

Conclusion

The present research demonstrates that adherence to honornorms is an important predictor of emotional reactions after an

10 We did not fit the full model with emotion as a repeated measuresfactor because only 11 participants were accurate for all emotional faces.This very small sample size would drastically reduce the power of themodel.

873BRIEF REPORTS

insult. Individuals with a strong adherence to these norms can reactmore aggressively and hostile after an affront. This could implythat these individuals respond to an insult with violent payback.Moreover, individuals with a strong adherence to honor norms alsoperceive more aggression and hostility in subsequent relevantstimuli. This could imply that these individuals perceive the personwho insults them as more aggressive, which may make an ap-praisal of hostile intent of the affront more probable. Both possibleconsequences may, separately or together, lead to a stronger esca-lation of honor-related issues. Research on honor may provideimportant insights on important societal issues, such as honor-based aggression and revenge.

References

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing andinterpreting interactions. Thousand Oaks, CA: Sage.

Caro Baroja, J. (1965). Honour and shame: A historical account of severalconflicts. In J. G. Peristiany (Ed.), Honour and shame: The values ofMediterranean society (pp. 80 –137). London: Weidenfeld and Nicolson.

Cohen, D. (1996). Law, social policy, and violence: The impact of regionalcultures. Journal of Personality and Social Psychology, 70, 961–978.

Cohen, D., & Gelfand, M. (2006). Events and norms: How events unfoldand what we should do. In C.-Y. Chiu & Y.-Y. Hong (Eds.), Socialpsychology of culture (pp. 161–196). New York: Psychology Press.

Cohen, D., & Nisbett, R. E. (1994). Self-protection and the culture ofhonor: Explaining southern violence. Personality and Social PsychologyBulletin, 20, 551–567.

Cohen, D., & Nisbett, R. E. (1997). Field experiments examining theculture of honor. The role of institutions in perpetuating norms aboutviolence. Personality and Social Psychology Bulletin, 23, 1188 –1199.

Cohen, D., Nisbett, R. E., Bowdle, B. F., & Schwarz, N. (1996). Insult,aggression, and the southern culture of honor: An “experimental eth-nography.” Journal of Personality and Social Psychology, 70, 945–960.

Ellsworth, P. C. (1994). Sense, culture, and sensibility. In S. Kitayama &M. R. Markus (Eds.), Emotion and culture: Empirical studies of mutualinfluence (pp. 23–50). Washington, DC: American Psychological Asso-ciation.

Firth, D. (1991). Generalized linear models. In D. V. Hinckley, N. Reid, &E. J. Snell (Eds.), Statistical theory and modelling (pp. 55– 82). NewYork: Chapman and Hall.

Fischer, A. H., Manstead, A. S. R., & Rodriguez Mosquera, P. M. (1999).The role of honour-related versus individualistic values in conceptual-ising pride, shame and anger: Spanish and Dutch culture prototypes.Cognition and Emotion, 13, 149 –179.

Fischer, D. H. (1989). Albion’s seed: Four British folkways in America.New York: Oxford University Press.

Frijda, N. H. (1986). The emotions. Cambridge, England: CambridgeUniversity Press.

Frijda, N. H., & Mesquita, B. (1994). The social roles and functions ofemotions. In S. Kitayama & M. R. Markus (Eds.), Emotion and culture:Empirical studies of mutual influence (pp. 51– 88). Washington, DC:American Psychological Association.

Liang, K. Y., & Zeger, S. L. (1986). Longitudinal data analysis usinggeneralized linear models. Biometrika, 73, 13–22. Lundqvist, D., Flykt,A., & Öhman, A. (1998). The Karolinska directed emotional faces.Stockholm: Karolinska Institute.

Mesquita, B., & Ellsworth, P. C. (2001). The role of culture in appraisal.In K. R. Scherer, A. Schorr, & T. Johnstone (Eds.), Appraisal processesin emotion (pp. 233–248). Oxford, England: Oxford University Press.

Miller, W. I. (1993). Humiliation and other essays on honor, socialdiscomfort, and violence. Ithaca, NY: Cornell University Press.

Nelder, J. A., & Wedderburn, R. W. M. (1972). Generalized linear models.Journal of the Royal Statistical Society, Series A, 135, 370 –384.

Nisbett, R. E., & Cohen, D. (1996). Culture of honor: The psychology ofviolence in the south. Boulder, CO: Westview.

Pitt-Rivers, J. (1965). Honour and social status. In J. Peristiany (Ed.),Honour and shame: The values of Mediterranean society (pp. 19 –77).London: Weidenfeld & Nicholson.

Pitt-Rivers, J. (1977). The fate of Shechem or the politics of sex: Essays inthe anthropology of the Mediterranean. Cambridge, England: Cam-bridge University Press.

Preacher, K. J., & MacCallum, R. C. (2002). Exploratory factor analysis inbehavior genetics research: Factor recovery with small sample sizes.Behavior Genetics, 32, 153–161.

Rodriguez Mosquera, P. M., Manstead, A. S. R., & Fischer, A. H. (2000).The role of honor-related values in the elicitation, experience, andcommunication of pride, shame, and anger: Spain and the Netherlandscompared. Personality and Social Psychology Bulletin, 26, 833– 844.

Rodriguez Mosquera, P. M., Manstead, A. S. R., & Fischer, A. H. (2002).The role of honour concerns in emotional reactions to offences. Cogni-tion and Emotion, 16, 143–163.

Roseman, I. J. (1984). Cognitive determinants of emotions: A structuraltheory. In P. Shaver (Ed.), Review of personality and social psychology(Vol. 5, pp. 11–36). Beverly Hills, CA: Sage.

Roseman, I. J., Wiest, C., & Swartz, T. S. (1994). Phenomenology, behav-iors, and goals differentiate discrete emotions. Journal of Personalityand Social Psychology, 67, 206 –211.

Scherer, K. R. (1984). Emotion as a multicomponent process: A model andsome cross-cultural data. In P. Shaver (Ed.) Review of personality andsocial psychology: Vol 5. Emotions, relationships and health (pp. 37–63). Beverly Hills, CA: Sage.

Scherer, K. R. (1997). The role of culture in emotion-antecedent appraisal.Journal of Personality and Social Psychology, 73, 902–922.

Smith, C. A., & Ellsworth, P. C. (1985). Patterns of cognitive appraisal inemotion. Journal of Personality and Social Psychology, 48, 813– 838.

Smith, E. R., & Semin, G. R. (2004). Socially situated cognition: Cognitionin its social context. Advances in Experimental Social Psychology, 36,53–117.

Stewart, F. H. (1994). Honor. Chicago: Chicago University Press.

Appendix

Experimental Set-up

The experiment was conducted in a train from Amsterdam toRotterdam. Dutch trains do not allow for people to walk next toeach other; hence, confederates did not factor in whether partici-pants were accompanied or not. The experiment was conductedoutside of peak hours in the Dutch railroads’ “Koploper,” a trainwith a hallway approximately 30 m long (approximately 98 feet).The length of the train permitted the confederates to completesubsequent bumps out of sight of the preceding participants. Care-ful measures were also taken to limit effects of social context onparticipants. For example, participants were also out of sight ofother train travelers who witnessed the bump when completing thequestionnaire.

Confederate instructions. All confederates were thoroughlyinstructed. Hans IJzerman was also in proximity of the confeder-ates, in the unlikely case the situation should escalate. Moreover,he also debriefed participants thoroughly. All confederates wereinformed on how to handle the situation in case any direct conflictmight occur. The following are the different instructions confed-erates received before embarking on the train.

Insulter instructions. “When one of the face raters signals you,you are to stand up, turn around, and bump the participant slightly.

874 BRIEF REPORTS

After the bump, you are to say, ‘Hey, kijk eens uit joh!’ (‘Hey,watch it!’) and keep walking in order not to escalate the situation.After the bump, you will be signaled by the person handing out thequestionnaires to sit down out of sight of the previous participant.After the previous participant completed the questionnaire, you areto apologize to this participant and explain this was part of anexperiment.”

Face-rater instructions. “You will watch the door for partic-ipants. When a participant enters the train and walks toward thedoor, you are to signal the insulter. After the insult, you will watchthe participant’s emotional response and rate this on a scale from1 to 7.”

Instructions for handing questionnaires. “After the participanthas been insulted, wait until he sits down. Then approach him andask him whether he will be willing to complete two questionnaires:a demographic questionnaire related to a marketing study con-ducted by the Vrije Universiteit Amsterdam for the Dutch railroadsand a questionnaire by a fellow Psychology Department student onthe moods of train travelers. Ostensibly, these studies are con-ducted together in order to save you and your friend time.”

Received March 24, 2006Revision received May 29, 2007

Accepted July 13, 2007 �

Correction to Engelmann and Pessoa (2007)

In the article “Motivation Sharpens Exogenous Spatial Attention” by Jan B. Engelmann and LuizPessoa (Emotion, 2007, Vol. 7, No. 3, pp. 668 – 674), the supplemental materials link is as follows:http://dx.doi.org/10.1037/1528-3542.7.3.668.suppDOI: 10.1037/1528-3542.7.4.875

875BRIEF REPORTS