The Role of Honor Concerns in Emotional Reactions to Offenses

COGNITION AND EMOTION, 2002, 16 (1), 143–163 The role of honour concerns in emotional reactions to offences Patricia M. Rodriguez Mosquera, Antony S.R. Manstead, and Agneta H. Fischer University of Amsterdam, The Netherlands We investigated the role of honour concerns in mediating the effect of nationality and gender on the reported intensity of anger and shame in reaction to insult vignettes. Spain, an honour culture, and The Netherlands, where honour is of less central significance, were selected for comparison. A total of 260 (125 Dutch, 135 Spanish) persons participated in the research. Participants completed a measure of honour concerns and answered questions about emotional reactions of anger and shame to vignettes depicting insults in which type of threat was manipulated. It was found that Spanish participants responded especially intensely to insults that threaten family honour, and that this effect of nationality on emotional reactions to threats to family honour was mediated by individual differences in concern for family honour. Both the ethnographic record and social psychological research testify to the fact that offences, such as humiliations and insults, have an especially strong impact in honour cultures (Cohen & Nisbett, 1994; Cohen & Nisbett, 1997; Cohen, Nisbett, Bowdle & Schwarz, 1996; Cohen, Vandello, & Rantilla, 1998; Miller, 1993; Murphy, 1983; Nisbett & Cohen, 1996; Peristiany, 1965; Pitt-Rivers, 1977; Stewart, 1994). Offences are often the object of intense emotional experiences, especially of angry feelings and expressions, in honour cultures. This effect of offences on emotion is related to the importance attached to reputation in such cultures. In particular, the status of one’s honour is strongly based on the maintenance of a good reputation (see e.g., Peristiany, 1965; PittRivers, 1977; Stewart, 1994). This implies a keen sensitivity to social approval Correspondenc e should be addressed to Patricia M. Rodriguez Mosquera, Social Psychology Program, Department of Psychology, University of Amsterdam, Roetersstraat 15, 1018 WB Amsterdam, The Netherlands; e-mail: [email protected] l The review and editorial process for this paper was handled by W. Gerrod Parrott. # 2002 Psychology Press Ltd http://www.tandf.co.uk/journals/pp/02699931.html DOI:10.1080/02699930143000167 144 RODRIGUEZ MOSQUERA, MANSTEAD, FISCHER and disapproval of one’s behavior and personality in honour cultures. Situations in which social respect is withdrawn, as in the case of offences, therefore undermine one’s reputation and threaten one’s claim to honour, thereby provoking intense emotional reactions. Empirical research on the determinants of emotional experiences in reaction to offences in honour cultures has, to date, mainly focused on the characteristics of offence situations that lead to emotion, and particularly to anger, in honour cultures. This research has documented that offences that: (a) take place in public; (b) question the reputation of one’s female relatives in terms of sexual shame; or (c) threaten masculinity lead to intense anger in honour cultures (see e.g., Nisbett & Cohen, 1996; Peristiany, 1965; Stewart, 1994). These types of offence threaten core honour values, such as the importance of protecting one’s reputation in public settings, thereby eliciting intense emotions in honour cultures. In other words, whether or not an offence has implications for the status of one’s honour plays an important role in determining how members of honour cultures react emotionally. Extrapolating from this line of argument, it seems reasonable to assume that the more one is concerned with maintaining honour, the more intense will be the emotional reactions to offences that jeopardise the status of one’s honour. In other words, it can be argued that emotional reactions to honour-threatening offences should be mediated by honour concerns, that is, by the extent to which one is concerned with maintaining honour by conforming to prescriptions of the honour code. Although a psychological concern for maintaining honour is usually assumed to be the factor underlying emotional responses to offences in honour cultures, its role in emotion has mostly been inferred on the basis of the strength of an emotional response (such as anger). The stronger the response, the greater is the presumed concern for honour. Honour concerns have never been directly assessed. The main objective of the present study was to address the role of honour concerns in emotional reactions to offences. We focused on a specific type of offence, namely verbal insults, and on two emotions that are closely related to the loss of honour, namely anger and shame. The research was carried out among young, middle class populations living in two countries which differ with respect to the significance attached to honour: Spain and The Netherlands. Previous cross-cultural studies on social values in these countries have established that honour and honour values (e.g., reputation, social interdependence, family honour) are more important in Spain than in The Netherlands (Fischer, Manstead, & Rodriguez Mosquera, 1999, study 1; Rodriguez Mosquera, 1999; Rodriguez Mosquera, Manstead, & Fischer, in press). This significance of honour and honour values in Spain is in line with ethnographic research on honour in that country (Gilmore, 1987; Gilmore & Gwynne, 1985; Murphy, EMOTION, HONOUR, AND OFFENCES 145 1983; Pitt-Rivers, 1965, 1977).1 In contrast, individualistic values that emphasise independence, autonomy, and the capacities and achievements of the self (e.g., ambition, capability) are more significant in the Netherlands than in Spain.2 There are two emotions that are especially associated with the loss of honour, according to the literature on honour: anger and shame. Anger has been studied in particular in relation to threats to male honour (Cohen & Nisbett, 1994, 1997; Cohen et al., 1996, 1998; Murphy, 1983; Nisbett & Cohen, 1996; Peristiany, 1965; Pitt-Rivers, 1977; Stewart, 1994). Honour cultures promote a view of manhood in which expressions of toughness, strength, and status in public behaviour are even more desirable than is the case in individualistic cultures (Gilmore, 1990). A refusal to submit to public humiliation is therefore a core characteristic of what it means to be a man in honour cultures. In the face of offences, this refusal is expressed in honour cultures by the experience and outward expression of male anger. In contrast to men’s emotional reactions to offences in honour cultures, women’s emotional reactions have been less systematically studied. In the literature on honour and offences women are usually portrayed as the instigators of angry reactions in men in offence situations, and as socialisation agents of such angry reactions in children (see e.g., Nisbett & Cohen, 1996). Previous research on honour and emotions has shown, nevertheless, that offences can also lead to female anger in honour cultures: When asked to recall autobiographical experiences of anger, both female and male Spanish participants reported that being offended by an intimate other was a common elicitor of anger (Rodriguez 1 Two objections can be raised to using ethnographies in social psychological research. First, ethnographies are often carried out in populations quite different from the populations used by social psychologists. Second, ethnographie s are often based on the discourse of informants whose ideologies may not correspond to their psychologies . We agree that care should be taken when deriving predictions for social psychologica l research solely on the basis of the conclusions of ethnographi c records, because the research goals, methodologies, and participant populations involved in the two types of research are often very different. However, in our view, neither objection can reasonably be levelled at the present research. First, our research predictions are based not only on the conclusions of ethnographi c research, but also on our own studies of honour in Spanish and Dutch culture (e.g., Fischer et al., 1999, Study 1), and on research conducted by Cohen and Nisbett and their collaborators (e.g., Cohen et al., 1996). Second, although early ethnographies on honour in the Mediterranean area were mostly carried out in rural communities (e.g., Peristiany, 1965), there has been recently a shift in focus in anthropologica l studies on honour, with research being carried out also in urban areas (e.g., Murphy, 1983; Wikan, 1984). Moreover, cultural anthropologica l research on honour has used a variety of methodologica l approaches besides the use of local informants, such as participant observation and demographi c analysis. 2 Cultural differences between honour and individualistic cultures are seen as relative rather than absolute. Honour and individualistic value orientations are not regarded as opposite poles of a single dimension, but rather as two separate dimensions. 146 RODRIGUEZ MOSQUERA, MANSTEAD, FISCHER Mosquera, Manstead, & Fischer, 2000). There are grounds, therefore, for expecting that offences would also lead to angry feelings on the part of women in honour cultures. Shame has been also associated with the loss of honour in the anthropological literature on honour. Shame has been described as an emotion that follows either disgraceful behaviour by the self or intimate others, or the withdrawal of social respect (Miller, 1993; Peristiany, 1965). Furthermore, ‘‘having a sense of shame’’ seems to be central in honour cultures, and in particular in Mediterranean honour cultures (see e.g., Gilmore, 1987; Peristiany, 1965; Pitt-Rivers, 1977). Shame in this context refers to an inner disposition or personal attribute of someone who has a sense of honour and is concerned with reputation issues. Feeling shame when social respect is withdrawn, as in the case of an offence, signals one’s attachment to the honour code and is moreover consistent with a self-definition in terms of ‘‘having a sense of shame’’. In addition, such an emotional reaction should reinforce one’s identity as a person who cares for his/ her honour. Because reputation is emphasised for both sexes in honour cultures, experiencing shame in reaction to an offence should be elicited in both women and men in these cultures. In sum, both anger and shame have been related to the withdrawal of social respect in the literature on honour. Anger in reaction to an offence implies a focus on others’ negative behaviour, on a lack of respect shown by the offender. Shame in reaction to an offence implies an inward focus, that is, a focus on one’s image being undermined in the eyes of others, which could lead to a lowering of self-esteem. We anticipate that both emotions are experienced more intensely in reaction to offences in honour cultures than in other cultures. Intense anger and shame will not be elicited by any offence in honour cultures, but rather in response to those offences in which honour issues are at stake, such as one’s family honour being jeopardised. In other words, the type of threat involved in an offence situation, the extent to which it is honour threatening, is assumed to be a moderator of the intensity of emotional reactions to offences in honour cultures. Moreover, the more significant honour concerns are for an individual, the more intense one’s anger and/or shame should be in reaction to honour-threatening offences. The core concerns related to the maintenance of honour (i.e., honour concerns), can be derived from the honour code. The honour code involves a set of values and norms that define honourable and dishonorable behaviour, and thereby prescribes those things about which an individual with a sense of honour should be concerned. The honour code stresses the importance of family honour, social interdependence, masculine honour, and feminine honour (Gilmore, 1987, 1990; Gilmore & Gwynne, 1985; Jakubowska, 1989; Nisbett & Cohen, 1996; Peristiany, 1965; Pitt-Rivers, 1977; Stewart, 1994). Family honour refers to values and norms concerning the protection of one’s family’s reputation. Being concerned with family honour implies caring for EMOTION, HONOUR, AND OFFENCES 147 social evaluations of one’s family, for the impact of one’s behaviour on one’s family honour, and for the defence of one’s family’s name. There is thus a strong interdependence between personal and family honour: The family’s collective reputation is a reflection of the reputation of its individual members, while the status of the family’s honour is shared by each family member. This strong interdependence between one’s own and one’s family’s honour means that one is dependent on the actions of one’s intimates, and how they are evaluated. Social interdependence refers to a group of values and norms that focus on the strengthening of social bonds and the maintenance of interpersonal harmony, such as generosity, honesty or hospitality. Social interdependence therefore implies wanting to live up to others’ expectations, to be honest, and to demonstrate reciprocity in one’s relations with others as significant concerns in honour cultures. These concerns can also be understood as a desire to express one’s integrity in social relations. Personal integrity (i.e., being loyal to one’s own values and principles), has also been referred to in the literature as a relevant concern in relation to the maintenance of honour (see e.g., Peristiany, 1965). 3 Being concerned with family honour and with one’s integrity in social relations are conceived of in honour cultures as relevant to the maintenance of both women’s and men’s honour. However, there are also masculine and feminine honour codes that define gender-specific concerns. The masculine honour code entails being concerned with one’s family’s well-being, the maintenance of authority over one’s family, and virility (e.g., being sexually active or being able to father many children). The feminine honour code, on the other hand, centers on sexual shame. Sexual shame entails the expression of restraint in sexual behaviour (e.g., maintaining virginity before marriage), modesty, and a sense of shame in women’s social relations with men, and decorum in dress. Moreover, the behaviour and reputation of female relatives in terms of sexual shame is an important determinant of the status of family honour. This means that a lack of sexual shame on the part of female relatives is an important antecedent for the loss of family honour, and that caring for one’s female relatives’ sexual behaviour becomes a central concern with regard to the maintenance of family honour. We assume that the core social values and the norms of the honour code are internalised to different degrees at the individual level, leading to individual differences in the extent to which an individual is attached to the honour code. 3 There exist different views on the role of personal integrity in the maintenance of honour. Some ethnographer s (e.g., Peristiany, 1965) have referred to it as part of the honour code, whereas other authors (Nisbett & Cohen, 1996; Stewart, 1994) have argued that personal integrity can be seen as a transcultural determinant of honour and prestige, implying that it is less characteristic of the honour code. We considered it important to include personal integrity in the present study in order to have a complete view of honour concerns. 148 RODRIGUEZ MOSQUERA, MANSTEAD, FISCHER Thus there will be individual differences in the extent to which one is concerned with family honour, maintaining integrity in social relations, masculine honour, and feminine honour. We propose that these honour concerns serve as mediators of the impact of honour-threatening offences on the intensity of anger and shame. Finally, our assumption that the type of threat involved in an offence situation influences how people respond emotionally implies that type of threat could be a moderator of cultural differences in emotional reactions to offences. It seems reasonable to assume that types of threat other than those that are honour-related are relevant to the self in cultures in which honour is of less central significance, such as individualistic cultures. In particular, offences in which individualistic concerns are at stake, such as when one’s independence or competence is questioned, should give rise to emotion in individualistic cultures. THE PRESENT STUDY Participants completed two sets of measures. The first was a measure of honour concerns, that is, family honour, integrity in social relations, feminine honour, and masculine honour. The second was a measure of the intensity of anger and shame in reaction to insult vignettes in which type of threat posed by the insult was manipulated. Honour concerns were operationalised in the form of items describing either a behaviour or a reputation that violates the concern in question. This was intended to reflect the fact that the maintenance of honour is dependent both on one’s actions and on one’s maintenance of a reputation that conforms to the honour code (see, e.g., Peristiany, 1965). The items were phrased in such a way as to express a lack of concern for family honour, integrity in social relations, masculine honour, or feminine honour. Further, honour concerns should be a central aspect of one’s identity in honour cultures. This means that self-esteem should be dependent on living up to honour concerns. Participants were therefore asked to rate the extent to which behaving or having a reputation as described in each item would damage their self-esteem, with a higher score expressing greater damage. This measure of honour concerns was constructed in order to: (1) investigate the effect of nationality and gender on honour concerns; and (2) be used as a mediating variable if nationality and gender proved to have effects on the reported intensity of anger and shame in reaction to honour-threatening offences. The second set of measures involved six vignettes in which verbal insults were described. Because type of threat was assumed to be a factor moderating the impact of offences on emotion in honour and individualistic cultures, we varied the type of threat posed by the insults described in the vignettes. Three vignettes posed threats to honour concerns (i.e., threat to family honour; threat to masculine honour; and threat to feminine honour). The other three posed threats to individualistic concerns (i.e., threat to competence; threat to autonomy; and threat to EMOTION, HONOUR, AND OFFENCES 149 assertiveness). Participants were presented with all six vignettes and asked to imagine as vividly as possible the situation described in each vignette happening to them and to report the extent to which it would elicit anger and shame. Our predictions were as follows. First, we expected Spanish participants to be more concerned than Dutch participants with family honour and integrity in social relations, and therefore to score higher on the items associated with these honour concerns. Spanish males should also be more concerned than Spanish females or Dutch participants with masculine honour; likewise, Spanish females should be more concerned than Spanish males or Dutch participants with feminine honour. Second, we expected type of threat in an offence situation to affect the intensity of the emotional reactions of Spanish and Dutch participants: Dutch participants should report more intense anger and shame than Spanish participants in reaction to the three vignettes in which threats to individualism were posed. By contrast, Spanish participants should report more intense anger and shame in reaction to the vignette in which family honour is threatened. Third, Spanish males were expected to report more intense anger and shame in reaction to the vignette in which masculine honour was threatened, by comparison with Spanish females or Dutch participants. Spanish females, on the other hand, were expected to report more intense anger and shame in reaction to the vignette in which feminine honour was at stake, by comparison with Spanish males or Dutch participants. Fourth, we expected individual differences in strength of honour concerns to mediate the predicted effects of nationality and gender on emotional reactions to the honour-threatening vignettes. METHOD Participants A total of 125 Dutch (61 females, 64 males) and 135 Spanish (62 females, 73 males) university students participated in the study. The mean age of the Spanish participants was 22.3 years, and the mean age of the Dutch participants was 21.6 years, the mean age of the whole sample being 21.9 years. Spanish participants were students at the Auto noma University of Madrid, and Dutch participants were students at the University of Amsterdam. All Spanish and Dutch participants were born and had grown up in Spain or the Netherlands, respectively. Spanish and Dutch participants were only included if the nationalities of both parents were Spanish or Dutch, respectively. Questionnaire and procedure Participants completed a two-part questionnaire. One part included the measures of strength of concern for family honour, integrity in social relations, masculine honour, and feminine honour. Each honour concern was represented by an 150 RODRIGUEZ MOSQUERA, MANSTEAD, FISCHER approximately equal number of items. The items are shown in Appendix A. Note that the word ‘‘honour’’ was not used in any of the items. Participants were asked to imagine as vividly as possible that they behaved or had the reputation described in each item, and to indicate the extent to which such a behaviour or reputation would damage their self-esteem by circling a point on a 7-point scale ranging from not at all (0) to very much (6). The items were presented in a random order. Part 2 of the questionnaire consisted of the six insult vignettes and emotion measures in reaction to the vignettes. The content of the vignettes was varied so as to manipulate type of threat represented by the insult. Three vignettes posed threats to honour concerns: being depicted as a disgraceful member of one’s family (threat to family honour); being depicted as unable to protect an intimate other in a threatening situation (threat to masculine honour); and being depicted as having various sexual partners (threat to sexual shame). The other three vignettes represented threats to individualism: being depicted as a failure in an academic context (threat to competence); being depicted as lacking autonomy and independence (threat to autonomy); and being depicted as someone who lacks assertiveness in social relations (threat to assertiveness). The vignettes were derived from real-life experiences as described in a previous interview study with Spanish and Dutch adults. In this interview study participants were asked to report autobiographical situations that had negatively affected their self-esteem. Content analysis of these descriptions showed that Dutch participants more often reported threats to self-esteem in which individualistic concerns were at stake, whereas Spanish participants more often reported self-esteem threats in which honour concerns were at stake. Situations that clearly reflected threats to individualism and threats to honour were selected for the present research. The vignettes described an act in which individualism or honour was made salient, followed by an insult made by someone else that was a criticism of the described act.4 The insult was always initially phrased in the same way: ‘‘If others were then to say to you . . .’’ All vignettes are shown in Appendix B. 4 Before the verbal insult was introduced some measures were taken to assess reactions to these self-esteem threatening situations. Thus we asked participants to report on a 7-point scale from not at all (0) to very much (6) the extent to which they would think about others’ opinions, and the extent to which they would feel shame and anger in reaction to these situations. A MANOVA with nationality, gender, and type of self-esteem threatening situation (individualistic or honour) was performed on these measures. The main effect of gender was significant: Female participants scored higher on shame and anger than did male participants in reaction to the situations. The interaction between gender and type of situation was also significant: Female participants scored higher on a measure of concern with others’ opinion and on shame than did male participants, but only in relation to the honour situations. However, the alphas of the measures of shame and anger within each set of selfesteem threatening situations were lower than .60, which led us to decide not to report these analyses. Cultural differences in emotional reactions to these self-esteem threatening situations were not found. As will be seen later, cultural differences did appear when insults were introduced into the vignettes. EMOTION, HONOUR, AND OFFENCES 151 Participants then answered five questions designed to assess the intensity of anger and shame in reaction to the insults. They were instructed to react specifically to the last part of the vignette (i.e., to the verbal insult). Three questions assessed the extent to which a vignette would elicit angry feelings, namely ‘‘To what extent would you feel enraged?’’ ‘‘To what extent would you feel insulted?’’ and ‘‘To what extent would this [i.e., the verbal insult] hurt your pride?’’5 Two questions assessed the extent to which a vignette would elicit feelings of shame: ‘‘To what extent would you feel shame?’’ and ‘‘To what extent would this [i.e., the verbal insult] damage your self-esteem?’’ These questions were answered on 7-point scales ranging from not at all (0) to very much (6). Order of presentation of vignettes was counterbalanced across participants. Order of presentation of the measure of honour concerns and the vignettes was also counterbalanced across participants. Spanish respondent s participated in this research on a voluntary basis. Dutch respondents were given either course credits or a small financial reward (Dfl5; approximately US$2.50) for participating in the study. All participants completed the questionnaire individually. Questionnaires were prepared in Dutch and then translated into Spanish by a native speaker. The conceptual and linguistic equivalence of the Spanish and Dutch versions of the questionnaire was established by a process of backtranslation performed by an independent linguistic expert fluent in both languages. RESULTS Effect of nationality and gender on honour concerns We first assessed the internal consistency of the four group of items aimed at assessing honour concerns, that is, family honour, integrity in social relations, masculine honour, and feminine honour. These four measures will be referred to henceforth as concern for family honour, concern for integrity, concern for masculine honour, and concern for feminine honour. Cronbach’s alphas were computed for each honour concern for the whole sample, as well as for the Spanish and Dutch sample separately. We therefore computed 12 Cronbach’s alphas. All alphas were higher than .70, with several being higher than .80. Next, 5 Previous research on honour and emotion suggests that a feeling of one’s pride being hurt is also associated with offences in honour cultures (Rodriguez Mosquera et al., 2000). In this research, Spanish and Dutch participants were asked to report autobiographica l experiences of pride (among other emotions). Some of the Spanish participants’ pride stories focused on what we labelled hurt pride. These stories centred on offences committed by intimate others, and this type of emotional reaction involved a negative rather than a positive phenomenologica l experience (e.g., negative thought s about the offender). Pride in this context can be understood as wounded honour or hurt dignity, and therefore as being closely related to feelings of being insulted in an offence situation. 152 RODRIGUEZ MOSQUERA, MANSTEAD, FISCHER scores on items within each honour concern measure were averaged, thereby obtaining an average measure of strength of each honour concern, and were entered into a MANOVA as multiple dependent variables. Nationality and gender served as independent factors. The multivariate main effects of nationality, F(4, 238) = 14.45, p < .001, and of gender, F(4, 238) = 10.62, p < .001, were significant. The means and standard deviations for each honour concern measure as a function of nationality and gender, together with the corresponding univariate F-values are shown in Table 1. The univariate nationality main effect was significant only in the case of family honour: Spanish participants scored higher than Dutch participants. The univariate gender main effect was significant only in the case of feminine honour: Female participants scored higher than male participants. Contrary to expectations, the multivariate interaction between nationality and gender was not significant. Finally, comparing the relative importance of the four honour concerns, it is worth noting that respondents in both countries appear to be least concerned with feminine honour and most concerned with their integrity in social relations. Anger and shame in reaction to the insult vignettes In order not to lose information about the specific threats to honour or threats to individualism at stake in each vignette, the vignettes were analysed separately. We first computed composite scores for anger and shame reactions to each vignette. The anger score was computed by averaging responses to the questions ‘‘To what extent would you feel enraged?’’ ‘‘To what extent would you feel insulted?’’ and ‘‘To what extent would this [the verbal insult] hurt your pride?’’ The shame score was computed by averaging responses to the questions ‘‘To what extent would you feel shame?’’ and ‘‘To what extent would this [the verbal TABLE 1 Effect of nationality and gender on honour concerns Nationality Spanish Concern for: Family honour Integrity Masculine honour Feminine honour Gender Dutch Female Male M (SD) M (SD) M (SD) M 5.35 5.94 3.99 2.67 (1.19) (0.75) (0.95) (1.27) 4.51 5.92 4.18 2.83 (1.08) (0.70) (1.00) (0.95) 4.99 6.01 4.03 3.18 (1.08) (0.69) (0.97) (1.06) 4.92 5.86 4.12 2.35 (SD) F1 (1.33) 31.66*** (0.75) <1.00 (0.99) 1.53 (1.04) 1.37 F2 <1.00 1.58 <1.00 37.78*** Note: Spanish: n = 135; Dutch: n = 125. Female participants: n = 123. Male participants: n = 137. F1 , F2 = Univariate effects correspondin g to the multivariate main effect of nationality and the multivariate main effect of gender, respectively, with 1, 241 degrees of freedom. *** p < .001. 153 EMOTION, HONOUR, AND OFFENCES insult] damage your self-esteem?’’ The internal consistency of these composite anger and shame scores was assessed by computing Cronbach’s alphas for each of the vignettes, both for the sample as a whole and for the Spanish and Dutch sample separately. All of the resulting 36 alphas were higher than .70, with several being higher than .80.6 These anger and shame scores were then entered into a MANOVA, with nationality and gender as factors. The results relating to the threat to family honour, threat to masculine honour, and threat to sexual shame vignettes are shown in Table 2. The results relating to TABLE 2 Threat to family honour, threat to masculine honour, and threat to sexual shame vignette s Nationality Spanish Emotion M (SD) Gender Dutch M (SD) Female M (SD) Male M (SD) F1 F2 Threat to family honour Anger 4.59 (1.71) Shame 4.14 (1.72) 4.49 (1.52) 3.53 (1.57) 5.05 (1.42) 4.35 (1.50) 4.08 (1.65) 3.40 (1.70) <1.00 10.09** 25.73*** 24.38*** Threat to masculine honour Anger 2.93 (1.86) Shame 2.39 (1.63) 3.11 (1.73) 2.25 (1.37) 2.73 (1.79) 2.01 (1.18) 3.27 (1.76) 2.61 (1.71) <1.00 <1.00 5.88* 10.23** Threat to sexual shame Anger 3.39 (2.06) Shame 3.14 (1.90) 3.46 (1.70) 3.26 (1.69) 4.51 (1.67) 4.09 (1.69) 2.45 (1.52) 2.41 (1.50) <1.00 <1.00 105.03*** 71.88*** Note: F1 = Univariate effects corresponding to the multivariate main effect of nationality for each vignette. F2 = Univariate effects corresponding to the multivariate main effect of gender for each vignette. Univariate effects for the threat to family honour and threat to masculine honour vignettes have 1, 256 degrees of freedom; univariate effects for the threat to sexual shame vignette have 1, 255 degrees of freedom. * p < .02; ** p < .01; *** p < .001. 6 On the basis of our earlier (unpublished) interview study on self-esteem threats in Spain and The Netherlands, a further insult vignette was created. This vignette depicted an insult in which the self is portrayed as lacking integrity in social relations (i.e., as being dishonest). This situation was more often described as threatening to self-esteem by Spanish than by Dutch participants in the interview study. However, the reliability of the composite score of shame for this vignette was lower than .60 (regardless of whether the reliability was computed for the whole sample, or for the Spanish and Dutch samples separately). For this reason, the results for this vignette are not fully reported here. Surprisingly, however, Dutch participants reported more anger in reaction to this vignette than did Spanish participants. Further, female participants reported more anger and shame in reaction to this vignette than did male participants. 154 RODRIGUEZ MOSQUERA, MANSTEAD, FISCHER the threat to competence, threat to autonomy, and threat to assertiveness vignettes are shown in Table 3. Threat to family honour vignette. The multivariate main effect of nationality was significant, F(2, 255) = 6.88, p = .001. The univariate effect was only significant for shame (see Table 2). As predicted, Spanish participants reported more shame in response to being portrayed as a disgraceful member of the family than did Dutch participants. The multivariate main effect of gender was also significant, F(2, 255) = 15.12, p < .001. The univariate effect was significant for both anger and shame (see Table 2). Female participants reported more anger and shame in response to this vignette than did male participants. The multivariate interaction was not significant. Threat to masculine honour vignette. The multivariate main effect of gender was significant, F(2, 255) = 5.11, p < .01. The univariate effect was significant for both anger and shame (see Table 2). Male participants reported more anger and shame in reaction to being portrayed as unable to protect an intimate other in a threatening situation than did female participants. Neither the TABLE 3 Threat to competence, threat to autonomy, and threat to assertiveness vignettes Nationality Spanish Emotion M (SD) Gender Dutch M (SD) Female M (SD) Male M (SD) F1 1.36 1.45 Threat to competence Anger 4.59 (1.64) Shame 3.88 (1.84) 4.83 (1.38) 3.69 (1.46) 5.12 (1.40) 4.22 (1.58) 4.33 (1.53) 3.40 (1.66) Threat to autonomy Anger 3.63 (1.70) Shame 3.28 (1.60) 4.53 (1.53) 3.74 (1.53) 4.61 (1.55 3.95 (1.56) 3.57 (1.64) 19.42*** 3.10 (1.50) 5.11** Threat to assertiveness Anger 3.89 (1.73) Shame 3.25 (1.61) 3.90 (1.46) 3.60 (1.45) 4.10 (1.55) 3.60 (1.46) 3.71 (1.63) <1.00 3.25 (1.60) 3.14* F2 17.71*** 15.86*** 27.25*** 19.31*** 3.80 3.24 Note: F1 = Univariate effects corresponding to the multivariate main effect of nationality for each vignette. F2 = Univariate effects corresponding to the multivariate main effect of gender for each vignette. Univariate effects for the threat to competence vignette have 1, 255 degrees of freedom, for the threat to autonomy vignette have 1, 256 degrees of freedom, and for the threat to assertiveness vignette have 1, 254 degrees of freedom. * p < .08; ** p < .03; *** p < .001. EMOTION, HONOUR, AND OFFENCES 155 multivariate main effect of nationality nor the multivariate interaction between nationality and gender was significant. Threat to sexual shame vignette. The multivariate main effect of gender was significant, F(2, 254) = 52.48, p < .001. The univariate effect was significant for both emotions (see Table 2). Female participants reported more anger and shame in response to being portrayed as being ‘‘easy to get off with’’ than did male participants. Neither the multivariate main effect of nationality nor the multivariate interaction between nationality and gender was significant. Threat to competence vignette. The multivariate main effect of gender was significant, F(2, 254) = 10.48, p < .001. The univariate effect was significant for both emotions (see Table 3). Female participants reported more anger and shame in reaction to this vignette than did male participants. Although the multivariate main effect of nationality was significant, F(2, 254) = 3.48, p < .04, none of the univariate effects was significant at the .05 level (see Table 3). However, both main effects were qualified by a significant multivariate interaction between nationality and gender, F(2, 254) = 3.23, p < .05. The univariate effect was significant for both anger, F(1, 255) = 4.57, p < .04, and shame, F(1, 255) = 5.69, p < .02. With regard to anger, an analysis of simple main effects revealed that Dutch male participants (M = 4.65, SD = 1.40) reported more anger when their competence was called into question than did their Spanish counterparts (M = 4.05, SD = 1.60), F(1, 257) = 6.15, p < .02. No significant differences were found between Spanish and Dutch female participants’ reported intensity of anger. Unexpectedly, and in relation to shame, an analysis of simple main effects indicated that Spanish female participants (M = 4.57, SD = 1.64) reported more shame in reaction to this vignette than did their Dutch counterparts (M = 3.85, SD = 1.45), F(1, 256) = 5.92, p < .02. Threat to autonomy vignette. The multivariate main effect of nationality was significant, F(2, 255) = 10.03, p < .001.The univariate effect was significant for both anger and shame (see Table 3). Dutch participants reported more anger and shame in reaction to being portrayed as a person who is not autonomous than did Spanish participants. The multivariate main effect of gender was also significant, F(2, 255) = 14.38, p < .001. The univariate effect was significant for both anger and shame (see Table 3). As with the other vignettes, female participants reported more anger and shame in reaction to this vignette than did male participants. Both main effects were qualified by a significant multivariate interaction between nationality and gender, F(2, 255) = 3.66, p < .03. The univariate effect was only significant for anger, F(1, 256) = 6.97, p < .01. Analysis of simple main effects revealed that Dutch male participants (M = 4.29, SD = 1.53) reported more anger in reaction to this vignette than did Spanish male participants (M = 2.95, SD = 1.49), F(1, 257) = 26.05, p < .001. There was 156 RODRIGUEZ MOSQUERA, MANSTEAD, FISCHER no significant difference between the anger scores of Spanish and Dutch female participants. Threat to assertiveness vignette. The multivariate main effect of nationality was marginally significant, F(2, 253) = 2.95, p < .06. The univariate effect was marginally significant for shame (see Table 3). Dutch participants reported more shame in reaction to being portrayed as not being assertive in social relations than did Spanish participants. Neither the multivariate main effect of gender nor the multivariate interaction between nationality and gender was significant. In summary, Spanish respondents reacted more intensely when their family honour was threatened, but (contrary to our expectations) there was no difference between Spanish and Dutch respondents’ reactions when masculine or feminine honour was threatened. As predicted, Dutch respondents reacted more strongly than did Spanish respondents to insults concerning their autonomy and assertiveness. Furthermore, Dutch male participants reacted more strongly than did their Spanish counterparts to insults concerning their competence and autonomy. On the whole, women reported more intense anger and shame in response to all threats, except the threat to masculine honour and the threat to assertiveness, suggesting that women are generally more emotionally reactive to social criticism than males are. Honour concerns as potential mediators In order to investigate the role of honour concerns in mediating the link between nationality, gender, and emotional reactions to the vignettes that depicted threats to honour, mediational analyses were performed following the procedure proposed by Baron and Kenny (1986). It has been already established that the independent variables (nationality and gender) had a significant impact on both the presumed mediator (honour concerns) and the outcome variable (emotional responses to the honour-threatening vignettes). Because nationality and gender differences were only found for concern for family honour and concern for feminine honour, respectively, further analyses are reported only in relation to these two honour concerns, and in relation to the corresponding threat to family honour and the threat to sexual shame vignettes. The next step in mediational analysis involves showing that there is a relation between the mediator and the dependent variables. Concern for family honour correlated significantly with the reported intensity of shame (r = .37; p < .001) in reaction to the threat to family honour vignette. Further, concern for feminine honour correlated significantly with the reported intensity of anger (r = .55; p < .001), and shame (r = .62; p < .001) in reaction to the threat to sexual shame vignette. Then a MANCOVA was performed using nationality as the factor, concern for family honour as the covariate, and the reported intensity of shame in EMOTION, HONOUR, AND OFFENCES 157 reaction to the threat to family honour vignette as the dependent variable. The covariate accounted for a significant amount of variance in the reported intensity of shame, F(1, 256) = 32.75, p < .001. The main effect of nationality was no longer significant after controlling for the covariate, F(1, 256) < 1.00, n.s. A further MANCOVA was performed using gender as the factor, concern for feminine honour as the covariate, and the reported intensity of anger and shame in reaction to the threat to sexual shame vignette as dependent variables. The covariate accounted for a significant amount of variance in the multivariate analysis, F(2, 249) = 53.68, p < .001. Univariate analyses showed that the covariate was significant for the reported intensity of both anger, F(1, 250) = 62.92, p < .001; and shame, F(1, 250) = 107.41, p < .001. However, the multivariate main effect of gender remained significant after controlling for the covariate, F(2, 249) = 30.25, p < .001. The effect of gender also remained significant in univariate terms for both anger, F(1, 250) = 60.70, p < .001; and shame, F(1, 250) = 30.07, p < .001. In sum, these results show that concern for family honour mediated the influence of nationality on the reported intensity of shame in response to insults that threaten one’s family honour. However, concern for feminine honour only partly mediated the impact of gender on the reported intensity of anger and shame in reaction to insults that threaten one’s sexual shame. DISCUSSION The main objective of the present study was to examine the role played by different honour concerns in affecting the intensity of emotional reactions to offences. Our predictions with regard to the effect of nationality and gender on honour concerns were partly confirmed. As expected, Spanish participants rated injuries to family honour, such as being unable to defend one’s family reputation or allowing others to insult one’s family, as more damaging to their self-esteem than did Dutch participants. This reflects a greater concern for family honour on the part of Spanish participants. These results also imply that self-esteem is to a greater extent interpersonally determined in Spain than in The Netherlands, and that it is especially dependent on the reputation and actions of those with whom one shares honour (i.e., one’s relatives). However, and in contrast with what would be expected on the basis of cultural anthropological studies (e.g., Gilmore, 1987; Gilmore & Gwynne, 1985; Murphy, 1983; Pitt-Rivers, 1965, 1977), no cultural differences or interactions between nationality and gender were found for the other honour concerns. There was a significant impact of gender on honour concerns, albeit only in the case of feminine honour concerns. Lacking sexual shame was rated as more damaging for self-esteem by both Spanish and Dutch female participants than by their male counterparts. It is somewhat surprising to find that the impact of gender on these concerns was not moderated by nationality. Moreover, two supposedly core 158 RODRIGUEZ MOSQUERA, MANSTEAD, FISCHER values in honour cultures, namely, the masculine and feminine honour codes, were both rated as being of relatively minor concern in relation to self-esteem. This may reflect the fact that more egalitarian attitudes towards men and women have become more commonplace in Spain in recent decades. The use of university students as participants makes it more likely that comparatively egalitarian attitudes were held by our Spanish respondents. The absence of significant cross-cultural differences in masculine and feminine honour concerns may also help to explain why transgressions of masculine and feminine honour codes did not evoke the expected differences in emotional reactions from Spanish and Dutch respondents. We will return below to this issue. Turning now to the expected role of type of threat in moderating Spanish and Dutch emotional reactions to offences, here too our expectations were only partly confirmed. As predicted, Dutch participants reported more intense anger and/or shame in response to insults portraying them as lacking autonomy, and as not being assertive in social relations, by comparison with Spanish participants. Furthermore, a significant nationality by gender interaction was found for two of the vignettes that threatened individualistic concerns: Dutch males reported more intense anger than did their Spanish counterparts in response to insults that portrayed them as not being competent in an academic context, or as lacking in autonomy. These results support our prediction that members of honour cultures would not react with intense emotion to all types of offence, and our argument that the relevance for the self of the type of threat posed by an offence is central to the understanding of cultural differences in emotional reactions to offences. Both men and women in individualistic cultures appear to be sensitive to offences that question their assertiveness in social relations, as compared to men and women in honour cultures. Furthermore, men in individualistic cultures, when compared to men in honour cultures, seem to be especially sensitive to insults that question their competence and autonomy. An unexpected finding was that Spanish female participants reported more intense shame than did their Dutch counterparts when they were portrayed by an insult as incompetent. The types of situations that Spanish female participants reported as threatening to their self-esteem in our earlier interview study may provide an explanation for this unexpected result. Spanish women who participated in the interview study were final year undergraduates who had already started working in or outside the university. These Spanish women referred quite often to situations in which their capacity to perform academic work successfully was called into question. Moreover, they believed that such expressions of doubt about their competence were often founded on their gender. This could explain why the Spanish female participants in the present study reacted more strongly to the threat to competence vignette. Our predictions were confirmed in relation to the threat to family honour vignette: Spanish participants reported more intense shame in response to this vignette than did Dutch participants. In this vignette, the self is accused of EMOTION, HONOUR, AND OFFENCES 159 bringing shame on his/her family, which could potentially damage family honour. These results are consistent with the fact that family honour was seen as a more important concern for one’s self-esteem in Spain as compared to the Netherlands, and further demonstrate the significance of family honour in Spanish culture.7 In relation to threats to masculine honour, the predicted nationality by gender interaction was not found: Spanish and Dutch male participants reported equivalent levels of anger and shame in response to being described as someone who is incapable of protecting an intimate other. These results are consistent with what was found in relation to concern for masculine honour. Taken at face value these results suggest that protecting one’s family and property, having authority over one’s family, and virility are equally important concerns for men in both honour and individualistic cultures. However, this absence of cultural difference may be partly due to the fact that the participants were students and therefore less concerned with family-related issues than an older sample would be. Differences between honour and individualistic cultures with respect to masculine honour concerns and emotional reactions to insults that threaten male honour might be more apparent in an older group of men who have greater family responsibilities. It may also be that our paper-and-pencil measures could not capture cultural differences in this domain, and that behavioural measures would have revealed cultural differences in masculine honour, although the plausibility of this explanation is somewhat weakened by the fact that we did find nationality effects on reactions to other vignettes. These explanations could be explored in future research. In the case of the threat to sexual shame vignette the results were consistent with those found in relation to concern for feminine honour. Emotional reactions to an insult that threatens one’s sexual shame were influenced by gender, but not by nationality. Thus, female participants reported more intense anger and shame in reaction to being described as being ‘‘easy to get off with’’ than did their male counterparts. These results are in line with previous cross-cultural research on social values in these countries, in which it was found that the cultural importance attached to feminine honour values varied as a function of gender but not as a function of nationality (Rodriguez Mosquera, 1999). 7 It could be argued that what we actually measured in the present research is a concern for familism rather than a concern for family honour . We believe, however, that familism and family honour are two different, albeit related, constructs that imply different types of psychologica l concerns. Familism has been defined as an orientation towards the welfare of one’s immediate and extended family (Gaines et al., 1997). This value orientation is measured by items that express one’s strong emotional ties to and identification with one’s family members, such as ‘‘I cherish the times that I spend with my relatives’’, and ‘‘I cannot imagine what I would do without my family’’. Although a concern for family honour also implies emotional ties to and identification with one’s family members, the primary focus of family honour is on the protection of the family’s reputation. This is conceptuall y distinct from the fundamentally affective construct of familism. 160 RODRIGUEZ MOSQUERA, MANSTEAD, FISCHER Although feminine honour is an honour domain that appears to be more significant for women than for men in both Spain and The Netherlands, other aspects of our results suggest that these cultures do differ with respect to the family implications of not living up to feminine honour concerns. This conclusion is based on the results relating to concern for family honour. The measure of this domain of concern included the item: ‘‘One’s sister or mother having the reputation of having diverse sexual relations.’’8 Spanish participants rated injuries to their family honour, based partly on such lack of sexual shame on the part of female relatives, as more damaging to their self-esteem than Dutch participants did. This suggests that the differences between Spain and The Netherlands with regard to feminine honour focus on the social implications of violating this honour code: In Spain these implications extend beyond the individual, and negatively affect both the individual honour of relatives and the collective honour of the family.9 This greater impact of female relatives’ sexual shame on self-esteem and family honour in Spain is consistent with ethnographic research conducted in Mediterranean honour cultures (Gilmore, 1987; Gilmore & Gwynne, 1985; Murphy, 1983; Peristiany, 1965; Pitt-Rivers, 1977). Our fourth and final hypothesis concerned the mediating role of honour concerns: We expected individual differences in strength of honour concerns to mediate the effects of nationality and gender on emotional reactions to the honour-threatening vignettes. There was good support for this prediction in relation to concern for family honour. Concern for family honour was positively related to the reported intensity of emotion: The more the self was concerned with family honour, the more intense was the shame in the face of insults threatening family honour. Furthermore, the effect of nationality on the reported intensity of emotional response to the vignette depicting a threat to family honour was no longer significant when controlling for individual differences in concern for family honour. The prediction that concern for feminine honour would mediate the impact of gender on emotional reactions to the threat to sexual shame vignette was not confirmed. Concern for feminine honour was significantly and positively correlated with the reported intensity of emotional response to this vignette, showing that the more the self is psychologically concerned with feminine honour, the more intense are one’s anger and shame in reaction to offences that 8 This item was included in this honour concern domain on theoretical grounds, and concern for family honour had good levels of internal consistency. Moreover, the inclusion of this item in this honour concern domain was also supported by empirical results: In an exploratory principal components analysis this item loaded on the same factor as the other family honour items. 9 A comparison between Spanish and Dutch participants’ scores on the item ‘‘One’s sister or mother having the reputation of having diverse sexual relations’’ further supported this conclusion: Spanish participants (M = 4.79, SD = 2.00) rated the lack of sexual shame of female relatives as more damaging for their self-esteem than did Dutch participants (M = 3.89, SD = 1.79), t(257) = 3.78, p < .001. EMOTION, HONOUR, AND OFFENCES 161 threaten sexual shame. However, the impact of gender on intensity of emotional reactions remained significant after controlling for individual differences in concern for feminine honour. Thus, concern for feminine honour only partly mediates the impact of gender on the reported intensity of anger and shame in reaction to the vignette in question. It nevertheless seems most unlikely that gender has a direct effect on intensity of emotional reactions to threats to sexual shame, without any psychological mediation. Future research should try to identify the nature of these psychological mediators. In conclusion, we regard the present study as advancing our understanding of the relations among emotion, honour, and offences. First, we established that violations of family honour concerns have a greater impact on self-esteem in honour cultures than do violations of other honour concerns. Moreover, offences that threaten one’s family honour seem to lead to more intense shame in honour cultures, as compared with offences in which other type of honour threats are involved. Second, it was shown that the type of threat posed by an offence is relevant to the understanding of cultural differences in emotional reactions to offences. Third, we found evidence that concern for family honour mediates the intensity of shame responses to situations in which family honour is threatened. Finally, there is suggestive evidence that the differences between honour and individualistic cultures in relation to feminine honour reside in the implications of lack of sexual shame for intimate others, rather than oneself. Manuscript received 25 September 1999 Revised manuscript received 28 July 2000 REFERENCES Baron, R.M., & Kenny, D.A. (1986). The moderator-mediato r variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Cohen, D., & Nisbett, R.E. (1994). Self-protection and the culture of honour: Explaining southern violence. Personality and Social Psychology Bulletin, 20, 551–567. Cohen, D., & Nisbett, R.E. (1997). Field experiments examining the culture of honor: The role of institutions in perpetuating norms about violence. 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 ethnography’’. Journal of Personality and Social Psychology, 70, 945–960. Cohen, D., Vandello, J., & Rantilla, A.K. (1998). The sacred and the social: Cultures of honor and violence. In P. Gilbert & B. Andrews (Eds.), Shame: Interpersona l behavior, psychopatholog y and culture (pp. 261–282). Oxford, UK: Oxford University Press. Fischer, A.H., Manstead, A.S.R., & Rodriguez Mosquera, P.M. (1999). The role of honor-related versus individualistic values in conceptualizing pride, shame and anger: Spanish and Dutch cultural prototypes. Cognition and Emotion, 13, 149–179. Gaines, S.O., Marelich, W.D., Bledsoe, K.L., Steers, W.N., Henderson, M.C., Granrose, C.S., Bara jas, L., Hicks, D., Lyde, M., Takahashi, Y., Yum, N., Rõ  os, D.I., Garcõ  a, B.F., Farris, K.R., & 162 RODRIGUEZ MOSQUERA, MANSTEAD, FISCHER Page M.S. (1997). Links between race/ethnicity and cultural values as mediated by racial/ethnic identity and moderated by gender. Journal of Personality and Social Psychology, 72, 1460–1476. Gilmore, D.D. (1987). Honor and shame and the unity of the Mediterranean. Washington, DC: American Anthropologica l Association. Gilmore, D. D. (1990). Manhood in the making: Cultural concepts of masculinity. New Haven, CT: Yale University Press. Gilmore, D.D., & Gwynne, G. (1985) (Eds.). Sex and gender in Southern Europe. [Special Issue]. Anthropology, 9. Jakubowska , L. (1989). A matter of honor. The world and I, April, 670–677. Miller, W.I. (1993). Humiliation and other essays on honour, social discomfort, and violence. Ithaca, NY: Cornell University Press. Murphy, M. (1983). Emotional confrontation s between Sevillano fathers and sons. American Ethnologist, 10, 650–664. Nisbett, R.E., & Cohen, D. (1996). Culture of honour: The psycholog y of violence in the South. Boulder, CO: Westview. Peristiany, J.G. (Ed.) (1965). Honour and shame: The values of Mediterranean society. London: Weidenfeld & Nicolson. Pitt-Rivers, J. (1965). Honor and social status. In J.G. Peristiany (Ed.), Honour and shame: The values of Mediterranean society (pp. 18–77). London: Weidenfeld & Nicolson. Pitt-Rivers, J. (1977). The fate of Shechem or the politics of sex: Essays in the anthropolog y of the Mediterranean. Cambridge, UK: Cambridge University Press. Rodriguez Mosquera, P.M. (1999). Honor and emotion: The cultural shaping of pride, shame and anger. Doctoral dissertation, University of Amsterdam. Rodriguez Mosquera, P.M., Manstead, A.S.R., & Fischer, A.H. (2000). The role of honour-relate d values in the elicitation, experience and communication of pride, shame and anger: Spain and the Netherlands compared. Personality and Social Psychology Bulletin, 26, 833–844. Rodriguez Mosquera, P.M., Manstead, A.S.R., & Fischer, A.H. (in press). Honor in Mediterranean and Northern Europe. Journal of Cross-Cultural Psychology. Stewart, F.H. (1994). Honor. Chicago, IL: Chicago University Press. Wikan, U. (1984). Shame and honour: A contestable pair. Man, 19, 635–652. APPENDIX A Items for each honour concern Concern for family honour ‘‘One’s family having a bad reputation’’, ‘‘Self damaging one’s family’s reputation’’, ‘‘Being unable to defend one’s family’s reputation’’, ‘‘One’s sister or mother having the reputation of having diverse sexual relations’’, and ‘‘Letting others insult your family’’. Concern for integrity ‘‘Betraying other people’’, ‘‘Not keeping up one’s word’’, ‘‘Lying to others’’, ‘‘Not being loyal to one’s values and principles’’, ‘‘Having the reputation of being dishonest with others’’, ‘‘Having the reputation of being someone who is not to be trusted’’, and ‘‘Being hypocritical’’. Concern for masculine honour ‘‘Not defending oneself when others insult you’’, ‘‘Not having authority over one’s family’’, ‘‘Being unable to maintain one’s family’’, ‘‘Not yet having had a sexual relationship’’, ‘‘Being incapable of EMOTION, HONOUR, AND OFFENCES 163 having children’’, ‘‘Being known as someone who does not have authority over family’’, ‘‘Being known as someone who cannot support a family’’, ‘‘Having the reputation of being someone without sexual experience’’, and ‘‘Everybody knowing that you are sterile’’. Concern for feminine honour ‘‘Having sexual relations before marriage’’, ‘‘Changing partner often’’, ‘‘Sleeping with someone without starting a serious relationship with that person’’, ‘‘Wearing provocative clothes’’, ‘‘Being known as having different sexual contacts’’, and ‘‘Being known as someone with whom it is easy to sleep with’’. APPENDIX B Insult vignettes Threat to family honour vignette. ‘‘You feel rejected by your own family. One of your uncles often makes negative comments about you, such as: ‘‘You bring shame on the family.’’ If others were then to say to you: ‘‘Even your own family is ashamed of you,’’ to what extent would you . . . ?’’ Threat to masculine honour vignette. ‘‘You have a partner and you are with this person in a cafe . Another person you do not know begins to annoy your partner. Your partner reacts quickly and before you can do anything the other person leaves. If others were then to say to you: ‘‘You are not even capable of protecting your own partner’’, to what extent would you . . . ?’’ Threat to sexual shame vignette. ‘‘You do not have a stable partner and you have a variety of sexual relations. If others were then to say to you: ‘‘It is easy to get off with you’’, to what extent would you . . . ?’’ Threat to competence vignette. ‘‘At the beginning of the trimester you promise yourself to study hard and to prepare for all your courses well, but at the end of the trimester you have low grades. If others were then to say to you: ‘‘You are worthless’’, to what extent would you . . . ?’’ Threat to autonomy vignette. ‘‘When you are with your parents they treat you as a little child who does not have an opinion or life of his or her own. They do things for you that you yourself can do because they think that you are not able to take any decisions. If others were then to say to you: ‘‘You are not very independent ’’, to what extent would you . . . ?’’ Threat to assertiveness vignette. ‘‘You share an apartment with someone. You have agreed to buy some things together, such as toilet paper, cleaning products, etc. The truth is that you are the one who almost always buys these things and your flatmate almost never does so. You believe that this person should buy more often, but you do not dare to say anything. If others were then to say to you: ‘‘You are very weak, you are not even able to stand up for yourself’’, to what extent would you . . . ?’’