Over recent years a considerable amount of active research has been carried out in the field of emotions in contexts of intergroup conflict and political violence (Branscombe & Doosje, Reference Branscombe and Doosje2004; Mackie, Silver, & Smith, Reference Mackie, Silver, Smith, Tiedens and Leach2004), as well as in relation to intergroup forgiveness (Cehajic, Brown, & Castano, Reference Cehajic, Brown and Castano2008; Hewstone et al., Reference Hewstone, Cairns, Voci, McLernon, Niens, Noor, Branscombe and Doosje2004; Tam et al., Reference Tam, Hewstone, Cairns, Tauch, Maio and Kenworthy2007). These studies provide an interesting framework which may shed light on the question of how to resolve intergroup conflicts which are both entrenched and threatened by political violence. Nevertheless, hardly any studies have focused on gender differences in relation to these questions in the intergroup context; thus, little is known, for instance, about gender differences in relation to empathy in the intergroup context (Cundiff & Komarraju, Reference Cundiff and Komarraju2008) or about gender differences in relation to fear of terrorism (Nellis, Reference Nellis2009).
The analysis of gender differences in relation to aspects such as empathy and forgiveness in contexts of intergroup conflict and political violence is extremely important. Yet, without doubt, it is an underestimated and understudied topic. Throughout this study we will analyze gender differences in relation to emotions and attitudes which may help improve intergroup relations and overcome political violence; we shall also highlight why we should care about these differences.
Specifically, this study analyzes gender differences in emotions, forgiveness and tolerance in a context of political violence and intergroup conflict (namely Basque Country, España). It also analyzes said differences in justification of violence in general and attitude towards the terrorist group ETA (Euskadi ta Askatasuna, which means Basque Country and Freedom in the Basque language).
Historical context
Before reviewing the existing literature on the issues to be studied, we will provide a brief overview of the political context in which the research was conducted. First of all, it is important to highlight the fact that the context has changed drastically over the last few years, following ETA’s declaration of a definitive cessation of its armed activities in October 2011.
In Basque society there are two issues which should be highlighted in relation to past intergroup conflict and violence. Firstly, the confrontation between different senses of identity (Basque and Spanish) should be taken into account, as well as the confrontation existing between Basque nationalists (who want Basque Country to gain a greater degree of independence from Spain) and non-nationalists or constitutionalists (who support the essence of the 1978 Spanish Constitution and defend the unity of Spain as a single country). Secondly, we must highlight the existence of a terrorist group, ETA. ETA has sought to gain independence for the Basque Country through violence and has been perpetrating violent acts for more than 40 years, causing over 800 deaths in Spain (Ormazabal, Reference Ormazabal2003). In relation to ETA’s violence, it is also important to highlight what was known as the Kale borroka (literally, street fighting). Kale borroka refers to acts of street violence perpetrated mainly by young ETA sympathizers, some of whom were actually ETA members. Moreover, we must also bear in mind the existence during the 1980s of the GAL, armed parapolice groups established to fight against supposed ETA members or sympathizers. The Basque government’s “Victims of Violence Motion” (November 2000) cited 66 deaths in Basque Country and 19 more in other parts of Spain at the hands of the GAL and other similar groups (Ormazabal, Reference Ormazabal2003). Finally, there have also been accusations of torture leveled against the police by prisoners linked to ETA. These accusations have been investigated by (among others) Amnesty International.
Gender differences in emotion
Returning to our study, and specifically to gender differences in emotions, the review by Brody and Hall (Reference Brody, Hall, Lewis, Haviland-Jones and Feldman Barrett2008) of self-report studies analyzing gender differences in the frequency and intensity of emotions found that (in keeping with popular belief) women report more intense positive emotions, such as joy or feelings of wellbeing. They also report more intense negative emotions such as sadness, fear, guilt and shame (Fischer, Rodríguez-Mosquera, van Vianen, & Manstead, Reference Fischer, Rodríguez-Mosquera, van Vianen and Manstead2004). Similarly, it has also been found that women report more intense fear of terrorism (Nellis, Reference Nellis2009). It seems that the differences between men and women in relation to positive emotions are clearest in those emotions linked to intimate interpersonal relationships (Brody & Hall, Reference Brody, Hall, Lewis, Haviland-Jones and Feldman Barrett2008).
In regard to empathy, when what is measured is empathy as an affective response, rather than just perspective taking, gender differences are observed, with women scoring higher in intensity (Etxebarria, Ortiz, Conejero, & Pascual, Reference Etxebarria, Ortiz, Conejero and Pascual2009; Lennon & Eisenberg, Reference Lennon, Eisenberg, Eisenberg and Strayer1992; Rueckert & Naybar, Reference Rueckert and Naybar2008). Also, it has been observed that women tend to score higher in ethnocultural empathy (Cundiff & Komarraju, Reference Cundiff and Komarraju2008), a type of empathy linked to empathy with people from ethnic groups and cultures different from one’s own.
In relation to pride, a positive self-conscious emotion, few studies exist which focus on gender differences. It has been found, however, that confidence and pride are more common in men than in women (Collins & Frankenhaeuser, Reference Collins and Frankenhaeuser1978). Another, more recent study carried out using two very large sample groups (Brebner, Reference Brebner2003), one Australian (2,199 Australian students, friends and relatives) and the other international (made up of 6,868 students from 41 different countries, including a group of 286 Australian students), revealed that among Australians, experiences of pride were significantly more frequent and intense in men than in women. However, in the international sample, no significant differences were found between the two genders.
In general, in the interpersonal field guilt tends to be felt more intensely by women (Etxebarria et al., Reference Etxebarria, Ortiz, Conejero and Pascual2009; Lutwak, Ferrari, & Cheek, Reference Lutwak, Ferrari and Cheek1998). Similarly, it has also been found that women tend to report more intense feelings of shame (Ferguson & Eyre, Reference Ferguson, Eyre and Fischer2000; Fischer et al., Reference Fischer, Rodríguez-Mosquera, van Vianen and Manstead2004).
Collective emotions (pride, guilt and shame, etc.) are emotions that people feel as members of certain groups with which they identify. These emotions are arousing increasing interest (Mackie & Smith, Reference Mackie and Smith2002; Tiedens & Leach, Reference Tiedens and Leach2004) and play an especially relevant role in the political and social context. Due to the absence of studies analyzing gender differences in these emotions in the political sphere, research into this field is of particular interest. In this study, we were specifically interested in analyzing collective guilt, shame and pride.
In relation to gender differences in anger, the results of those studies which analyze this question through self-report measures are inconsistent (Brody & Hall, Reference Brody, Hall, Lewis, Haviland-Jones and Feldman Barrett2008). While many fail to find any differences between men and women in regard to anger (Averill, Reference Averill1983; Fischer et al., Reference Fischer, Rodríguez-Mosquera, van Vianen and Manstead2004), others do, such as one in which men were found to score higher in the anger-out dimension and women in the anger-in dimension (Iqbal & Ahmad, Reference Iqbal and Ahmad1993). Anger-in refers to the experience of anger in which anger is denied and repressed, while anger-out refers to the experience in which anger is freely expressed. Finally, in her review on the question, Kring (Reference Kring and Fischer2000) states that in light of the studies analyzed, it is not possible to conclude either that men feel more anger than women or that women feel more anger than men, although it is also not possible to conclude that no differences exist between the genders in this respect.
Gender differences in attitude towards violence
Also, in regard to aggressive behavior, which is often the result of anger, Crick et al. (Reference Crick, Werner, Casas, O’Brien, Nelson, Grotpeter, Markon and Bernstein1999) provide empirical evidence that “relational” aggression (i.e. behaviors which hurt others, harming or threatening relationships or feelings of acceptance, friendship or inclusion in the group) is more common among girls than among boys. However, other data contradict this conclusion, failing to find these differences (Saket, Reference Saket2005) and some studies have even found directly opposing ones (David & Kistner, Reference David and Kistner2000). Whatever the case, in relation to other forms of aggression, ample evidence exists of a greater frequency of these behaviors among men. In general, the studies carried out in different countries and with different age groups have found that men engage in more aggressive behaviors than women (Eagly & Steffen, Reference Eagly and Steffen1986; Etxebarria, Apodaca, Eceiza, Fuentes, & Ortiz, Reference Etxebarria, Apodaca, Eceiza, Fuentes and Ortiz2003; Tangney & Dearing, Reference Tangney, Dearing, Masling and Bornstein2002).
On a different note, we are not aware of any empirical studies that have analyzed the differences between men and women in relation to their support or rejection of violent groups.
Gender differences in forgiveness
Baumeister, Exline, and Sommer (Reference Baumeister, Exline, Sommer and Worthington1998) define forgiveness explicitly as an emotional attitude. Specifically, they believe that to forgive someone means to cease feeling anger or resentment. In fact, research supports a close relationship between emotions and forgiveness, particularly the negative relationship between anger and forgiveness (Tam et al., Reference Tam, Hewstone, Cairns, Tauch, Maio and Kenworthy2007).
In accordance with that proposed by Baumeister et al. (Reference Baumeister, Exline, Sommer and Worthington1998), we believe that, in the intergroup context, to forgive is to cease feeling negative emotions such anger or contempt towards the other group’s members. However, based on the definition established by Enright and the Human Development Study Group (Reference Enright, Kurtines and Gerwitz1991), we also believe that to forgive goes beyond ceasing to feel negative emotions and implies feeling empathy towards the other group.
In relation to the association between emotions and forgiveness, in the interpersonal field a positive relationship has been found between empathy and forgiveness of others (Konstam, Chernoff, & Deveney, Reference Konstam, Chernoff and Deveney2001; Macaskill, Maltby, & Day, Reference Macaskill, Maltby and Day2002), as well as between guilt and forgiveness, and between forgiveness and a reduction of anger (Konstam et al., Reference Konstam, Chernoff and Deveney2001). Similarly, in the field of collective violence, a positive relationship has been found between empathy and intergroup forgiveness (Moeschberger, Dixon, Niens, & Cairns, Reference Moeschberger, Dixon, Niens and Cairns2005), between collective guilt and intergroup forgiveness (Hewstone et al., Reference Hewstone, Cairns, Voci, McLernon, Niens, Noor, Branscombe and Doosje2004) (both studies carried out in Northern Ireland) and between collective guilt and apology in Australia (McGarty et al., Reference McGarty, Pedersen, Leach, Mansell, Waller and Bliuc2005).
So, what about gender differences in relation to forgiveness? Two meta-analyses that researched the question found different results. One meta-analysis (Miller, Worthington, & McDaniel, Reference Miller, Worthington and McDaniel2008), which included a total of 70 studies exploring gender differences in this area, concluded that women are more likely to forgive than men; however, since the effect size was small (d = .28), this conclusion should be viewed with caution. Also, of the 70 studies reviewed, most focused on the general concept of forgiveness of others (59). Other data contradict the conclusions drawn by this study. Specifically, another meta-analysis (Fehr, Gelfand, & Nag, Reference Fehr, Gelfand and Nag2010) has recently been published which finds a very different set of results. This meta-analysis, which analyzes 175 studies, concludes that men and women do not differ at all in regard to tendency to forgive. This meta-analysis focuses on individual acts of interpersonal forgiveness between a single victim and a single offender.
Despite the widespread nature of the stereotype of women as being more inclined to apologize than men (a stereotype which coincides with the aforementioned gender differences in empathy and guilt), very few studies actually analyze this question (Holmes, Reference Holmes1989; Schumann & Ross, Reference Schumann and Ross2010). In a recent study analyzing the offenses and apologies registered daily by a sample group of men and women, women were found to report offering more apologies than men, but they also reported committing more offenses (Schumann & Ross, Reference Schumann and Ross2010). The authors suggest that this difference may be due to the different threshold that men and women have for considering certain behaviors offensive, a threshold which is lower in the case of women.
Gender differences in tolerance
Finally, let us turn our attention to political tolerance. In Psychology, the majority of theoretical and empirical studies on political tolerance define the term as a willingness to defend the human rights of different sociopolitical groups, especially those that one likes least (Avery, Reference Avery1988; Sotelo, Reference Sotelo2000). Similarly, tolerance may also be characterized as a willingness to respect the opinions and practices of others which deviate from the norm (Ben-Ze’ev, Reference Ben-Ze’ev2001).
Studies which have analyzed gender differences in political tolerance are few and far between. Moreover, the few studies that have been conducted fail to point to any definite conclusion: some find that men are more tolerant than women (Golebiowska, Reference Golebiowska1999), others fail to find any differences (Sotelo, Reference Sotelo2000) and yet others claim that women are more tolerant than men (Avery, Reference Avery1988; Lozano & Etxebarria, Reference Lozano and Etxebarria2007). In specific terms, Lozano and Etxebarria (Reference Lozano and Etxebarria2007) analyzed gender differences in tolerance in a sample of Basque adolescents. The results revealed that girls scored higher than boys in tolerance in general, as well as in all of the subscales of tolerance, including that related to political ideas.
However, as Avery (Reference Avery1988) suggests, it may be that the inconsistencies observed between the various studies which have analyzed gender differences in political tolerance reflect a different conceptualization of the term itself. For example, this author argues that some research has focused almost exclusively on the right to dissent, whereas other research includes measures with some items referring to women’s rights. Moreover, these inconsistencies must be assessed also in light of the cultural differences which exist between different contexts. Is a nationalist group in Basque Country really comparable with one in Northern Ireland or the United States? Thus, the political groups towards which tolerance is directed may have very different meanings and may provoke very different reactions in different countries, making direct comparisons difficult and conclusions almost impossible.
In conclusion, considering the importance of emotions in contexts of intergroup conflict and political violence (Branscombe & Doosje, Reference Branscombe and Doosje2004; Mackie et al., Reference Mackie, Silver, Smith, Tiedens and Leach2004), as well as in relation to other aspects such as intergroup forgiveness (Cehajic et al., Reference Cehajic, Brown and Castano2008; Hewstone et al., Reference Hewstone, Cairns, Voci, McLernon, Niens, Noor, Branscombe and Doosje2004; Tam et al., Reference Tam, Hewstone, Cairns, Tauch, Maio and Kenworthy2007), in this study, we aimed to analyze the differences between women and men in all these questions. In light of the results found by the studies on gender differences reviewed in this section, there seems to be sufficient empirical support to conclude that differences do in fact exist in the emotional experience of men and women, and that these differences are especially evident in relation to certain emotions and may also exist in other aspects such as attitude towards violence. Given the differences found between women and men in the interpersonal field, it seems logical to conclude that similar differences may also exist in other fields, such as the political sphere. Indeed, the few studies that we found which focus on the intergroup context support this conclusion; it seems that women feel more intense ethnocultural empathy (Cundiff & Komarraju, Reference Cundiff and Komarraju2008), as well as greater fear of terrorism than men (Nellis, Reference Nellis2009).
But, why should we care about these differences? The identification of differences in questions which are important to peaceful and harmonious coexistence and solidarity, etc., such as empathy, attitude to violence and forgiveness, for example, is no trivial matter. Identifying such differences leads to the conclusion that we cannot afford to overlook the importance of whether initiatives which aim to overcome violence are led by men alone, women alone or mixed gender groups.
The reason for these differences, however, is a totally different question. Some authors have pointed to the different socialization of girls and boys as one possible origin of the differences found in this respect (Brody & Hall, Reference Brody, Hall, Lewis and Haviland1993). This includes both differences in peer socialization and differences in family socialization. However, other elements such as language development (which occurs earlier in girls) (Brody & Hall, Reference Brody, Hall, Lewis and Haviland1993), temperament, cultural values and display rules, etc. must also be taken into account (Brody & Hall, Reference Brody, Hall, Lewis and Haviland-Jones2000). Moreover, when talking about differences in emotions and attitudes in the socio-political sphere, we should probably also include other factors, such as political socialization, which may be different in men from in women.
Objectives and hypotheses
The general research project within which this study is located has two basic objectives. Firstly, it aims to analyze the relationship between diverse emotional variables and identity, ideology, segregation and various other variables linked to victimization. And secondly, it aims to analyze the relationships between all the aforementioned variables and various measures of attitude towards forgiveness, tolerance and, specifically, attitude towards ETA. The results related to both these objectives will be presented in two papers which we hope to publish soon. The first paper will include the results relating to emotional variables and forgiveness, while the second, shorter paper, will present the results relating to tolerance and attitude towards ETA.
In addition to the two aforementioned objectives, the project also provided an ideal opportunity for exploring a third question - our third objective: gender differences in emotions and attitudes related to intergroup conflict and political violence, an intriguing and understudied topic. It is this question that is explored in this study.
Regarding the hypotheses established in this present study, it should be pointed out that for certain variables, exploratory hypothesis were formulated. This is the case in relation to positive and negative emotions in general, tolerance and forgiveness.
We will now proceed to present the hypotheses upon which the study was based.
-
1. Firstly, we believed that women would experience fear for political reasons more intensely than men.
-
2. We believed that women would feel more intense empathy with diverse groups.
-
3. We also believed that women would experience more intense guilt (including collective guilt).
-
4. We hypothesized that women would score higher in collective shame.
-
5. We hypothesized that men would score higher in collective pride.
-
6. Based on studies which have found that men engage in more aggressive behaviors than women, we hypothesized that women would report a lower intensity of both hostile emotions towards people who think differently from themselves and of Schadenfreude. Schadenfreude is a German loanword that describes an emotion for which there is no word in English (or indeed in Spanish or Basque). It can be defined as “pleasure derived from the misfortunes of others”. Thus, this emotion is felt when an event is misfortunate for another person and, for this very reason, is a source of pleasure to the person experiencing the emotion (Ortony, Clore, & Collins, Reference Ortony, Clore and Collins1988).
-
7. Again, based on studies which have found that men engage in more aggressive behaviors than women, in relation to the justification of violence in general, we believed that men would justify it more than women.
-
8. As for attitude to ETA, given that this attitude may be strongly influenced by identity and political ideology, over and above the possible influence of gender, we hypothesized that there would be no differences between women and men.
Method
Participants
The sample comprised 728 people. All participants had been living for at least one year in either Basque Country or Navarra; mean time living in Basque Country or Navarra was 25.98 years (SD = 13.01). Participants were aged between 17 and 72 (M = 29.15, SD = 12.81). 45.5% were men and 54.5% were women.
The aim was to ensure that our sample was as diverse as possible, particularly in regard to national identity and political ideology. Consequently, the sampling process was non-probabilistic, and even, to a certain degree, deliberate. Thus, in addition to the accessible population (e.g. university students), we also included people belonging to particularly important or significant groups within the political and social life of Basque Country. Table 1 contains a brief description of the sample group in relation to identity and ideology; specifically: identification with the Spanish, identification with the Basques, sympathy towards Basque nationalism and sympathy towards leftwing or rightwing politics.
Table 1. Description of the Sample in relation to the following Variables: Identification with the Spanish, with the Basques, Sympathy with Basque Nationalism and Sympathy for (leftwing or) rightwing politics
ª In the variable Sympathy for (leftwing or) rightwing politics, 1 indicated strong sympathy for the left and 7 strong sympathy for the right.
Instrument
Given the specific aims of the current study, most of the variables examined came from scales created ad hoc. Three criteria were followed during the creation of the variables. In some cases, we followed a rational criterion (arithmetic mean of the scores), based on both theory and our knowledge of the true situation. In these cases, the criteria were established basically by the authors of the current study. Subsequently, the internal consistency coefficient was calculated. In other cases, in which the set of items was fairly diverse and we had no unequivocal criterion, we opted to follow an empirical criterion using exploratory factor analyses (varimax rotation). In these analyses, the saturation minimum was set at .3. Finally, in certain other cases, we followed a mixed criterion, i.e. we grouped some items into a variable using their arithmetic mean, while the remaining items were reduced to variables by means of factor analyses. In addition to these three criteria, some other variables were created from a single item.
Emotional variables
Table 2 sums up the long process by which the emotional variables were created. The Table shows the name of the variable and gives an example of an item. It also specifies the criterion followed to create the variable, the number of items used and the Cronbach’s alpha if the variable was isolated using a rational criterion. However, some clarifications are required.
Table 2. Description of the Instrument: Emotional Variables
Firstly, the items from which the majority of the variables were created (Positive and Negative emotionality, Fear for political reasons, Indifference, Collective pride, Collective shame, Empathy with those who suffer, Empathy with those who suffer and think like I do, Empathy with those who suffer even though they think differently from me, Negative emotions towards those who think differently from me and Schadenfreude) allude to “the current political and social situation in Basque Country”.
Collective guilt, which had been measured through a single item and referred to guilt felt over violent acts perpetrated by other Basque citizens, was subsumed into the variable Guilt over ETA’s violence, by means of a factor analysis.
Furthermore, all the items from which the emotional variables were created referred to intensity and were measured on a 7-point response scale (1 = not at all, 7 = a lot), with the exception of those which resulted in the variable Positive and Negative political-emotional climate. The scale from which these last two variables derived, the Political-Emotional Climate Scale, deserves special comment. This scale was adapted from the Emotional Climate Scale by Páez et al. (Reference Páez, Ruiz, Gailly, Kornblit, Wiesenfeld and Vidal1997) in accordance with both the dimensions of the concept of emotional climate described by de Rivera (Reference de Rivera and Strongman1992), and the specific characteristics of the Basque political situation. This new scale consisted of 18 items (4 of which were excluded) and asked participants to assess the social climate in Basque Country. The response scale used was the same as for Páez et al.’s Emotional Climate Scale (Reference Páez, Ruiz, Gailly, Kornblit, Wiesenfeld and Vidal1997) (1 = not at all and 5 = a lot).
Given the importance of empathy, we believe it is necessary to describe in more detail the process by which the following variables were created: Empathy with ETA victims, Empathy with prisoners and Empathy with the Basque nationalist political sector. 14 items were designed to assess empathy towards different groups. In these items, participants were asked: “To what extent do you feel sorry for these people?” referring to fourteen different groups of people, e.g. “ETA victims” and “Families of Basque political prisoners”. Participants were asked to respond on a 7-point scale (1 = not at all sorry and 7 = very sorry). The exploratory factor analyses revealed three factors for analysis: Empathy with ETA victims (in which the items with the most weight were those linked to sympathy for the suffering of ETA victims); Empathy with prisoners (in which the main items were those related to sympathy for the suffering of prisoners’ families); and finally, Empathy with the Basque nationalist political sector (in which the main items were those related to sympathy for the suffering of members of Basque nationalist parties).
Attitude variables
Justification of violence. Participants were asked to what extent they believed that the use of violence could be justified. Respondents could choose between five graduated and exclusive answers: “Never. The use of violence is never justified”, “In extreme circumstances, for example genocide or self-defense, the use of violence may be justified”, “In some circumstances, such as war, the use of violence may be justified”, “In numerous circumstances, for example in a country with a democratic regime that fails to recognize a nation’s right to self-determination, the use of violence by this nation may be justified” and finally “In the majority of cases the use of violence can be justified”.
Attitude towards ETA. Participants were asked about their “attitude towards ETA today”. Again, respondents were asked to choose between five graduated and exclusive answers: “Total rejection. I reject both their violent means and their ends”, “I reject their violent methods although I share their aims”, “I neither reject nor defend them, I am indifferent”, “I support them, although there are certain aspects I would criticize” or “I totally support them”.
Forgiveness towards different groups. Participants were asked: “If the necessary conditions were to arise in the future, to what extent do you believe these people should be forgiven?”, followed by a list of 19 different groups (for example: “ETA”, “Gobierno Español” and the members of each of the political parties of Basque Country). A 7-point response scale was provided for each group (1 = they should not be forgiven at all and 7 = they should be totally forgiven). Three factors were obtained from the exploratory factor analyses: Willingness to forgive ETA and the perpetrators of street violence (in which the principal items were willingness to forgive the perpetrators of street violence and ETA); Willingness to forgive the police and the Spanish Government (in which the principal items were willingness to forgive the national police/civil guard, the Spanish Government and the French Government); and finally, Willingness to forgive the Basque nationalist sector (in which the main items were willingness to forgive the members of the various Basque nationalist parties).
Need for apology and forgiveness. We included two questions on this issue: “Do you believe it is necessary for certain people or groups to apologize before peace and reconciliation can be achieved in Basque Country?” and “Do you believe it is necessary for certain people or groups to forgive (other people or groups) in order to achieve peace and reconciliation in the Basque Country?”. The answers to these questions were simple yes/no responses. The exploratory factor analysis revealed a single factor: Need for apology and forgiveness.
Tolerance. We asked participants about the degree to which they accepted each of the following five identities: “Spanish only”, “Basque only”, “equally Spanish and Basque”, “more Spanish than Basque” and “more Basque than Spanish”. Respondents answered on a 7-point scale (1 = totally reject it and 7 = totally accept it). Previously, participants had been asked to choose their own identity from among these options. We created a variable called Acceptance of different identities, which was the result of the mean degree of acceptance by participants of the four identities which were different from their own. Participants were also asked about their willingness to “compromise on some of their political objectives in order to achieve peace in Basque Country” (1 = not at all willing and 7 = very willing).
We carried out a factor analysis with the variable Acceptance of different identities and the question about compromising on certain political objectives, obtaining a single factor which we termed Tolerance.
In addition to all these measures, the instrument also included others relating to national identity, political ideology, victimization and ideological segregation, which will not be analyzed in this study, but rather in other papers still pending publication.
Procedure
A concerted effort was made to ensure that all political ideologies and national identities were represented in the study.
The tests were administered to university students, who completed the questionnaires individually during class time. In addition to this, a series of groups were also chosen on the basis of their political and social importance within Basque Country, and in November 2005 they were sent an introductory letter requesting their collaboration in the study. These groups included: political parties, associations working in favor of victims of terrorism, prisoner support associations, pacifist associations and human rights groups, associations both for and against Basque nationalism and security forces. Although certain groups were very reluctant to respond to the questionnaire, and some even refused outright (meaning that not all the groups we contacted ended up participating in the study), in the end we achieved a relatively broad range of participants from all ideologies, including members of the most extreme groups, which had initially proven most reluctant to participate in the research project.
After contacting the selected groups over the telephone and having gained their consent, we delivered the questionnaires in individual envelopes. Said questionnaires were completed individually and returned either in person or by conventional mail after approximately 10 days.
In all cases the questionnaires were completed anonymously.
The data were collected between November 2005 and February 2006. The political and social context in which the collection process took place was characterized by a high level of uncertainty and optimism resulting from a long period with no attacks (since May 2003) and from certain rumors regarding a possible ETA ceasefire. Consequently, we made every effort to ensure that the data were collected as diligently as possible. A short time later, on 23 March 2006, ETA declared a ceasefire (although it did not declare the definitive cessation of its armed activities until October 2011).
Results
Bearing in mind the age difference between the women and men in the sample group (mean ages 26.52 and 32.42 respectively), and in light of our suspicion regarding possible gender differences in the variables related to national identity, political ideology and victimization, various mean difference analyses were conducted.
The results revealed significant differences in age, ideologized identity (a variable made up of items relating to national identity and political ideology) and having suffered as a result of street violence or from any of ETA’s violent activities. The women in our sample were not only younger than the men, t(659) = 6.25, p = .001, d = .44, 95% CI [4.05, 7.76], they also scored higher in ideologized identity (greater degree of sympathy with Basque nationalism and the political left, greater degree of identification with Basques and a lesser degree of identification with the Spanish), t(571) = –4.84, p = .001, d = –.36, 95% CI [–.54, –.23]. Furthermore, women reported having suffered less as the result of ETA’s violence, t(364) = 4.79, p = .001, d = 0.35, 95% CI [.25, .60]. To calculate the Cohen’s d the standard deviation of the male respondents was used as a reference in all cases.
In order to analyze gender differences in the variables considered in this study, we decided to use MANCOVAs, the best method for reducing the chances of making Type I errors and for controlling multicolinearity between dependent variables. Through this kind of analysis we were able to study differences between women and men, controlling the effect of the three covariates. We also conducted ANCOVAs. As recommended by Field (Reference Field2011) in these cases, we also conducted discriminant analyses in order both to gain a better understanding of the relationships which exist between dependent variables and to reduce the chances of making Type I errors. The covariates were included along with the rest of the emotional and attitudinal variables in the discriminant analyses. In this section, we will first present the results of the MANCOVAs. Subsequently, we will present the general results of the discriminant analyses. Finally, we will present only the results of those variables which were significant in the ANCOVAs and which also appeared in the function of the discriminant analyses with a weight of .30 or more (standardized canonical discriminant function coefficients).
In specific terms, we carried out four MANCOVAs and four discriminant analyses. The grouping of the variables in each of the analyses was guided at all times by theoretical criteria. The first one contained general emotional variables: positive and negative emotionality, fear for political reasons, indifference, collective pride, collective shame, and positive and negative political-emotional climate. The second group included empathy measures, negative emotions towards those who think differently from oneself and Schadenfreude, and guilt measures. The third group included justification of violence and attitude towards ETA. And, finally, the last group contained measures of forgiveness and tolerance.
Of the total sample group of 728 participants, 28 were eliminated for having a score in one of the dependent variables which was above or below the mean by 3.5 times the standard deviation. Furthermore, the analyses revealed a high number of lost cases (between 200 and 400 approximately).
Gender Differences in General Emotional Variables
The MANCOVA Footnote 1 revealed a significant effect, which indicates differences between men and women in emotionality, F(8, 451) = 3.22, p = .001, η² partial = .054. The MANCOVA was followed up with discriminant analyses, which revealed one discriminant function, and explained 100% of the variance, canonical R² = .17. This function significantly differentiated women from men, Λ = 0.83, χ²(11) = 83.68, p = .001, positively for men and negatively for women.
Significant differences were found in the ANCOVAs in three emotional variables, with a weight of over .30 in the discriminant analyses: positive emotionality, fear for political reasons and indifference. Men reported more intense positive emotionality than women in relation to the political situation, F(1, 458) = 5.35, p = .021, η² partial = .012; weight in the discriminant function (standardized canonical discriminant function coefficient) = .35. As we had hypothesized, women reported more intense fear for political reasons, F(1, 458) = 9.58, p = .002, η² partial = .02; weight in the discriminant function = –.34. And finally, men reported more intensity in indifference, F(1, 458) = 6.06, p = .014, η² partial = .013; weight in the discriminant function = .36. In relation to indifference, we should take into account that Levene’s test was significant, F(1, 461) = 6.31, p = .012, which means that the result may be affected by the correction of heteroskedasticity. Table 3 shows the marginal means of the ANCOVAs of this section for women and men.
Table 3. Differences between Men and Women in General Emotional Variables: Marginal Means
Note. The response scale was from 1 to 7 in all cases except that of positive and negative political-emotional climate, which was from 1 to 5.
* In these variables significant differences were observed between men and women in ANCOVAs, with a minimum weight of .30 in the discriminant analyses.
Gender Differences in Empathy, Negative Emotions towards those who Think Differently from Oneself and Schadenfreude, and Guilt
The MANCOVA revealed a significant effect, F(10, 367) = 2.69, p = .003, η² partial = .068. Box’s test Footnote 2 of equality of covariance matrices was significant, F(55, 436684) = 1.37, p = .037. The discriminant analysis revealed one discriminant function, and explained 100% of the variance, canonical R² = .18. This function significantly differentiated women from men, Λ = 0.82, χ²(13) = 72.26, p = .001, positively for women and negatively for men.
Of all the variables included in the analyses, three revealed significant differences in the ANCOVAs, with a weight of .30 or over in the discriminant analyses: empathy with prisoners, empathy with those who suffer and think like oneself, and Schadenfreude. Women reported more intensity in empathy with prisoners, F(1, 376) = 11.84, p = .001, η² partial = .031; weight in the discriminant function = .36. Women also reported more empathy with those who suffer and think like oneself, F(1, 376) = 9.91, p = .002, η² partial = .026; weight in the discriminant function = .30. For their part, men reported more intense Schadenfreude, F(1, 376) = 4.41, p = .036, η² partial = .012; weight in the discriminant function = –.31. Levene’s test was significant for empathy with those who suffer and think like oneself, F(1, 379) = 10.48, p = .001, indicating that correcting for heteroskedasticity may affect the result. Table 4 shows the marginal means of the ANCOVAs of this section for women and men.
Table 4. Differences between Men and Women in Empathy, Negative Emotions towards those who Think Differently from Oneself and Schadenfreude, and Guilt: Marginal Means
Note. The response scale was from 1 to 7 in all cases except that of empathy with ETA victims, empathy with prisoners, empathy with the Basque nationalist sector and guilt over ETA’s violence, in which factorial scoring was used.
* In these variables significant differences were observed between men and women in ANCOVAs, with a minimum weight of .30 in the discriminant analyses.
Gender Differences in Justification of Violence and Attitude towards ETA
The MANCOVA revealed no significant effect, F(2, 402) = 1.81, p = .165, η² partial = .009. Given that the variables included in this analysis are fairly different (the first refers to general justification, not linked to any specific context, while the second refers to attitudes towards a specific terrorist group, which may be influenced by identity, ideology and victimization, etc.), and in light of the importance of both questions, we decided to study the results of the ANCOVAs. Upon analyzing the ANCOVAs, no significant differences were found in any of the variables, although in relation to the justification of violence, the p was near to significance level F(1, 403) = 3.58, p = .059, η² partial = .009 (men scoring higher than women). Motivated mainly by the importance of the questions analyzed here, and given that the differences between men and women in the justification of violence reached marginal probability, we decided to examine the results of the discriminant analysis. The discriminant analysis revealed one discriminant function, and explained 100% of the variance, canonical R² = .11. This function significantly differentiated women from men, Λ = 0.89, χ²(5) = 48.60, p = .001. The function was positive for men and negative for women. Justification of violence had a value in the standardized canonical discriminant function coefficient of .30. Table 5 shows the marginal means of the ANCOVAs of this section.
Table 5. Differences between Men and Women in Justification of Violence and Attitude towards ETA: Marginal Means
Note. The response scale was from 1 to 5.
Gender Differences in Forgiveness and Tolerance
The MANCOVA was significant, F(5, 253) = 2.71, p = .021, η² partial = .051, indicating differences between men and women in either forgiveness or tolerance or a combination of both. Box’s test was significant, F(15, 247156.1) = 3.21, p = .001. The discriminant analysis revealed one discriminant function (100% of the explained variance), canonical R² = .16. This function significantly differentiated women from men, Λ = 0.84, χ²(8) = 43.83, p = .001, positively for women and negatively for men.
We found only one relevant variable in the ANCOVAs and discriminant analysis: Need for apology and forgiveness. Women considered to a greater extent than men both apology and forgiveness to be necessary in order to achieve peace and reconciliation, F(1, 257) = 6.24, p = .013, η² partial = .024; weight in the discriminant function = .46. For the aforementioned variable we should take into account that Levene’s test was significant, F(1, 260) = 11.22, p = .001, which means that correcting for heteroskedasticity may affect the result. Table 6 shows the marginal means of the ANCOVAs of this section for women and men.
Table 6. Differences between Men and Women in Forgiveness and Tolerance: Marginal Means
Note. In all variables factorial scoring was used.
* In these variables significant differences were observed between men and women in ANCOVAs, with a minimum weight of .30 in the discriminant analyses.
Discussion
As stated at the end of the introduction, the study presented here forms part of a broader research project which aims to analyze the relationship between variables such as identity, ideology and ideological segregation and diverse emotional variables and the willingness to forgive. Also, within the framework of this research project, which was carried out in a context of intergroup conflict and violence, the aim was to analyze the relationships existing between emotional variables and willingness to forgive, among others.
However, one of the other objectives of the research project was to analyze gender differences in the variables being studied, and it is this analysis that is presented in this paper. Given the absence of research analyzing these differences in contexts of political violence, this study is of particular interest.
The main questions we faced at the start of the research project were as follows: are the gender differences in emotions found in studies carried out in the interpersonal field also applicable to emotions regarding the political context and political violence? And what about gender differences in attitudes which are relevant in such contexts, such as forgiveness, attitude towards violence and tolerance? While some of the results found in our study (e.g. more intensity in fear or more intensity in some types of empathy among women) were similar to those found in the interpersonal field, others were slightly different and deserve to be studied in more detail in the future. For example, the absence of differences in guilt and some measures of empathy suggest that in the intergroup context, the emotional experiences of women and men may be influenced by other factors, such as identity or ideology, factors which render gender less important and reduce, eliminate or even invert the differences found in the interpersonal context.
The following is a discussion of the different results obtained in comparison with the hypotheses established on the basis of our review of the existing literature.
Gender Differences in General Emotional Variables
Women scored lower than men in intensity of positive emotionality. This result is somewhat surprising given that, in general, previous studies show that women tend to score higher in positive emotions such as joy and wellbeing (Brody & Hall, Reference Brody, Hall, Lewis, Haviland-Jones and Feldman Barrett2008). Nevertheless, it may be that this result is not actually as incongruous as it may first appear. The differences in positive emotions found in the studies cited by Brody and Hall (Reference Brody, Hall, Lewis, Haviland-Jones and Feldman Barrett2008) allude mainly to the interpersonal context. Thus, it may be that in the collective sphere, and more specifically in that relating to political violence, these differences are reduced, or even inverted, as in the case of this study.
Moreover, in accordance with our hypothesis, women scored higher in fear for political reasons. This is consistent with the results of other studies which found that women feel negative emotions such as sadness and fear (Fischer et al., Reference Fischer, Rodríguez-Mosquera, van Vianen and Manstead2004), as well as fear in the field of political violence (Nellis, Reference Nellis2009), more intensely.
On the whole, the results discussed so far suggest that the emotional experience of women in the political context is more negative than that of men: they have less intense positive emotions and more fear for political reasons. This may also imply that women take less interest in the political sphere than men, and engage in certain avoidant behaviors in this field. For their part, perhaps men find active participation in politics more satisfying because of the emotions they experience in this respect, an occurrence which would explain, in part at least, the greater presence of men in the political field. However, men scored higher in indifference. This result, and its implications, deserve further study.
According to previous studies carried out in the interpersonal sphere, women tend to experience shame more intensely than men (Ferguson & Eyre, Reference Ferguson, Eyre and Fischer2000; Fischer et al., Reference Fischer, Rodríguez-Mosquera, van Vianen and Manstead2004), while men tend to experience more intensity in pride than women (Brebner, Reference Brebner2003; Collins & Frankenhaeuser, Reference Collins and Frankenhaeuser1978). However, we failed to find any gender differences in relation to either collective shame or collective pride. With respect to pride, we should take into account that in Brebner’s study (2003), gender differences were only found in the Australian sample (not in the international one).
The lack of consistency between our results and some results found by previous studies with respect to shame and pride may be due to either cultural differences or to the fact that what we were analyzing in our study was collective shame and pride, not personal shame or pride. We are not aware of any other paper analyzing gender differences in collective pride or collective shame with which to compare our results. In any case, it may be possible that the gender differences observed in pride or shame as personal emotions are simply not present in collective pride or shame, emotions in relation to which socialization and the “feeling rules” (Hochschild, Reference Hochschild1983) differ much less from one gender to the other.
Gender Differences in Empathy, Negative Emotions towards those who Think Differently from Oneself and Schadenfreude, and Guilt
Although in light of the results of many studies women tend to be considered as more empathetic than men in general (Cundiff & Komarraju, Reference Cundiff and Komarraju2008; Etxebarria et al., Reference Etxebarria, Ortiz, Conejero and Pascual2009; Lennon & Eisenberg, Reference Lennon, Eisenberg, Eisenberg and Strayer1992; Rueckert & Naybar, Reference Rueckert and Naybar2008), our results show that they are also more empathetic in some types of empathy in the political sphere (empathy with those who suffer and think like oneself and empathy with prisoners), a finding which has important implications. Empathy encourages behavior based on solidarity and the desire to help. It inhibits aggression and fosters forgiveness. However, we should not forget that, although at a descriptive level we observed that women scored higher in most of the empathy measures, only two of them were significant in both ANCOVAs and discriminant analyses.
While the analyses revealed no significant differences in negative emotions towards those who think differently from oneself (for example, disdain, anger, etc.), differences were found in relation to Schadenfreude. Based on studies which have found that men engage in more aggressive behaviors than women, not just physically but verbally also (Eagly & Steffen, Reference Eagly and Steffen1986; Etxebarria et al., Reference Etxebarria, Apodaca, Eceiza, Fuentes and Ortiz2003; Tangney & Dearing, Reference Tangney, Dearing, Masling and Bornstein2002), we hypothesized that women would score lower in these two variables which involve a large dose of hostility. However, despite the fact that women scored lower in both emotional variables, the difference was only significant in the case of Schadenfreude. Schadenfreude, when aimed at members of a group (intergroup Schadenfreude), may have played a pivotal role in some of the worst atrocities of the twentieth century (Spears & Leach, Reference Spears, Leach, Tiedens and Leach2004), such as, for example, those committed against the Jews in Nazi Germany. Although Schadenfreude is a passive emotion and cannot explain the active persecution of rival groups, only the celebration of their suffering, it may provide a context in which aggression or human rights violations are tolerated.
In general, the fact that women reported less intense Schadenfreude and scored higher in certain measures of empathy suggests that, in contexts of intergroup conflict and violence, the presence of women in initiatives aimed at overcoming conflict may help prevent aggression against certain groups and foster respect for human rights. Nevertheless, it is important to clarify here that no differences were found between the genders in either general empathy or respect for all groups. It is a question which requires further analysis in future studies.
In relation to guilt, we had hypothesized that women would score higher than men. Although research into this question suggests that in general, in the interpersonal sphere, women tend to experience more intense guilt than men (Etxebarria et al., Reference Etxebarria, Ortiz, Conejero and Pascual2009; Lutwak et al., Reference Lutwak, Ferrari and Cheek1998), in our study no differences were observed either in guilt over the situation of prisoners or in guilt over ETA’s violence. As we stated at the beginning of this section, since what we are talking about here are feelings of guilt that are closely linked to ideological questions and identity-based issues, it may be that other variables, such as ideology and identity, take precedence here over gender.
Gender Differences in Justification of Violence and Attitude towards ETA
Although the MANCOVA was not significant, the ANCOVA in justification of violence in general was very near significance level and the discriminant analysis was significant, with men justifying the use of violence to a greater extent than women. However, no gender differences were found in relation to attitude to ETA. Based on numerous studies which have found a greater degree of aggressiveness in men than in women (Eagly & Steffen, Reference Eagly and Steffen1986; Etxebarria et al., Reference Etxebarria, Apodaca, Eceiza, Fuentes and Ortiz2003; Tangney & Dearing, Reference Tangney, Dearing, Masling and Bornstein2002), we had hypothesized that men would have a greater tendency than women to justify violence. As for the violence perpetrated by ETA, given that what we are talking about here is an attitude to a specific type of violence imbued with many political and ideological connotations (the fruit of each individual’s education, socialization, etc.), we hypothesized that gender differences would diminish, or even disappear altogether. The results of the study provide some support for these hypotheses. However, the result relative to the justification of violence in general should be interpreted with caution, since neither the MANCOVA or the ANCOVAs revealed any significant differences between women and men.
Gender Differences in Forgiveness and Tolerance
No gender differences were found in relation to the specific forgiveness measures of our study (willingness to forgive ETA and the perpetrators of street violence, willingness to forgive the police and the Spanish Government and willingness to forgive the Basque nationalist sector), in accordance with Fehr et al.’s meta-analysis (2010). Nevertheless, differences between men and women were found in relation to the need for forgiveness and apology in order to achieve peace and reconciliation, with women showing a greater tendency to perceive forgiveness and apology as necessary elements for achieving peace and reconciliation. This last result was consistent with the results of Miller et al.’s meta-analysis (2008) and Schumann and Ross’s study (2010). Given the enormous value of this finding, we believe it is worth exploring in more detail. In light of this, we believe that women may play a key role in the development of processes of forgiveness in contexts of intergroup conflict and violence, such as the one in Basque Country.
In relation to gender differences in political tolerance, the results fail to reveal any differences between men and women. However, in another study carried out in the same context (Lozano & Etxebarria, Reference Lozano and Etxebarria2007), women were found to score higher than men in tolerance of political ideas. This divergence between the two sets of results may be due, among other factors, to the age difference between the two samples, since in Lozano and Etxebarria’s study the participants were adolescents while in our study they were adults. Nevertheless, a more in-depth interpretation of the results suggests that gender differences may, in general, be more acute in relation to generic issues, such as the justification of violence analyzed in this study or tolerance of political ideas as in the study by Lozano and Etxebarria (Reference Lozano and Etxebarria2007); they may also be less evident in relation to more specific issues, such as attitude to a specific violent group or political tolerance of a specific social group, questions which are influenced by a multitude of factors such as socialization, social identity and ideology, etc.
Although this study contributes new findings to the study of gender differences and political psychology, it does have some limitations.
Firstly, the data collection process took place during a specific time period, namely the end of 2005 and the beginning of 2006. This circumstance must be taken into consideration during any interpretation of the findings. Nevertheless, we believe that the majority of the results are fully valid, since although the intensity of the variables may have varied somewhat, there is no reason to believe that the gender differences found have changed significantly over the last few years.
The fact that most of the variables were measured with scales created ad hoc for the current study also implies some limitations. When designing the present research study, we decided to analyze, for example, empathy or forgiveness in relation to specific groups which are important in this context. However, this same specificity which provides our study with a more realistic context also makes it harder to replicate in other contexts outside Basque Country.
The use of self-reports also implies a series of limitations such as, for example, the social desirability bias (for example, in the questions about Schadenfreude and certain forms of empathy). In this sense, we tried to mitigate these effects at the beginning of the questionnaire by asking respondents to be honest in their replies and by assuring them of the anonymous and confidential nature of the whole data collection process. In addition to the social desirability bias, the study of gender differences using this kind of measure is often questioned, with critics alluding to the possible influence of stereotypes (Brody & Hall, Reference Brody, Hall, Lewis, Haviland-Jones and Feldman Barrett2008). Although it is true that stereotypes may influence responses to the self-reports, prompting men and women to respond in accordance with what is expected of them, why is it that we assume that this influence is absent throughout the whole course of men and women’s development and socialization, during which it may result in real differences in their experiences and emotional intensity? Furthermore, as suggested in a study by Etxebarria et al. (Reference Etxebarria, Ortiz, Conejero and Pascual2009), it is not entirely clear that stereotypes prompt women to present their emotions as being more intense than they actually are. This study, which analyzed gender differences in guilt, found that both men and women admitted to not having been totally honest in their responses, presenting their emotions as less intense than they actually were.
Finally, one last limitation should also be borne in mind in relation to the data analyses: missing data. Although the samples finally used in the analyses were large enough, these analyses nevertheless revealed a high number of lost cases, between 200 and 400 approximately. Despite this, however, we have no reason to believe that this loss of subjects was anything other than random. Furthermore, in this study, no estimation of population parameters was carried out and nor was any prior sampling design conducted; rather, we used the available sample (attempting to ensure that it was as diverse as possible). For all these reasons, we decided against using specific methods for handling the missing data (Schafer & Graham, Reference Schafer and Graham2002).
In short, we believe it is necessary to continue analyzing gender differences in emotions and attitudes such as forgiveness in contexts of political violence, which are of great importance to peaceful, harmonious coexistence.
Specifically, and bearing in mind the new era which Basque Country entered following ETA’s declaration of a definitive cessation of its armed activities in October 2011, it would be extremely interesting to continue the line of research opened by this present study. More concretely, it would be interesting to conduct a similar study either now or in the near future, with special attention being to paid to two aspects. Firstly, and in general terms, the study should analyze once again the same emotional and attitudinal variables that were studied in this paper, in order to ascertain whether or not the differences found here are replicated. And secondly, it would be interesting to carry out a more in-depth study of those variables that play a pivotal role in improving relations between people who belong to radically opposed groups; here, special attention should be paid to empathy, Schadenfreude, justification of violence and forgiveness. In relation to empathy, given that we only found differences in two types of empathy, it would be very interesting to study those types or areas in which gender differences do exist in more detail.
Finally, we do not rule out the possibility of using other methodologies such as the qualitative method, which may be particularly appropriate for analyzing such intimate and delicate aspects as those explored in this study.
In general, based on the results of the study, we can conclude that in the intergroup context, while some of our findings are similar to those found in the interpersonal sphere, relevant differences also exist which need to be studied in greater detail.
Women are often excluded from the negotiating table during conflict resolution processes and are relegated to the margins of reconstruction and reconciliation initiatives. It may be that the differences found in this study, for example that women experience some types of empathy more intensely, feel less intense Schadenfreude or perceive forgiveness and apology as more necessary for achieving peace, can be extrapolated to other contexts of intergroup conflict and violence. If this is indeed the case, although women’s lower intensity of positive emotions and greater intensity of fear for political reasons may make it difficult, societies may benefit from women playing a more important role in this kind of process, and in short, in their occupying a more prominent place in all those initiatives which seek to prevent or combat violence and bring about a rapprochement between conflicting groups.
This research project was supported in part by a grant from the Spanish Ministerio de Educación, Cultura y Deporte (Project code: AP2002–0839) at the Universidad del País Vasco/Euskal Herriko Unibertsitatea and thanks to an Assistant Professorship contract at the Universidad Pública de Navarra.
