Collective Innocence and Collective Guilt in the Persecution of Gypsies: From Deviant-Active Minority to Victimized Minority

: Gypsies have been a persecuted minority for more than five centuries. Our paper examines how exposing Spaniard participants to information according to which they personally, or their ingroup collectively, discriminate against the Gypsy minority influences (i) their attitudes toward the Gypsy minority (e

gay/lesbian (55%); Gypsy (56%); or transgender or transsexual (65%).Similar results were observed for the question "if one of your colleagues at work is...", where Gypsies were rated as the most uncomfortable of all the groups on the list.Similarly, a representative survey on ethnic prejudice in Spain assessed attitudes in a pool of 7,161 students and professors from primary and secondary schools regarding various minorities and out-groups relevant to that country (e.g.Gypsies, Moors-Arabs, Blacks, Jews, Indians, Latin Americans, Protestants, foreigners, Portuguese, and Americans).The results again showed that Gypsies were the minority that provoked the greatest rejection (Calvo, 1995, p. 91).In this survey, 45% of respondents openly admitted to personally having antipathy to, and prejudice against, Gypsies.In sum, the Gypsy community today, like six centuries ago, seems to be a prototypical group with "tribal stigma", as Goffman (1963) would say.
What social explanations are proposed from this marginalization of Gypsies?Are Gypsies the ones who deviate from the mainstream and therefore it is their own fault (i.e., majority´s collective innocence), or are they victims of the prejudice and discrimination they receive from the majority (i.e., collective guilt)?In other words, are Gypsies perceived as a deviant minority or as a victimized minority?

Deviant Minority and Innocent Majority vs Victimized Minority and Guilty Majority
Based on the theory of minorities as victims (Moscovici & Pérez, 2007; 2009; Pérez, Molpeceres, Ghosn, Chulvi, 2022), a change in social explanations was observed as beginning in the 1980s.Most of the groups that had been considered deviant, marginal or anomic gained the status of victims entitled to social acceptance.In fact, since the enhanced focus on human rights, such as equality and justice, achieved by civil rights activists in the 1960s, there has been a societal change in the social representations of ethnic and racial prejudices, as well in the social representations of minorities.That majority began to admit that racism was real and unjust, and, by extension, that it was a cause of the marginalization of the minorities that were its target.The accusation of being prejudiced, racist or a "discriminator" began to create a moral conflict in the majority of people.In causal attribution terms, this new representation of ethnic and racial prejudices involves a decreasing sua culpa-that is, less victimization of minorities-and an increasing nostra culpa, an assumption of guilt or responsibility by the majority in the marginalization of minorities.The minority was no longer seen as merely deviant but came to be seen as the victim of the majority´s discrimination.Moreover, minorities have realized that they have this new moral power (Pérez & Molpeceres, 2018).They see that if the majority recognizes them as victims of discrimination, or of unfair historical mistreatment, then they begin to have rights.Social movement denouncing hierarchical microaggressions is a confirmation of this new moral power of minorities (e.g., Sue, 2010; Young, Anderson and Stewart, 2015).
This new era of victimized minorities began to increase in the 1980s.A new societal ethos is now observed, in which governments, institutions, organizations and society in general are prone or willing to face the history of unjust abuse that their ancestors inflicted on certain minority groups (see, for example, Accattoli, 1998;Barkan, 2000;Brooks, 1999;Howard-Hassmann, 2014).The recognition of the victimized minorities, together with the current obsession of society with the victims, as pointed out by Erner (2006; for a related issue on microaggressions, c.f., Campbell and Manning, 2014), has led to collective guilt.As several studies show, this collective guilt increases the desire to apologize and provide restitution or compensation to the victimized minority (e.g.Mallett & Swim, 2004; Miron, Warner, & Branscombe, 2011; Stewart, Latu, Branscombe, & Denney, 2010; for a review, see Wohl, Branscombe, & Klar, 2006).The novelty concerning issues of discrimination and racism against minorities is that the majority stops seeing the minority as something "natural", stops explaining them by the nature of the minorities themselves, and begins to see themselves as a perpetrator of harm to minorities.
The question then arises concerning the factors that have it made possible for the majority to move from collective innocence to collective guilt.Other texts (e.g.Moscovici & Pérez, 2009;Pérez & Molpeceres, 2018) have pointed out factors such as the moral shock produced by the Nazi genocide, the institutionalization of the Universal Declaration of Human Rights (1948), the non-applicability of statutes of limitation for crimes against humanity, group rights versus individual rights, and, most relevant from our point of view, the active struggle of minorities for civil rights and for the extension of democratic principles to everyone equally.All these factors, and indubitably others, have ended up transforming the relationships between the majority and minorities.A political relationship has passed on to an ethical relationship which not only condemns, until it almost becomes a taboo, the discrimination against minorities but in a further step also introduces the ethics of collective social responsibility, for which it is not only about avoiding personal discrimination against minorities but about vindicating rights for them.This ethical paradigm framing the relationships between the majority and minorities, would not only transform the majority but also the minorities themselves.
Minorities have ceased to be seen as mere deviants and have come to be seen as victims of majority practices.But this ethical change in the majority-minority relationship does not seem to be applied automatically to a minority like the Gypsies.It seems that presenting the historical information of the injustice suffered by Gypsies is not enough for the majority to recognize them as their victims.For example, in a previous study (Moscovici & Pérez, 1999), university students were given a thousand-word summary on the legislation and measures against Gypsies applied in several European countries since the 15th century.After reading the summary, in the sua culpa experimental condition, participants were induced to think of Gypsies as a deviant minority.On the contrary, in the mea culpa condition, participants were induced to see Gypsies as a victimized minority, i.e. measures against Gypsies would cause them to distrust majoritarian society and would propel them to marginalize themselves.In a third control condition, participants were not given the summary on anti-Gypsy legislation to read.The most striking result was that no difference was observed between the control condition and the victimized minority conditions on the attitudes and social distance towards Gypsies.With respect to these two conditions, it was the deviant minority condition that induced more negative attitudes and greater social distance towards Gypsies.As an important conclusion, the study showed that it is not enough to present information on the history of the persecution to which the Gypsies have been the target for them to be recognized as victims.
Following this line of research, in the experiment presented in this article, participants were given a brief summary on the history of persecution and injustice suffered by the Gypsy minority for more than five centuries.This historical information can lead to social explanations in which Gypsies are perceived either as a deviant minority-blaming them for their own marginalization (i.e.majority´s collective innocence)-or as a victimized minority-blaming their marginalization on the discrimination done by the majority (i.e.majority´s collective guilt).The main hypothesis of the experiment is that in addition to knowing the history of the discrimination of the minority, it is necessary that the members of the majority recognize the existence of a collective discrimination.Without this change in the representation of the collective discrimination norm by the majority, the information per se of the history of the minority persecution is not enough to grant them the status of victim.In general terms, it would be said that the era of victims does not happen because historians have brought the discrimination of minorities to light but because the majority recognizes itselfor is led to recognize itself-as an agent of discrimination.

Overview and Hypothesis
The experiment used a 2 (target of discrimination: Gypsy minority vs. ad hoc out-group) x 2 (accusation of having discriminated against the Gypsy minority, or the ad hoc out-group, at collective vs. individual level) factorial design.The main objective of the experiment consisted in varying the collective or individual level of discrimination to which the Gypsies are a target.For that purpose, each participant was either individually accused of having been racist against the Gypsies, or they received a collective accusation that they, like the rest of the majority in-group, had been racist against the Gypsies.
The participants then read a text that recalled the persecution suffered by the Gypsies for five centuries throughout their history in Spain.Grounded in the discrimination-level hypothesis of perpetrating majority-victimized minority, collective discrimination rather than individual discrimination will facilitate a greater recognition of Gypsies as victims.The recognition of the minority as a victim could generate a greater moral conflict within the majority, that is, a greater collective guilt.Indeed, several studies (e.g.Mallett & Swim, 2004;Miron et al., 2011;Stewart et al., 2010;Wohl et al., 2006) have shown that collective guilt promotes requests for forgiveness and favorable attitudes towards compensation for the discrimination suffered by minorities.It was then predicted that the recognition of Gypsies as victims, allow more favorable attitudes towards restitution for this minority.It was secondly hypothesized that there would be a transformation of the representation of the minority from deviant to victim: the collective representation of the majority as racist should change the representation of a deviant minority into that of a victimized minority.Related to this transformation, a lower victimization of Gypsies was expected as explanation for their situation of marginalization and a greater inclination of the participants to assume nostra culpa instead of sua culpa for the social marginalization that this minority has suffered and still suffers.Two more experimental conditions were induced, in which the participants were accused of discriminating, either individually or collectively, against an ad hoc out-group.This was done to control the moral conflict arising from the new ethical relationship that governs perpetrating majority-victimized minority relationship, and not from the simple social undesirability and threat to social identity that the accusation of being a discriminator entails.(Faul, Erdfelder, Buchner, & Lang, 2009) shows that the sample of 131 participants provided the power of .80 (α = .05)todetect a small to moderate effect size of f = 0.25.Preliminary analyses indicated that neither the sex nor the age of participants produced a main effect or interaction with the other variables.

Ethical Statement
The study was guided by the Declaration of Helsinki's standards and ethics guidelines of University of Valencia.All participants read an informed consent statement and agreed to participate in the study.Their participation was anonymous and voluntary, and they did not receive any remuneration.The materials and data that support the findings of this study are public accessible via Open Science Framework at https://osf.io/ds3yc/Procedure The experiment was carried out in two sessions, one day apart.The first session was only there to be able to give the bogus feedback in the second day.In this session only the standard Minimal Group Paradigm matrices test was applied (Tajfel, Billig, Bundy, & Flament, 1971).This test was completed by 149 participants.In the second session 131 participants were involved, with a dropout of 18 overall cases (N = 7 in the Gypsy Minority condition and N = 11 in the Ad-hoc Group condition, see design of independent variables below).The second session began by informing participants of the supposed "results" of the previous day.This allowed manipulation of the independent variable, collective vs. individual accusation of discrimination, to be introduced.The participants then read a text (composed of 583 words), entitled The Gypsies: A historically persecuted minority.This text recalled the persecution suffered for five centuries by the Gypsies throughout their history in Spain.The text confirmed that the situation of marginalization in which Gypsies still live is a consequence of the discrimination to which they are subjected by the Payos (i.e.Spaniards, "Payos" in Spanish).The measurements of the dependent variables were taken, and at the end of the experiment a detailed debriefing was carried out.

Design and Manipulation of the Independent Variables
The experiment used a 2 (target of discrimination: Gypsy minority vs. ad hoc out-group) x 2 (accusation of having discriminated against the Gypsy minority, or the ad hoc out-group, at collective vs. individual level) factorial design.
Target of discrimination: Gypsy minority vs. ad hoc out-group.In the matrices test, participants had to distribute points between one person in their group and one person from another group.They were told that they had been divided into two groups "for administrative purposes of the study".In the Gypsy minority condition, points were divided between one person from the majority Payos group (to which all the participants belonged) and one person from the Gypsies minority group.In the ad hoc groups condition, they had to divide the points between one person in their Alpha group (to which all the participants were assigned, without them knowing), and one person from the out-group, the Omega group.
Collective vs. individual level of discrimination.The second independent variable was manipulated in the second session, a day after carrying out the matrices test.It was explained to all participants that the matrices they had completed the day before measured three intergroup behaviors: in-group favoritism ("giving more points to a person from their own group than to a person from another group"), parity ("treating the two groups equally, giving the same amount of points to the two people") and discrimination against the other group ("leaving the other group with the lowest number of possible points, even if it involves a cost in absolute terms for their own group").Those behaviors were illustrated with one of the matrices they had previously seen.
After this explanation, it was reported in the condition of collective level of discrimination, 85% of the responses of the people in their group had been discriminative against the Gypsy minority or against the Omega out-group, depending on the experimental condition.In the individual level of discrimination condition, each participant privately received their personal result and, by doing so, was informed that 85% of their responses on the matrices test had been discriminative against the Gypsy minority or the Omega out-group, depending on the experimental condition.In this condition, participants did not receive any information on the behavior of their group.The feedback provided to participants was independent of their actual responses.In order to ensure that the feedback they received seemed realistic to them, they were explained that all the 9 matrices had been subjected to an overall None of the participants expressed any doubt or suspicion at all about such explanation.

Dependent Variables
Estimation of attitudes of the majority towards the Gypsies.Participants had to estimate the social distance that they thought the majority (Payos) usually adopted regarding Gypsies in a 10-item test developed by us.For example, some items in the test were: "For most people, having a Gypsy friend is frowned upon"; "Most people do not want to have a Gypsy family as a neighbor"; and, "For most people, a Payo who marries a Gypsy drops down in status".Participants answered each item on a scale of 1 (totally disagree) to 7 (totally agree).An index was created grouping the ten items (M= 4.79; SD = 1.31;Cronbach's alpha = .93.).A high score indicated a negative perception of the majority´s attitudes towards Gypsies.
Recognition of Gypsies as historical victims.A list of 15 minority groups was presented (see the list in the results section below), which included the Gypsy minority.The task was explained in the following manner: "Here is a list of minorities, presented in alphabetical order.All of them have been historical victims.What you are asked to do is to order them according to which groups you estimate are more or less deserving of being recognized as historical victims; that is, that their current bad situation is explained by the mistreatment and discrimination they have suffered in the past, due to a perennial stigma ascribed to them".
Collective guilt for the persecution of Gypsies.This scale was adapted from the Collective Guilt Scale from Branscombe, Slugoski and Kappen ( 2004 Participants answered these items on a scale of 1 (completely agree) to 9 (completely disagree).An index was created grouping all five items (M = 4.38, SD = 1.31), so that a low score indicated a greater sense of the participant´s guilt, or internal conflict, with respect to the Gypsies' fate.Cronbach's alpha for this scale was 0.57.Although the internal consistency of this scale is low, significant outliers were not observed.A principal components analysis shown that the five items loaded in a same factor (loadings greater than .42).
Attitudes towards the compensation of Gypsies.This scale is composed of eight items.The items measure attitudes to positive discrimination and the compensation of Gypsies for the historical discrimination to which they have been subjected.It includes items such as: a) We have to grant special rights to Gypsies to erase that horrible past from memory; b) At least 2% of management positions in public institutions must be reserved for Gypsies; c) The state has to compensate Gypsies for their persecution in the past; and, d) It is necessary to provide more financial aid for Gypsies.These items were scored on a scale of 1 (completely agree) to 9 (completely disagree).An index was created, grouping the eight items.A low score indicated more favorable attitudes towards the compensation of Gypsies (M = 6.12,SD = 1.52).Cronbach's alpha for this scale was .82.
Victimization of Gypsies for their marginalization.This scale was taken from Pérez and Mugny (1993) and is composed of the following five items ).An important element of this scale for the victimization of Gypsies due to their marginalization is to see the minority as an entity isolated from the majority of society and disunited from the human community, and to reduce the explanation to one implying their fault, that is, to attribute their separation to the idea that they do not want to profess the values of the majority mainstream.Items on this scale were directly aimed at measuring to the degree to which, compared to the Payos, Gypsies are represented as having less desire to improve and less interest in values such as schooling, technology, politics, or environmental protection.These are values that perform as cultural truisms (McGuire, 1964) of prosperity for the majority.The more participants insist that the Gypsies are unconcerned about these mainstream values, the greater the victimization of the Gypsies due to their marginalization (i.e.Gypsies are blamed more for their own situation of marginalization).

Preliminary Analysis
The results of the matrices test were analyzed following the guidelines of Bourhis, Sachdev and Gagnon (1994).The  1971) were used merely as a procedure to operationalize the "accusation" that the participants had discriminated (individually vs. collectively) against either a historical minority (Gypsies) or an ad hoc out-group.In any case, a one-way analysis of variance, according to the type of categorization (historical groups Payos/Gypsies vs. ad hoc groups Alpha/Omega) for each of the intergroup behaviors measured in the matrices showed that there were no significant differences (all Fs < 1) in any of the intergroup behaviors measured by matrices (maximum differentiation, maximum joint profit, maximum ingroup profit, in-group favoritism and parity).Table 1 shows the means of these behaviors.Despite the absence of differences, and as far as the manipulation of discriminating accusations can acquire one meaning or another, depending on the real behavior of the participants in these matrices, these intergroup behaviors of the individual were introduced as covariates in all the analyses that followed.Even though the results were similar when they were not introduced as covariates, its degree of relationship with the different measures helps calibrate the intergroup meaning of each scale.Note.The higher the score, the more strength in the first term of the two opposing intergroup behaviors.SD in parentheses.MD = maximum differentiation; MJP = maximum joint profit; MIP = maximum in-group profit; FAV = in-group favoritism; and, P = parity

Statistical Analysis
Each dependent variable was analyzed through an ANCOVA, 2 Target of discrimination (Gypsies vs. ad hoc out-group) x 2 Level of discrimination (collective vs. individual), with the six intergroup behaviors in Table 1 as covariates.Although the results are practically the same without introducing intergroup behaviours as covariates, their presence in the analyses is a criterion of validity of the intergroup significance of each dependent variable.

Estimation of the attitudes of the majority towards Gypsies.
The ANCOVA for this measure showed that participants perceived the majority as having more negative attitudes towards Gypsies in the condition of collective level of discrimination (M = 5.01, SD = 1.28) than in the condition of individual level of discrimination (M = 4.54, SD = 1.32,F(1, 120) = 6.06, p = .015,η 2 = .048).It is also observed that in the condition in which the target of discrimination was the Gypsy minority, they estimated that the majority of people have more negative attitudes towards Gypsies (M = 5.05, SD = 1.22) than when the target of discrimination was the ad hoc out-group (M = 4.52, SD = 1.37,F(1, 120) = 5.81, p = .017,η 2 = .046).None of the intergroup behaviors in the matrices test had an effect on this measure ps > .25.

Recognition of Gypsies as historical victims.
After reading the extract on the history of the persecution suffered by the Gypsies, participants had to determine the ranking of 15 minorities according to which minorities they considered the most (score 1) to the least (score 15) worthy of being recognized as a historical victim.The average ranking (see figure 1) turned out to be the following: black people (M = 4.34), Jews (M = 5.09), women (M = 5.18), homosexuals (M = 5.56), immigrants (5,82) ).A t-test shows that the eighth position, in which the Gypsies were ranked, differs from both adjacent minorities (Moors, t/130 = 2.006, p = .047;and AIDS patients, t/129 = -1.985,p = .049).Overall, these results confirm the resistance to recognizing Gypsies as more significant historical victims than other six ethnic and social minorities included in the ranking (black people, Jews, women, homosexuals, indigenous people and Moors), who were all of them recognized as more significant victims than Gypsies.The ANCOVA on the ranking assigned to the Gypsies showed, firstly, that the covariate intergroup strategy of social differentiation in the matrices was significantly related to the recognition of the Gypsies as victims (Beta = .22,F(1, 120) = 5.75, p = .018for MD vs MJP+MIP; and Beta = -.29,F(1, 120) = 10.48,p = .002for MJP+MIP vs MD).Both effects indicate that the greater the intergroup strategy of social differentiation in the matrices test, the smaller the recognition of Gypsies as a victimized minority.In addition, the interaction of the two manipulated independent variables was significant (F(1, 120) = 3.97, p = .049,η 2 = .032).As can be seen in Figure 2, under the condition of collective level of discrimination against Gypsies, participants tended to recognize the minority as a victim (M adj = 6.93,SE = .50)more than in the condition of individual level of discrimination against Gypsies (M adj = 8.15, SE = .54,p = .099,95% CI [-2.68, 0.23]).The participant levels of discrimination against the Omega outgroup, either collectively (M adj = 7.51, SE = .52)or individually (M adj = 6.59,SE = .58),did not vary in their recognition of Gypsies as victims (p > .24).The planned contrast also revealed that the difference between the two conditions of individual level of discrimination was significant, p = .05,95% CI [-3.12, -0.002]; that is, the individual level of having discriminated against the Gypsies showed less recognition of the Gypsies as victims.No other contrast was significant.

Attitudes towards the compensation of Gypsies.
The ANOVA on attitudes towards the compensation of Gypsies showed that the covariate intergroup strategy of maximum joint gain was significantly related to the compensation of Gypsies (β = -.08,F(1, 120) = 8.75, p =. 004, for MJP+MIP vs MD): the more the maximum joint gain was chosen in the matrices, the more the participants were in favor of compensating the Gypsies.A main effect was also observed where the participants showed to be more in favor of compensating Gypsies in the condition of collective level of discrimination (M adj = 5.84, SE = .17)than in the condition of individual level of discrimination (M adj = 6.45,SE = .19,F(1, 120) = 5.27, p = .023,η 2 = .042)No other difference in attitudes towards the compensation of Gypsies was significant (ps > .45).

Victimization of Gypsies for their marginalization.
The covariate intergroup strategy of maximum differentiation was shown to be related to attitudes towards the victimization of Gypsies for their marginalization (β = -.07,F(1, 120) = 4.21, p = .042,for MD vs MJP+MIP): the more the intergroup strategy of maximum differentiation was chosen, the more the participants attributed blame to Gypsies for their own situation.The interaction between the two manipulated independent variables was also significant, F(1, 120) = 5.41, p = .022,η 2 = .043.As can be seen in Figure 4, pairwise comparisons showed that the participants in collective level of discrimination against Gypsies attributed less victimization to them for their marginalization (M adj = 4.86, SE = .

DISCUSSION
The results have confirmed the discrimination-level hypothesis of perpetrating majority-victimized minority.They have shown that the de facto collective level having discriminated against Gypsies, compared to having discriminated against them at the individual level, increased the recognition of Gypsies as victims and gave rise to a greater sense of collective guilt.These two variables are related to the most favorable attitudes towards the compensation of Gypsies for the historical mistreatment to which they have been subjected.The contrasts indicate that these effects are significant, especially when contrasted with the individual level of having discriminated against the minority.A belief in the existence of a collective racism increases the recognition of Gypsies as victims.These responses are not due to a mere strategy of improving the moral identity of the in-group questioned by the experimental manipulation, since the conditions in which the participants were at collective level informed of an unfair treatment to the out-group did not influence the attitudes towards Gypsies.It was the collective level of the majority for having discriminated against the minority, and not the collective level of the in-group for having discriminated against an ad hoc out-group, that led to a representation of Gypsies as victims.
The second hypothesis concerned the transformation of the representation of the minority from deviant to victim: highlighting the existence of a collective discrimination against the minority should cause a transformation of the representation of a deviant minority into a victimized minority -less victimization of Gypsies for their own situation of marginalization.The results confirmed this hypothesis, since in the condition of collective level of discriminating against Gypsies participants placed a lesser degree of guilt on Gypsies themselves for their marginalization, than in the condition of individual level of discriminating against Gypsies, or in the condition of collective discrimination against the out-group.Therefore, the social representation of the causes of the marginalization of the minority is transformed in the condition of collective level of discrimination, and from seeing the minority as deviant, it begins to be seen as a victim.
A particular theoretical concern of our experiment was how, or at what level, collective guilt is resolved.As observed in several studies, collective guilt increases the desire to apologize and to provide restitution or compensation to the victimized minority (e.g.Mallett & Swim, 2004;Miron et al., 2011;Stewart et al., 2010; for a review, see Wohl et al., 2006).Even though contrition and penitence generate a good conscience, do they, however, induce a new representation of the object that incites the majority to sin?Our question is whether, in addition to these manifest attitudes, the social representation of the minority is transformed.A basic component of collective innocence is victimization, that is, to believe that the minority has a set of essential characteristics that incite or justify their own discrimination.
There are different ways to measure victimization.In our experiment, it is assumed that an important element of the victimization of Gypsies for their marginalization is to see the minority not only as an isolated entity of the majority of society, disunited from the human collective, but also to reduce the explanation to sua culpa, that is, to attribute this separation to the minorities themselves who profess marginal values for the societal mainstream.The measure of the victimization of Gypsies for their marginalization in these experiments was directly aimed at measuring the degree to which, compared to the Payos, Gypsies are represented as having less desire to overcome and less interest in facets of life such as schooling, technology, policy, or environmental protection.Representing Gypsies as having a lower degree of interest in these issues means seeing them differently in topics that are fundamental for the majority.No doubt these measures should be improved, and others should be developed in future research.A further important limitation was the low reliability for collective guilt scale.This scale was adapted from the Collective Guilt Scale from Branscombe et al., (2004).Clearly more research is needed to find out if the guilt towards the Roma minority has differential nuances with respect to the guilt generated by other minorities.

CONCLUSION
The active struggle of minorities in the '50s and '60s for civil rights and for the extension of democratic principles to everyone equally, have ended up transforming the relationships between the majority and minorities.Step by step a political relationship has passed on to an ethical relationship, i.e., a moral perspective from which the majority regards their own behaviors towards social minorities.This resulted in an immorality judgment of discriminatory attitudes and behaviors that had long been regarded as natural.In the '80s a new moral representation of discriminated minorities, a new category of minoritiesvictimized minoritieshas appeared.Minorities have ceased to be seen as mere deviants and have come to be seen as victims of majority practices.In short, it made possible for the majority to move from collective innocence to collective guilt.
carried out with 131 students (70.2% women) from a University in Spain (age M = 20.26,SD = 3.64).A G* power analysis construct an algorithm that supported the inference about the (collective or individual) discrimination behavior.
), and contained the following five items: (a) I feel guilty about the harmful actions of Payos (i.e.Spaniards) towards Gypsies; (b) I feel guilty about the negative things other Payos have done to Gypsies; (c) I feel regret for some of the things Payos have done to Gypsies; (d) I might feel guilty about bad things done to Gypsies by the Payos; and, (e) I feel compassion for the Gypsies (reverse scored).

Figure 1 .
Figure 1.Ranking of 15 minorities considered the most (score 1) to the least (score 15) worthy of being recognized as a historical victim

Figure 2 .
Figure 2. Recognition of Gypsies as victims (higher scores show less recognition of Gypsies as a victimized minority)

Table 1 :
Strength of each intergroup behavior according to the type of categorization.