Seeking revenge or seeking reconciliation how concern for social image and felt shame helps explain responses in reciprocal intergroup conflict european journal of social psychology

As such, members of the involved groups in or after a reciprocal conflict can focus on their group as a perpetrator or as a victim Mazziotta, Feuchte, Gausel, & Nadler, 2014; SimanTov-Nac

R E S E A R C H A R T I C L E

Seeking revenge or seeking reconciliation? How concern for social-image and felt shame helps explain responses in reciprocal

intergroup conflict

Nicolay Gausel*, Colin Wayne Leach†, Agostino Mazziotta‡ & Friederike Feuchte§

* Department of Psychosocial Health, Faculty of Health and Sport Sciences, University of Agder, Grimstad, Norway

† Department of Psychology, University of Connecticut, Storrs, USA

‡ University of Hagen, Germany

§ Monrovia, Liberia

Correspondence

Nicolay Gausel, University of Agder,

Postboks 422, 4604 Kristiansand, Norway.

E-mail: nicolay.gausel@uia.no

Received: 16 February 2015

Accepted: 5 January 2017

http://dx.doi.org/10.1002/ejsp.2295

Keywords: shame, guilt, social-image,

revenge, reconciliation, victim, perpetrator

Friederike Feuchte, Independent

Researcher.

The raw data for this study is stored with the

GESIS Data Archive for the Social Sciences

with the doi: http://dx.doi.org/10.7802/1331

Abstract

In conflicts with reciprocal violence, individuals belong to a group that has been both perpetrator and victim In a field experiment in Liberia, West

Africa, we led participants (N = 146) to focus on their group as either

perpetrator or victim in order to investigate its effect on orientation towards inter-group reconciliation or revenge Compared to a perpetrator focus, a victim focus led to slightly more revenge orientation and moderately less reconciliation orientation The effect of the focus manipulation on revenge orientation was fully mediated, and reconciliation orientation partly mediated, by viewing the in-group’s social-image as at risk Independent of perpetrator or victim focus, shame (but not guilt) was a distinct explanation

of moderately more reconciliation orientation This is consistent with a growing body of work demonstrating the pro-social potential of shame Taken together, results suggest how groups in reciprocal conflict might be encouraged towards reconciliation and away from revenge by feeling shame for their wrongdoing and viewing their social-image as less at risk As victims and perpetrators are widely thought to have different orientations to inter-group reconciliation and revenge, we suggest that work on reciprocal conflicts should account for the fact that people can belong to a group that

has been both perpetrator and victim.

Many analyses of intergroup conflict distinguish

between perpetrators and victims It is often assumed

to be obvious which is which However, in the real

world, it is not always so clear For instance, in

reciprocal conflicts—in places like Syria, Egypt, the

Balkans, Northern Ireland, and Rwanda—each group

has perpetrated acts of violence against the other

group and suffered as victims of such acts (e.g., Brym

& Araj, 2006; Staub, 2006; Stevenson, Condor, &

Abell, 2007) As such, members of the involved

groups in or after a reciprocal conflict can focus on

their group as a perpetrator or as a victim (Mazziotta,

Feuchte, Gausel, & Nadler, 2014; SimanTov-Nachlieli

& Shnabel, 2014) To examine how focusing on one’s

group as a perpetrator or as a victim effect the

orientation towards reconciliation and revenge, we

led individuals to focus on their group as either

perpetrator or victim in a field experiment in

Liberia—where civil wars engulfed the society in

devastating reciprocal conflicts between 1989 and

2003 (Cain, 1999)

As much research on reconciliation assume that

perpetrators can be motivated to reconcile because it

can re-establish their moral standing (for reviews, see Nadler, 2012; Nadler & Shnabel, 2015), we examined whether a perpetrator focus led those in this reciprocal conflict to be more oriented to reconciliation As an angry, hostile desire for revenge is common in recipro-cal conflict, we also examined orientation to revenge Because of the reciprocal violence, victims have little reason to expect the adversary to identify as a perpetra-tor obliged to reconcile or make reparation (Brym & Araj, 2006; Carlsmith, Darley, & Robinson, 2002; Mikula, 1993; SimanTov-Nachlieli & Shnabel, 2014); thus, revenge should seem especially likely for those who focus on their group as a victim To explain these contrasting effects of perpetrator versus victim focus

on orientation to revenge and reconciliation, we examined Gausel and Leach’s (2011) notion of perceived risk to group social-image as a mediator A social-image at risk stokes fears of exclusion and isolation—the most serious of social threats (Gausel, Leach, Vignoles, & Brown, 2012; Gausel, Vignoles, & Leach, 2016) Thus, greater perceived risk to social-image should help explain orientation to revenge (more prevalent in victim-focus) and lesser perceived risk

should help explain orientation to reconciliation (more

prevalent in perpetrator-focus)

As important as perpetrator and victim focus are to

group member’s inter-group orientation, group-based

emotions such as shame about wrongdoing should

serve as additional explanations of reconciliation

orientation Consistent with a growing body of work

on the pro-social orientation of group-based shame

(for reviews, see Gausel & Leach, 2011) in contexts

where improvement is possible (Leach & Cidam,

2015), we expect shame to help explain orientation

to reconciliation in this reciprocal conflict As each

group is both perpetrator and victim, feeling ashamed

about the serious and substantial immorality

committed by one’s group should predict believing

that repair of the relationship is possible (Leach &

Cidam, 2015) as well as empathy, compassion, and

the otherwise pro-social orientation to the out-group

necessary to a reconciliation orientation (Gausel

et al., 2012, 2016)

Victim-Focus versus Perpetrator-Focus: Revenge

or Reconciliation?

A good deal of previous work on inter-group

reconcilia-tion has been grounded in the needs-based model

(Nadler, 2012; Nadler & Shnabel, 2008) Although the

basic tenets of the needs-based model are well

sup-ported by prior research (for reviews, see Lund, 2003;

Machakanja, 2010; Nadler, 2012; Nadler & Shnabel,

2015), the model assumes a clear-cut distinction

between perpetrator and victim Such a distinction is

unlikely in reciprocal conflict because the opposing

parties have been both perpetrator and victim (Brym

& Araj, 2006; Kanyangara, Rimé, Paez, & Yzerbyt,

2014; Stevenson et al., 2007) Because of the more fluid

perpetrator and victim positions in reciprocal conflict,

examinations of reconciliation in such contexts seem

to require additional explanations of the role of

perpetrator and victim positions (Mazziotta et al.,

2014; SimanTov-Nachlieli & Shnabel, 2014) This seems

especially important as a focus on the group as a victim

in reciprocal conflicts can encourage a hostile and

sometimes violent orientation towards revenge and

thereby perpetuate a vicious circle of violence (Brym

& Araj, 2006; Kanyangara et al., 2014; Nadler &

Shnabel, 2015)

Victim-Focus and Revenge

At its heart, viewing oneself as a victim means that one

views another party’s actions as causing lasting damage

to oneself (Baumeister, Stillwell, & Wotman, 1990;

Chaikin & Darley, 1973; Mikula, 1993) Although some

victims manage to forgive perpetrators (e.g., Bandura,

1999; Macaskill, Maltby, & Day, 2002) and to move

towards reconciliation, in many instances, victims are

angry, hostile, and oriented towards revenge (Brym &

Araj, 2006; Carlsmith et al., 2002; Mikula, 1993; Nadler

& Shnabel, 2015) Victims seems especially likely to

eschew reconciliation for revenge when they have little reason to expect their antagonists to view themselves as perpetrators who should apologize and thereby empower victims (see Bandura, 1999; Baumeister

et al., 1990; Scheff, 1994) This seems likely in cases of reciprocal conflict where perpetrators can legitimately also view themselves as victims (Mazziotta et al., 2014; Nadler & Shnabel, 2015; SimanTov-Nachlieli & Shnabel, 2014) When there is little chance of apology from those reluctant to identify as perpetrator, victims can experience the inter-group relation as a ‘stalemate’ that can be best resolved through taking revenge (see Brym & Araj, 2006; Gausel, 2013; Nadler & Shnabel, 2015; Retzinger, 1991; Scheff, 1994)

Thus, in the sort of reciprocal conflict that we exam-ine here, we expect individuals led to focus on their group as a victim, rather than a perpetrator, to be more oriented towards revenge We think that this is likely explained by victims viewing their social-image as someone to respect being put at risk of serious, perhaps irreparable, damage (Leach & Cidam, 2015) As Gausel and Leach (2011) discussed in their recent review, a social-image put at risk is the psychological experience where an individual is concerned about how others view them in terms of morality/immorality Hence, it

is the situation in which the individual fear that their all-important need to belong, to be respected, and to

be accepted by others may go unfulfilled if these others

find out about their immorality or the immorality in which they are associated (i.e., through their group-affiliation) For a victim, one example of this process can be that others think that the victim is responsible for their own suffering; that is, that the suffering is just and seen as a natural consequence of their self-caused situation (Gausel, 2013) Hence, a concern for social-image is the concern that others will condemn one, and thus find one to be unworthy of their respect and inclusion (Gausel & Leach, 2011) A social-image put

at risk for victims has long been known to orient people towards retaliatory hostility (e.g., Brym & Araj, 2006; Lewis, 1971; for a review, see Leary, Twenge, & Quinlivan, 2006), and sometimes even bloody revenge (Scheff, 1994)

Although this is the first time to present an explana-tion of why people who focus on themselves as victims are oriented towards revenge, some initial support can

be found in two experiments reported by Gausel et al (2016) Even though they did not examine the orienta-tion to revenge that we examine here, they showed in their experiments on social-image concerns that putting individuals in situations where others might learn of their serious wrongdoing led them to worry about being isolated, losing respect, and being rejected by others In these studies, and in parallel studies of group image (Gausel et al., 2012), perceived risk to social-image (i.e., the fear of losing others respect, and be isolated and rejected by them) was consistently linked

to anti-social, self-defensive motivation, such as covering up the wrong or wanting to avoid others Hence, in the case of reciprocal conflict that we examine

here, perceived risk to social-image should link with the

anti-sociality of being oriented to revenge against

per-ceived perpetrators

Perpetrator-Focus and Reconciliation

According to Mazziotta et al (2014), if people belong to

a group that has been both perpetrator and victim in a

reciprocal conflict, they prefer to construe themselves

as victims rather than as perpetrators (see also Nadler

& Shnabel, 2015; SimanTov-Nachlieli & Shnabel,

2014) This means that it may be difficult in reciprocal

conflicts to get the antagonists to focus on themselves

as perpetrators as we know that people tend to avoid

being labelled as perpetrators (for reviews, see Leach,

Zeineddine, & Čehajić-Clancy, 2013; Mikula, 1993;

Noor, Shnabel, Halabi, & Nadler, 2012), as perpetrators

are, by definition, immoral (Chaikin & Darley, 1973)

If, however, individuals in a reciprocal conflict can be

led to focus on their group as a perpetrator, then they

will implicitly also admit to immorality Admitting to

perpetration will likely increase orientation to reconcile

with the perceived victim (Mazziotta et al., 2014;

SimanTov-Nachlieli & Shnabel, 2014) because it will

serve to re-establish their morality and their standing

as moral actors who can be trusted to operate within

the community (Nadler & Shnabel, 2015) Put in terms

of Gausel and Leach’s (2011) thinking, a perpetrator

focus (i.e., an admittance of immorality) should lead to

a greater orientation to reconcile with victims because

a chance to restore their standing as moral actors will

be increased the more they feel empathy with the

vic-tims and the more they want to repair the relationship

that has been damaged by their immorality This means

that the perpetrators perceived risk to their group’s

social-image is diminished by their increased

willing-ness to restore their morality through moral

reconcilia-tory orientations

Hence, when perpetrators believe that their

social-image in the eyes of others is less at risk (and thus that

they are likely to be socially included and valued) effort

at reconciliation can in fact re-establish their moral

standing and cement their social belonging as suggested

by Gausel and Leach (2011) and Nadler (2012) In other

words, a relatively safe social-image enables those

focused on themselves as perpetrators to pursue

recon-ciliation in the belief that it is possible to improve their

relations with the victims (see Leach & Cidam, 2015)

Shame: A Distinct Explanation of Reconciliatory

Orientation

As important as perpetrator and victim focus and

perceived risk to social-image are to group member’s

orientation to inter-group reconciliation, there are also

other explanations As the groups in a reciprocal conflict

like the civil war in Liberia all took part in serious and

sustained immorality, members of the groups can all

feel the intense moral self-reproach of shame regardless

of their focus on their group’s role in the conflict (for

reviews, see Gausel & Leach, 2011; Tangney & Dearing, 2002) Thus, in a reciprocal conflict, group members can feel shame about their group even if they are not focused on the group as a perpetrator per se Given this, shame about the immorality done by one’s group in the reciprocal conflict should be a consistent basis of recon-ciliation orientation, independent of whether one is focused on one’s group as a perpetrator or a victim

In contrast to the view that shame is maladaptive and defensive (e.g., Nadler, 2012; Tangney & Dearing, 2002), many recent investigations of shame have found that shame for immorality motivates efforts to repair harm done (Berndsen & McGarty, 2012; Brown & Čehajić, 2008; Brown, Gonzalez, Zagefka, Manzi, & Čehajić, 2008; Gausel et al., 2016; Imhoff, Bilewicz, & Erb, 2012; Leach, Iyer, & Pedersen, 2006) and to object to ongoing immorality (Berndsen & Gausel, 2015) Already in 2006, Gausel found that felt shame for ingroup perpetration against minorities significantly predicted a desire to repair the damage done to them Building on this, Gausel et al (2012) found that shame felt by ethnic Norwegians for their country’s mistreat-ment of Norwegian Gypsies (‘Tatere’) predicted want-ing to repair this relationship by communicatwant-ing contrition to the victims and offering restitution And

a recent meta-analysis of studies of individual and group-based shame by Leach and Cidam (2015) established that shame has a moderate-sized link to pro-sociality and self-improvement in those circum-stances where the failure appears more reparable Thus, we expect that individual’s shame about their group’s immorality in the reciprocal conflict of the Libe-rian civil war will serve as a distinct explanation of rec-onciliatory orientation, whether or not they focus on their group as perpetrator or victim As each group is both perpetrator and victim, feeling shame about the harm caused by one’s group should be associated with believing that repair of the relationship is possible (Leach & Cidam, 2015) as well as empathy, compassion, and the otherwise pro-social orientation necessary to a reconciliation orientation towards the out-group (Gausel et al., 2012, 2016)

The Current Study

In order to test our expectations, we returned to a field experiment conducted in Liberia, Africa The republic

of Liberia has been plagued by 14 years of civil wars

Of the approximately 2.5 million people who lived in Liberia before the wars began in 1989, 85% were killed, internally displaced or became refugees (Cain, 1999) Thus, it is clear that most of the population has been profoundly affected by an extreme, long-standing, and reciprocal inter-group conflict

Parts of this large-scale study were reported in a manuscript by Mazziotta et al (2014), which focused

on how survivors (i.e., a person that has lived through the war) construe their group’s role and how this influ-ences their willingness to reconcile with the other parties of the conflict As this is a difficult to collect

sample of survivors of civil war, we found it legitimate to

include several different sets of measures in the study to

examine different research questions For the sake of

clarity, we illustrate how the measures are used across

the two manuscripts in Table 1

As outlined above, we expected participants who

were led to focus on their in-group as victim to be more

oriented to revenge than perpetrators In contrast, we

expected participants who were led to focus on their

in-group as perpetrator to be more oriented to

reconcil-iation than victims We also expected that victims would

be more concerned for the risk to the group’s

social-image than perpetrators, and that the concern for the

risk to the group’s social-image should increase the

orientation towards revenge and decrease the

orienta-tion towards reconciliaorienta-tion In addiorienta-tion, we expected

that shame about the harm done in the civil war should

serve as a distinct explanation of a reconciliatory

orientation However, as there is some debate that the

feeling of guilt can also motivate reconciliatory repair

(e.g., Brown et al., 2008; Imhoff et al., 2012; Leach

et al., 2006), we investigated both felt guilt and shame

As felt shame often outweighs the less profound and less

identity-relevant guilt (Berndsen & Gausel, 2015;

Gausel & Leach, 2011; Lewis, 1971), we expected felt

shame to be more related to reconciliatory orientation

than guilt

Method Participants and Procedure

The study was conducted in the outskirts of Monrovia,

the capital of Liberia, in collaboration with a local

NGO One hundred and forty-six Liberians (73 women,

6 participants did not indicate their gender; Mean age:

27.8, range: 16–51 years, 34 participants did not indicate

age) participated in exchange for a small monetary

compensation.1Participants were approached at

com-munity centres and at their homes by three trained

Liberian research assistants, and asked if they would like

to take part in a study in which questions about the war

and reconciliation would be addressed Participants

were informed that participation in the study was

entirely voluntary, that it was anonymous and that they

could stop filling out the questionnaire at any time

without disclosing their reasons and without any

disadvantage

In the first part of the questionnaire participants

in-dicated their age, then they were randomly assigned to

either a victim-focus or a perpetrator-focus condition

In the victim-focus (N = 69) condition, participants

were asked to: ‘Please take some time and think about

an episode during the war where people from your

‘tribe’ (a colloquial term for ethnic group) have been

harmed… ,’ and then write down an episode of harm

in which their own ethnic group was the victim In

the perpetrator-focus (N = 77) condition, participants

were asked to: ‘Please take some time and think about

an episode during the war where people from your

‘tribe’ have caused harm… ,’ and then write down an

episode of harm in which their own ethnic group was the perpetrator After this, participants answered

a series of questions presented with response scales

that ranged from 1 = ‘not at all’ to 7 = ‘very much’ At

the end of the study, all participants were thanked and thoroughly debriefed.2

Measures Revenge orientation As much research and theorizing on revenge orientation focuses on both hostile anger and the desire for revenge (e.g., Brym & Araj, 2006; Gausel, 2013; Mikula, 1993; Retzinger, 1991), we measured both We measured Revenge (α = 60) with two items: ‘Sometimes I wish for bad things to happen to that ‘tribe’ ’ and ‘That ‘tribe’ should pay for what they did.’ We measured Anger (α = 67) with three items adopted from Gausel et al (2012) only now directed towards an outgroup: ‘I feel angry with members of that ‘tribe’ ’, ‘I feel irritated with members

1 We did not decide on a fixed sample size in advance as we did not

know how many participants we could reach in Liberia However, as

we would deploy Structural Equation Modelling to analyse the data,

it was important for us to at least reach a suitable sample-size for our

somewhat complex structural regression model (i.e., at least five

partic-ipants per parameter; see Bentler & Chou, 1987).

2 As reported by Mazziotta et al (2014), participants did not differ in es-timation of collective victimization and collective perpetration before being assigned to the two conditions In support of the focus-manipulation, a content analysis conducted by Mazziotta et al (2014) yielded that the victim-focus participants wrote down significantly more episodes of victimization than did the perpetrator-focus condition participants, and the perpetrator-focus participants wrote down signifi-cantly more episodes of perpetration than did the perpetrator-focus condition participants This provided support to the manipulation of fo-cus For the interested reader, it may be worth noting that in the perpetrator-focus condition, participants tended to report fewer details about the episodes, and mentioned fewer episodes where themselves were involved as well as listing more reasons that led to the episodes (see Mazziotta et al., 2014) Even though it did not affect the manipula-tion, it can be interpreted as attempts to distance from the perpetrator label (Leach et al., 2013; Mazziotta et al., 2014; Mikula, 1993; Noor

et al., 2012).

Table 1 Illustration of the measures used

Variables

MS 1(published

in EJSP)

MS 2(current) Willingness to engage in cross-group

contact

x

of that ‘tribe’ ’ and ‘I am annoyed with members of

that ‘tribe’.’

Reconciliation orientation As an orientation

towards reconciliation can be interpreted as

hypocrit-ical without a sincere empathic orientation towards

the ‘other’ (e.g., Batson et al., 2003), we measured

Empathy (α = 76) with three items3: ‘I can

under-stand the point of view of people from other ‘tribes’ ’,

‘I feel sorry for people from other ‘tribes’ when they

are having problems’ and ‘When I see someone from

other ‘tribes’ treated unfairly, I feel pity for them.’

Relationship repair (α = 74) was measured with four

items: ‘One can repair the social bond between my

‘tribe’ and that ‘tribe’ ’, ‘The social connection

between us can be fixed’, ‘The relationship between

the ‘tribes’ has suffered: but the wounds can be

healed’ and ‘It is possible to fix the damaged social

connection we have.’

Shame and guilt Our measure of Shame

(α = 71) included three items from prior research

(Gausel et al., 2012, 2016; for a review, see Tangney

& Dearing, 2002): ‘When I think about what my

‘tribe’ has done, I feel disgraced’, ‘When I think about

what my ‘tribe’ has done, I feel ashamed’, and ‘When

I think about what my ‘tribe’ has done, I feel

humili-ated.’ The measure of Guilt (α = 64) consisted of

two items based in prior research (Gausel et al.,

2012, 2016; Leach et al., 2006): ‘I feel guilty when I

think about what my ‘tribe’ has done’ and ‘I feel guilty

because of what my ‘tribe’ did.’

Risk to image The measure Risk to

social-image (α = 64) was based on the measure of Gausel

et al (2012, 2016) and included the items, ‘Other

‘tribes’ might not have the same respect for my tribe

because of this’ and ‘I think my ‘tribe’ could be

isolated from the moral community because of this’.

Gausel et al (2012, 2016) found perceived risk to

social-image to be indicative of a concern for

condem-nation by others and feelings of rejection and

isolation by others Thus, it is conceptually similar to

a fear of exclusion and the loss of social bonds (Gausel & Leach, 2011)

Results Confirmatory Factor Analysis—The Measurement Model

We used AMOS 22 to test our hypothesized measure-ment model of shame, guilt and risk to social-image in

a Confirmatory Factor Analysis Adopting a conserva-tive approach, we did not allow any items to cross-load on the factors and we did not allow the error terms to be correlated, except from the three latent factors that were allowed to correlate As expected, our measurement model fit the data very well (see Kline, 2005): χ2 (11) = 7.22, p = 78, χ 2 /df = 0.66, IFI > 1, CFI = 1, RMSEA = 000, AIC = 55.22) As can be seen in

Figure 1, our seven items loaded uniquely and strongly onto their expected factors (standardized λ’s ≥ 59; all

p’s < 001) indicating that each factor was well defined

by its items Despite the fact that latent variables nor-mally carry with them higher correlations due to lack

of unreliability, the three correlations between the

latent factors ranged from 13, p = 28 to 34, p = 012 and 61, p < 001, confirming that our measurement

model specified three different constructs

Other possible models Our measurement model proved to be superior to four other possible models First, our measurement model fit better than a model that col-lapsed shame and guilt into one factor where risk to social-image made up the other factor, χ2(2) = 56.18,

p < 001 Second, it fit better than a model that collapsed

shame and risk to social-image into one factor and guilt as

a separate, second factor, χ2 (2) = 51.1, p < 001 Third, it

fit better than a model that collapsed guilt and risk to social-image into one factor and had shame as a separate factor, χ2 (2) = 13.61, p = 001 Finally, it fit better than a

single-factor model where all items loaded onto one big

‘unpleasant feeling’ factor, χ2 (3) = 87.24, p < 001 In

conclusion, we could rest assured that our measurement model was superior to these other possible models that did not distinguish among our three hypothesized constructs

Possible cross-loading of items Even though one should be very cautious about interpreting the meaning of a single-item (Gausel & Salthe, 2014)— especially when it is removed from a uni-dimensional (and multi item) scale (Carmines & Zeller, 1979)— some readers might wonder whether the word

‘humiliation’ belongs more to the risk of social-image, rather than to the shame, scale Thus, we allowed the item ‘humiliated’ to cross-load onto the risk to social-image factor in an alternative model This did not improve model fit, χ2 (1) = 0.09, p = 76, and it

demonstrated that the ‘humiliated’ item loaded extremely weakly (standardized λ = .03, p = 76)

on the risk to social-image factor In addition, some

3 The Mazziotta et al (2014) paper used a 2-item measure that

investi-gated a strict affective-empathy: ‘I feel sorry for people from other ‘tribes’

when they are having problems’ and ‘When I see someone from other

‘tribes’ treated unfairly, I feel pity for them.’ The current paper, however,

wanted to investigate a more perspective-oriented type of empathy

with a third item that would go well with a reconciliatory orientation:

‘I can understand the point of view of people from other ‘tribes’.’ For

the sake of clarity: The results (including the pattern of regressions)

and the conclusions did not vary depending on the operationalization

of empathy in neither paper when we re-ran all the main analyses for

this current paper and the Mazziotta et al (2014) paper In fact,

intro-ducing this third perspective item made our model fit slightly worse

on all the fit indices Fit using the current measure of empathy (as

re-ported in the ms) that includes the third cognitive item: χ2

(9) = 17.53, p = 041, χ 2 /df = 1.95, IFI = 960, CFI = 954, RMSEA = 081.

Fit using a 2-item measure of empathy (identical to Mazziotta et al.,

2014): χ2 (9) = 15.37, p = 081, χ 2 /df = 1.71, IFI = 970, CFI = 966,

RMSEA = 07.

might argue that shame is about loss of respect and

that the single-item ‘loss of respect’ might cross-load

onto the shame factor Hence, we allowed the item

‘loss of respect’ to cross-load onto the shame factor in

an alternative model This did not improve fit, χ2

(1) = 0.54, p = 46, but it demonstrated that the ‘loss

of respect’ item was unrelated to the shame factor

(standardized λ = 07, p = 44) Hence, our alternative

cross-loading models demonstrated very clear support

for our hypothesized measurement model

Experimental Effects

The descriptive statistics and intercorrelations of all

measures are shown in Table 2

Risk to social-image An analysis of variance

(ANOVA) demonstrated that the focus manipulation

affected perceived risk to social-image, F (1, 142) = 5.31, p = 023,partialη2= 04.4As expected, perceived risk

to social-image was significantly higher in the victim

(M = 3.68, SD = 1.37) than the perpetrator condition (M = 3.09, SD = 1.67).

Reconciliation orientation A multivariate ANOVA (MANOVA) on empathy and relationship repair showed our manipulation had a multivariate effect

on these two measures of reconciliation orientation,

F(2, 138) = 15.91, p < 001,partialη2= 19 More specifi-cally, the focus manipulation had a significant effect on participant’s empathy with other ethnic groups,

F (1, 139) = 30.68, p < 001,partialη2= 18, as empathy

was significantly higher in the perpetrator (M = 5.13,

4There was no multivariate effect on guilt or shame, F (2, 136) = 1.33,

p = 27,partial η2< 02.

Fig 1: Confirmatory Factor Analysis, measurement mode Solid lines indicate significant relationships, p < 05

Table 2 Scale inter-correlations and descriptive statistics

1 Victims/perpetrators (+/ )

Note: N = 146 Higher Mean scores indicate higher levels of agreement (range = 1–7).

SD = 1.56) than in the victim condition (M = 3.66,

SD = 1.59) Also, as expected, the manipulation

had a significant effect on relationship repair,

F (1, 139) = 15.79, p < 001,partialη2= 10, as relationship

repair was higher in the perpetrator (M = 5.01, SD = 1.57)

than the victim condition (M = 4.06, SD = 1.07).

Revenge orientation A MANOVA on anger and

revenge showed that our manipulation had a

multivariate effect on these two measures of revenge

orientation, F (2, 139) = 4.15, p = 018, partialη2= 06

The focus manipulation had a significant univariate

effect on participants’ orientation towards revenge,

F (1, 140) = 8.37, p = 004,partialη2= 06, as the revenge

orientation was significantly higher (p = 004) in the

victim (M = 3.90, SD = 1.48) than in the perpetrator

condition (M = 3.10, SD = 1.77) However, there was

no significant effect on anger, F (1, 140) = 86, p = 36,

partialη2 < .01, although victims’ anger (M = 3.96,

SD = 1.29) was higher than perpetrators’ anger

(M = 3.72, SD = 1.70).

Structural Equation Modelling

To examine the proposed explanations of revenge and

reconciliation orientations, we specified a ‘structural

regression model’ (see Kline, 2005) using Structural

Equation Modelling with AMOS 22 software This

‘structural regression model’ (also known as a ‘hybrid

model’) approach has the benefit that it allows for fewer

variables, which in turn gives more statistical power as

one has more participants per parameter (i.e., it allows

for a better representation of the participants’

responses) Randomly missing data (Rubin, 1976; see Kaplan, 2009) were imputed with mean levels in order

to enable bootstrapping analyses (10 000 resamples, see Preacher & Hayes, 2008) As our study has two conditions (or categorical variables), we used effect coding (victim condition = +1, and perpetrator condition = 1) to trace the main effects of the experimental conditions on our dependent measures (Rutherford, 2001) We also specified perceived risk

to social-image as a manifest variable that could predict the latent factor of revenge orientation (consisting of two positively correlated composite

indicators; anger and revenge, r = 32, p < 05) and

the latent factor of reconciliation orientation (consisting of two positively correlated composite

indicators; empathy and relationship repair, r = 60,

p < 05) simultaneously Hence, our ‘structural

regression model’ allowed the four outcome manifest variables to indicate a theoretically sound latent factor each corresponding to a specific motivation that is allowed to be predicted by a causal path model Despite a significant chi-squared, our hypothesized model, χ2 (9) = 17.13, p = 047, χ 2 /df = 1.90, fit the

data very well (Kline, 2005) as underlined by several

other fit indices (IFI = 962, CFI = 958, RMSEA = 079,

AIC = 87.13).

In the upper half of Figure 2, one can see that risk to social-image was a significant, positive, predictor of

revenge orientation, β = 54, p = 003 In fact, risk to

social-image fully mediated the link between victim-focus and revenge orientation (Standardized

Point Estimate (SPE) = 115, SE = 07, Unstandardized Point Estimate (USPE) = 051, SE = 04, p = 035, 95% BC CI 001, 140), explaining as much as 33%

Fig 2: Structural regression model, victim and perpetrator conditions predicting feelings and orientations Solid lines indicate significant relationships, p < 05

of its variance.5 Also, in the upper half of Figure 2,

and in line with our expectations, risk to social-image

was a significant, negative, predictor of reconciliation

orientation, β = .26, p = 006 (SPE = 018, SE = 042,

.044).6Hence, perceived risk to social-image increased

orientation towards revenge and decreased orientation

towards reconciliation

As we wanted to investigate the distinct explanations

of shame and guilt on reconciliation orientation, we also

specified shame and guilt as manifest variables In the

lower half of Figure 2, one can see, as expected, that felt

shame was an additional positive predictor of

reconcili-ation orientreconcili-ation, β = 32, p < 001 Also, as expected,

felt shame was unrelated to revenge orientation,

β= .02, p = 82 In line with our expectations, felt guilt

was outweighed by felt shame, as it was unrelated to

reconciliation orientation, β = .06, p = 48, and to

revenge orientation, β = .16, p = 18.7In sum, shame

served as a distinct explanation of orientation towards

reconciliation that was unrelated to perpetrator- or

victim-focus, β = 08, p = 36.

Discussion The Measurement Model

As hypothesized, the Confirmatory Factor Analysis

demonstrated that our measures (of guilt, shame and

perceived risk to social-image) were distinct Indeed,

our measurement model was superior to six different

alternatives For instance, our alternative cross-loading

models showed that single items could not be removed

from their hypothesized scale without distorting the

meaning of the scale (see Carmines & Zeller, 1979;

Gausel et al., 2016) The validation of our measures

enabled to maximize degrees of freedom, and thus

statistical power, in our structural regression model by

treating the key variables as manifest rather than latent

variables

Perpetrator versus Victim Focus

Participants who focused on their group as a perpetrator

of harm expressed significantly more empathy with

other ethnic groups that had suffered during the war,

than did those in the victim-focus condition And they

were significantly more inclined to believe that repair

of the social bond with the other groups was possible

These findings are consistent with recent research showing that individuals can respond pro-socially to the perpetration of immorality by their group (Berndsen & Gausel, 2015; Berndsen & McGarty, 2012; Brown & Čehajić, 2008; Gausel & Brown, 2012), and they are consistent with Nadler and Shnabel’s (2008) theorizing that perpetrators often desire moral redress and forgiveness perhaps in an effort

to reestablish the group as a moral actor deserving of inclusion in the community (Nadler, 2012; see also Gausel & Leach, 2011)

In contrast, participants encouraged to have a victim-focus expressed significantly less empathy with other ethnic groups and were significantly less inclined to believe that the relationships could be repaired Our results also support the notion that victims tend to have less concern for the suffering of others (Retzinger, 1991; Scheff, 1994) and less desire for reconciliation when they focus mainly on their own suffering as victims (Brym & Araj, 2006; Mazziotta et al., 2014; Noor, Brown, Gonzalez, Manzi, & Lewis, 2008; SimanTov-Nachlieli & Shnabel, 2014) Our results also support Nadler’s (Nadler, 2012; see also Nadler & Shnabel, 2015) theorizing that the victim role is a complicated one in reciprocal conflicts

In support of the idea that identifying as a victim can lead to anti-social motivations (e.g., Brym & Araj, 2006; Mikula, 1993; Wohl & Branscombe, 2008), enabling a victim-focus made participants express greater orientation towards revenge against the perceived perpetrators As such, our results lends support to Brym and Araj (2006), Kanyangara et al (2014) and Retzinger’s (1991) arguments that victims

in reciprocal conflict can seek revenge Also, in a replication and extension of the findings of Mazziotta

et al (2014), SimanTov-Nachlieli and Shnabel (2014) showed that in the context of the intractable Israeli-Palestinian conflict, viewing one’s group as a victim promoted anti-social tendencies towards the adversary Our results are also consistent with the idea that victims may use revenge in an attempt to resolve

a ‘stalemate’ and to recover the social-image put at risk

by victimization (see Brym & Araj, 2006; Mikula, 1993) Risk to social-image In line with our expectations, individuals with a victim-focus had greater concern that their social image was at risk This finding supports Bandura (1999) and Brym and Araj’s (2006) argumen-tation that victimization is experienced as a painful loss

of social status (see also Baumeister et al., 1990; Mikula, 1993) that can severely damage social-image in the eyes

of others Hence, as expected, victim’s revenge orienta-tion was based in the concern that their group’s social-image was at risk This is consistent with Gausel and Leach’s (2011) argument that perceived damage to a group’s social-image leads to anti-social motivation, including hostile lashing out (see also Scheff, 1994) In this way, for a victim, being disrespected and isolated seems a bigger threat to their social standing than their own immoral perpetration Seeking revenge may

5When analysed in isolation from shame and guilt: SPE = 089, SE = 06,

USPE = 040, SE = 033, p = 021, 95% BC CI 001, 120.

6 Without accounting for guilt and shame, risk to social-image served as

a partial explanation for the lesser reconciliation orientation

(SPE = 050, SE = 031, USPE = 034, SE = 024, p = 023, 95% BC

CI .095, 003).

7 Separate analyses of the victim-focus condition and the

perpetrator-focus condition seem to imply that shame is more predictive of

reconcil-iation for perpetrators than for victims, but due to the low statistical

power offered in these separate analyses, we refrain from providing a

certain conclusion about this

therefore be an attempt to ‘settle the score’ with the

perpetrator by making them suffer a similar kind of

victimization (for a discussion, see Brym & Araj, 2006;

Retzinger, 1991), and by such, force the perpetrators

to respect them

Shame As expected, felt shame (but not guilt)

proved to be a distinct explanation of reconciliation

orientation This goes against the more traditional view

that shame is a maladaptive emotion that leads to

defensiveness and avoidance (e.g., Brown et al., 2008;

Nadler, 2012; Tangney & Dearing, 2002) Instead, it

supports recent theorizing (Gausel & Leach, 2011) and

nuanced empirical investigations of shame (Gausel

et al., 2012, 2016) that explain that shame can predict

pro-sociality, especially when moral or other failure is

viewed as reparable (Leach & Cidam, 2015) Thus, there

is a growing body of evidence that felt shame is related

to the motivation to object to ongoing immorality

(Berndsen & Gausel, 2015), to protest in-group

immo-rality (Shepherd, Spears, & Manstead, 2013) and a

contrite desire to repair and compensate victims

(Gausel, 2006; Gausel et al., 2012, 2016) In sum, the

results offer important support, in an under-studied

context, that group-based shame predicts reconciliation

As shame is an intense form of moral self-reproach

(Gausel & Leach, 2011), shame should motivate efforts

to repair the self and social-relations whether the group

is viewed as perpetrator or victim In this way, shame

provides a more constant basis of motivation to

recon-cile in reciprocal conflict or in other instances of moral

wrongdoing

Possible Limitations

One possible limitation in our study is the lack of a

control condition that did not focus individuals on their

group as victim or perpetrator In theory, this could

have helped us investigate whether the focus

manipula-tion increased or decreased revenge or reconciliamanipula-tion

orientation from a presumed baseline of no perpetrator

or victim focus However, in the context of the

recipro-cal conflict that we examined, the groups were both

perpetrator and victim In this civil war, as in most, the

uninvolved third party position was a luxury not

afforded to any Thus, a control condition of no victim

or perpetrator focus would have had little ecological

validity in this context More pragmatically, previous

research and theory strongly suggests that those in a

reciprocal conflict prefer to focus on themselves as

victims rather than perpetrators (Mazziotta et al.,

2014) As such, a no-focus control condition would

have likely been a reproduction of the victim-focus

condition without the experimental control and

internal validity afforded by the victim-focus condition

Another possible limitation might be that our study

asked participants to recall the experience of

perpetra-tion and victimizaperpetra-tion This can potentially be different

from a situation where an external source (e.g., a

ther-apist) is the reminder One can imagine that a helper is

able to aid the victim come to terms with the victimiza-tion (or perpetravictimiza-tion) without evoking a desire for angry revenge Although we acknowledge this, Retzinger (1991) underlined that helpers in reciprocal conflicts must often intervene to hinder an orientation towards angry revenge Moreover, she warns that the desire for revenge can be directed, not only onto the other in the conflict but also against helpers (see also, Lewis, 1971) Thus, the possible difference between an external versus an internal reminder may be more apparent than real

Concluding Thoughts

We realize that it is important to validate victim’s suffer-ing However, we would like to warn against simplistic thinking concerning reciprocal conflicts In many cases, reciprocal conflicts are marked by mutual humiliation and attacks (Brym & Araj, 2006; Retzinger, 1991; Staub, 2006; Stevenson et al., 2007) This feature of reciprocal conflict complicates the clear-cut understanding of who

is the victim and who is the perpetrator Also, Mazziotta

et al (2014) warned that both parties in a reciprocal conflict might construe themselves as the more sympa-thetic victim and resist the more suspect role as perpe-trator, which is synonymous with responsibility for harm (see also Leach et al., 2013; Mikula, 1993; Noor

et al., 2012) According to our results, this victim-focus may motivate revengeful rather than a reconciliatory orientation in all parties involved in reciprocal conflict This is of serious concern

Hence, it is important to take into account that a precondition for reconciliation is respect of the involved parties and their different perspectives (Mazziotta et al., 2014; Nadler, 2012), and to remember that these perspectives may elicit different orientations; some oriented towards reconciliations, some oriented towards revenge By also encouraging the involved parties to focus on their role as perpetrators of immoral-ity and harm, we may reduce the perceived risk to the group’s social image and also more directly increase their orientation to the empathic relationship repair seemingly necessary for true reconciliation

References

Bandura, A (1999) Moral disengagement in the

perpetra-tion of inhumanities Personality and Social Psychology

Review, 3, 193–209 https://doi.org/10.1207/s15327957

pspr0303_3 Batson, C D., Lishner, D A., Carpenter, A., Dulin, L., Harjusola-Webb, S., Stocks, E L., Gale, S., … Sampat, B (2003) “ As you would have them do unto you”: Does imagining yourself in the other’s place stimulate moral

action? Personality and Social Psychology Bulletin, 29,

1190–1201 https://doi.org/10.1177/0146167203254600 Baumeister, R F., Stillwell, A., & Wotman, S R (1990) Victim and perpetrator accounts of interpersonal conflict:

Autobiographical narratives about anger Journal of

Personality and Social Psychology, 59, 994–1005 https://doi.

org/10.1037/0022-3514.59.5.994

Bentler, P M., & Chou, C P (1987) Practical issues in

struc-tural modeling Sociological Methods & Research, 16, 78–117.

https://doi.org/10.1177/0049124187016001004

Berndsen, M., & Gausel, N (2015) When majority members

exclude ethnic minorities: The impact of shame on the

desire to object to immoral acts European Journal of Social

Psychology, 45, 728–741 https://doi.org/10.1002/ejsp.2127

Berndsen, M., & McGarty, C (2012) Perspective taking and

opinions about forms of reparation for victims of historical

harm Personality and Social Psychology Bulletin, 38,

1316–1228 https://doi.org/10.1177/0146167212450322

Brown, R., & Čehajić, S (2008) Dealing with the past and

facing the future: Mediators of the effects of collective guilt

and shame in Bosnia and Herzegovina European Journal of

Social Psychology, 38, 669–684 https://doi.org/10.1002/

ejsp.466

Brown, R., Gonzalez, R., Zagefka, H., Manzi, J., & Čehajić, S

(2008) Nuestra culpa: Collective guilt and shame as

predictors of reparation for historical wrongdoing Journal

of Personality and Social Psychology, 94, 75–90 https://doi.

org/10.1037/0022-3514.94.1.75

Brym, R J., & Araj, B (2006) Suicide bombing as strategy

and interaction: The case of the second intifada Social

Forces, 84, 1969–1986 https://doi.org/10.1353/sof.2006.

0081

Cain, K L (1999) The rape of dinah: Human rights, civil war

in Liberia, and evil triumphant Human Rights Quarterly, 2,

265–307 https://doi.org/10.1353/hrq.1999.0022

Carlsmith, K M., Darley, J M., & Robinson, P H (2002)

Why do we punish? Deterrence and just deserts as

motives for punishment Journal of Personality and Social

Psychology, 83, 284–299

https://doi.org/10.1037/0022-3514.83.2.284

Carmines, E G., & Zeller, R A (1979) Reliability and validity

assessment Beverly Hills, CA: Sage Publications.

Chaikin, A L., & Darley, J M (1973) Victim or perpetrator?

Defensive attribution of responsibility and the need for

or-der and justice Journal of Personality and Social Psychology,

25, 268–275 https://doi.org/10.1037/h0033948

Gausel, N (2006) Gravity keeps my head down, or is it maybe

shame? The effects of Essence and Reputation in Shame on

defensive responses, anger, blame and pro-social emotions.

Unpublished Master Thesis, University of Sussex

Gausel, N (2013) Self-reform or self-defense?

Understand-ing how people cope with their moral failures by

under-standing how they appraise and feel about their moral

failures In M Moshe, & N Corbu (Eds.), Walk of shame

(pp 191–208) Hauppauge, NY: Nova Publishers

Gausel, N., & Brown, R (2012) Shame and guilt – do they

re-ally differ in their focus of evaluation? Wanting to change

the self and behavior in response to ingroup immorality

The Journal of Social Psychology, 152, 1–20 https://doi.org/

10.1080/00224545.2012.657265

Gausel, N., & Leach, C W (2011) Concern for self-image

and social-image in the management of moral failure:

Rethinking shame European Journal of Social Psychology,

41, 468–478 https://doi.org/10.1002/ejsp.803

Gausel, N., & Salthe, G (2014) Assessing natural language:

Measuring emotion-words within a sentence or without

a sentence? Review of European Studies, 6, 127–132.

https://doi.org/10.5539/res.v6n1p127

Gausel, N., Leach, C W., Vignoles, V L., & Brown, R (2012) Defend or repair? Explaining responses to in-group moral failure by disentangling feelings of shame, inferiority, and

rejection Journal of Personality and Social Psychology, 102,

941–960 https://doi.org/10.1037/a0027233 Gausel, N., Vignoles, V L., & Leach, C W (2016) Resolving the paradox of shame: Differentiating among specific appraisal-feeling combinations explain

pro-social and self-defensive motivation Motivation and

Emotion, 40, 118–139

https://doi.org/10.1007/s11031-015-9513-y Imhoff, R., Bilewicz, M., & Erb, H P (2012) Collective regret versus collective guilt: Different emotional reactions to

his-torical atrocities European Journal of Social Psychology, 42,

729–742 https://doi.org/10.1002/ejsp.1886 Kanyangara, P., Rimé, B., Paez, D., & Yzerbyt, V (2014) Trust, individual guilt, collective guilt and dispositions toward reconciliation among Rwandan survivors and prisoners before and after their participation in

postgenocide gacaca courts in Rwanda Journal of Social

and Political Psychology, 2, 401–416 https://doi.org/

10.5964/jspp.v2i1.299

Kaplan, D (2009) Structural equation modeling: Foundations

and extensions Los Angeles, London, New Dehli, Singapore:

Sage Publications

Kline, R B (2005) Principles and practice of structural equation

modeling New York, NY: Guilford Press.

Leach, C W., & Cidam, A (2015) When is shame linked to

constructive approach orientation? A meta-analysis

Jour-nal of PersoJour-nality and Social Psychology, 109, 983–1002.

https://doi.org/10.1037/pspa0000037 Leach, C W., Iyer, A., & Pedersen, A (2006) Guilt and anger about in-group advantage as explanations of the willingness

for political action Personality and Social Psychology Bulletin,

32, 1232–1245 https://doi.org/10.1177/014616720628

9729 Leach, C W., Zeineddine, F B., & Čehajić-Clancy, S (2013) Moral immemorial: The rarity of self-criticism for previous generations’ genocide or mass violence

Journal of Social Issues, 69, 34–53 https://doi.org/ 10.1111/josi.12002

Leary, M R., Twenge, J M., & Quinlivan, E (2006) Interper-sonal rejection as a determinant of anger and aggression

Personality and Social Psychology Review, 10, 111–132.

https://doi.org/10.1207/s15327957pspr1002_2

Lewis, H B (1971) Shame and guilt in neurosis New York, NY:

International Universities Press

Lund, G (2003) “Healing the nation”: Medicolonial discourse and the state of emergency from apartheid to

truth and reconciliation Cultural Critique, 54, 88–119.

https://doi.org/10.1353/cul.2003.0036 Macaskill, A., Maltby, J., & Day, L (2002) Forgiveness of self

and others and emotional empathy The Journal of Social

00224540209603925

Machakanja, P (2010) National healing and reconciliation in

Zimbabwe: Challenges and opportunities Cape Town:

Insti-tute for Justice and Reconciliation

Mazziotta, A., Feuchte, F., Gausel, N., & Nadler, A (2014) Does remembering past ingroup harmdoing promote post-war cross-group contact? Insights from a field-experiment

in Liberia European Journal of Social Psychology, 44, 43–52.

https://doi.org/10.1002/ejsp.1986