This is an open access article distributed under the terms of the Creative Commons
Attribution License (“This is not anyone’s land but ours. This is the land of the Jewish people.”
Jewish-American blogger; Shemony, May 30, 2017
“This land is our land, no one has the right to do what they want in this land.”
Palestinian refugee; Fox, May 14, 2018
The quotes above illustrate two important issues regarding territorial conflicts. First, they differ from other conflicts in the sense that the rival groups feel that their own group is the more rightful, and sometimes the only rightful, owner of the land in question. Second, territorial claims are of great importance to those involved in the conflict. Such ownership perceptions and the importance of the territory for the groups involved might make support for reconciliation more difficult (Bar-Tal & Bennink, 2004), which renders territorial conflicts long-lasting and difficult to resolve (Toft, 2014). Despite the apparent importance of territory to the people living in conflict areas, previous research has addressed the issue predominantly from a macro-perspective, investigating for example territorial claims of political leaders in conflict resolution processes (Kaufman, 2009; Newman, 2006). Although scholars have theorized about the role of ownership perceptions, control, and territorial indivisibility in the reconciliation process (Bar-Tal, 2000; Goddard, 2006; Hassner, 2007), there is a lack of empirical research about individual perceptions of territorial ownership and whether these impede support for reconciliation.
In the psychological literature on possession, the term collective psychological ownership refers to people’s sense that an object, place, or idea belongs to their ingroup and that their ingroup has the exclusive right to decide what happens to it (Verkuyten & Martinovic, 2017). In this paper, we study collective psychological ownership among the inhabitants of different conflict areas, and have three research aims. First, to advance our understanding of collective psychological ownership of a territory, we examine whether it is different from place attachment – another, mainly emotional, tie people can have with a place (Hidalgo & Hernández, 2001). Second, we investigate how ingroup identification relates to collective psychological ownership as well as place attachment. Third, we examine whether people’s sense of collective ownership, rather than their attachment to the territory, relates to their support for reconciliation, where reconciliation is understood broadly as the development of positive relations between members of the conflicting groups. Taken together, we propose that collective psychological ownership (but not place attachment) is an important mediating mechanism that links ingroup identification with reconciliation support (e.g. Baysu, Coşkan, & Duman, 2018; Çelebi, Verkuyten, Köse, & Maliepaard, 2014). We study the contexts of Kosovo, Cyprus, and Israel and the Palestinian territories, and focus on one group in each context: Serbs, Greek Cypriots, and Israeli Jews.
Collective Psychological Ownership
Psychological ownership refers to an individual’s sense that they personally own an object, place, or idea; this sense can exist independently from legal ownership (Merrill, 1998; Snare, 1972). For instance, one can feel that an office is theirs, while it is actually owned by one’s company. Feelings of ownership extend beyond the individual level when groups experience a sense of collective ownership (Pierce & Jussila, 2010). When there is a feeling of ‘us’, there can also be a feeling of ‘ours’.
Perceptions of collective ownership can occur in the context of organizations, neighborhoods, and entire countries. These perceptions can have consequences for one’s attitudes towards one’s ingroup (Pierce & Jussila, 2010), but, importantly, also for how one relates to other groups (Verkuyten & Martinovic, 2017). A defining characteristic of (perceived) ownership is the right to control the object by prohibiting others from using it (gatekeeper right; Merrill, 1998; Snare, 1972). Conflicts over territories often emerge because two or more groups perceive that the territory is theirs. These feelings legitimize the exclusion of the other group from the territory, and might be central in shaping intergroup dynamics (Verkuyten & Martinovic, 2017).
People can develop various ties with places, not only based on ownership perceptions. An emotional attachment is known in the literature as place attachment, “a positive affective bond between an individual and a specific place” (Hidalgo & Hernández, 2001, p. 274). We argue that both collective psychological ownership and place attachment link people to places, but in different ways. Place attachment addresses feelings of belonging to a place (‘I or we belong here’), while collective psychological ownership refers to the perception that a place belongs to a group (‘the place belongs to us’). Collective psychological ownership involves the need for autonomy and control of one’s environment, and is accompanied by a bundle of rights and a gatekeeper right in particular (Verkuyten & Martinovic, 2017). The owner can grant another group access to the territory, but can also exclude outgroups. In contrast, place attachment involves the need to belong and is mainly an emotional, proximity-maintaining bond to a place, not accompanied by the wish to have control over the place (Scannell & Gifford, 2010). Consequently, we expect that collective psychological ownership and place attachment are distinct constructs that differently relate to reconciliation.
Collective Psychological Ownership, Place Attachment, and Reconciliation
In geopolitical regions that have experienced long conflicts, reconciliation is a prerequisite for harmonious coexistence between the conflicting groups, and therefore a relevant intergroup dynamic to investigate (Bar-Tal & Bennink, 2004). Achieving reconciliation in disputed territories is a complex endeavour. From a relational perspective, one must be willing to forgive, promote positive relations with the adversary, and expect harmonious coexistence in the future with a group that has long been perceived to be an ‘enemy’ (ibid). We address these three aspects of reconciliation and propose that a sense of collective ownership is related to lower reconciliation.
To achieve reconciliation, conflicting groups need to find an agreement on the territorial question. This can either be reached by dividing the territory, or one group needs to give up control over it. A greater sense of collective ownership of the territory hinders any incentive to compromise since that would imply giving up one’s territorial entitlements. As long as both conflicting groups perceive the territory as being rightfully theirs – and therefore having the exclusive right to decide about it (Merrill, 1998; Verkuyten & Martinovic, 2017) – it is unlikely that group members are willing to reconcile. Empirically, in the context of the Israeli-Palestinian conflict, Rosler, Sharvit, and Bar-Tal (2018) have shown that, among Israeli Jews, denial of Israel’s occupation of Judea and Samaria (a proxy for ingroup ownership perceptions – the territory is not occupied, it rightfully belongs to ‘us’) was related to less support for compromise.
In contrast, due to the absence of control-oriented feelings in place attachment, we do not expect this affective bond with the territory to represent a barrier to reconciliation. Research has shown that territorially attached people tend to trust others more than the non-attached ones (Lewicka, 2011). Feeling at home can imply a stronger willingness to include others who might also feel at home in the same place. Accordingly, we argue that feelings of belonging can be shared more easily, whereas the perception that a place is owned by ‘us’ tends to be more exclusive. Therefore, only psychological ownership but not place attachment is expected to relate to lower reconciliation.
Ingroup Identification
If collective psychological ownership is an obstacle to reconciliation, the next important question is who is more prone to hold a sense of ownership. Studies in the context of ethnic (Çelebi et al., 2014; Noor, Brown, Manzi, & Lewis, 2008; Van Tongeren, Burnette, O’Boyle, Worthington, & Forsyth, 2014) and religious (Baysu et al., 2018) group conflicts suggest that ingroup identification is important for understanding intergroup reconciliation. We propose that ingroup identification is also relevant for understanding who perceives more collective ownership, and that feelings of collective ownership mediate the relation between ingroup identification and reconciliation.
To have feelings of collective ownership, one needs to feel part of the collective: to feel that something is ‘ours’ one needs a sense of ‘we’ (Pierce & Jussila, 2010). Self-categorization theory (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) posits that categorizing oneself as a group member provides a lens for seeing and understanding the world in group terms. Individuals then tend to shift their beliefs in the direction of their ingroup’s beliefs (Hornsey, 2008). The ingroup provides a source of social influence (Turner, 1991). Moreover, high group identifiers tend to justify perceived rights and entitlements of their ingroup (Jost, Ledgerwood, & Hardin, 2008; Martinovic & Verkuyten, 2013), and (perceived) ownership may serve as a justification of such rights and entitlements to the territory (Merrill, 1998). Empirical studies showed that, even in contexts where territorial ownership is not contested (in Finland, the Netherlands, and the United States) majority members with a stronger sense of national identity were more likely to report ingroup collective ownership of the country (Brylka, Mähönen, & Jasinskaja-Lahti, 2015; Martinovic & Verkuyten, 2013; Wright, 2018). Similarly, in territorial conflicts, individuals who strongly identify with a group might be more prone to feel that their group collectively owns the territory. Perceptions of collective ownership are consequently expected to mediate the negative relation between ingroup identification and reconciliation.
Further, higher ingroup identifiers might also be more strongly attached to the contested territory. Territories tend to be of high value for national or ethnic groups and become symbols of the group identity throughout history (Lewicka, 2008; Low, 1992). For instance, places can be important for groups because of the presence of historical monuments or because of significant past events. According to social identity theory, when people identify more with a group, the group, along with the places that are of collective symbolic value, gain particular importance (Tajfel & Turner, 1986). Empirical research has shown that, in Ukraine, participants with a stronger national identity reported stronger attachment to places with a symbolic meaning for their nation (Lewicka, 2008). This relation should be particularly observed when the territory is contested and can, along with its symbolic value, be lost (Low, 1992; Scannell & Gifford, 2010). Therefore, we expected to find a positive relationship between ingroup identification and place attachment.
The Present Research
Bringing the theoretical arguments together, we hypothesize that in conflict contexts, collective psychological ownership and place attachment are two distinct constructs (H1), that higher identifiers have stronger feelings of collective ownership of the disputed territory (H2) and of place attachment (H3), and that collective ownership feelings – but not place attachment – relate to less reconciliation (H4). Only collective psychological ownership is thus expected to mediate the relation between ingroup identification and reconciliation. In Study 1 we focus on promoting positive relations and willingness to forgive as two aspects of reconciliation, and in Study 2 on reconciliation expectations.
The hypotheses were tested in the conflicts between Serbs and Albanians regarding Kosovo, Greek Cypriots and Turkish Cypriots regarding the northern part of Cyprus, and Israeli Jews and Israeli Arabs/Palestinians regarding Israel and the Palestinian territories. We focus on one of the two conflicting groups per setting – the group that we could get access to. In Study 1 we addressed the Kosovo conflict from the perspective of Serbs living in Serbia. To test whether our findings can be conceptually replicated, in Study 2, next to a new sample of Serbs in Serbia, we considered Serbs in Kosovo, Greek Cypriots in the southern parts of Cyprus, and Jews in Israel.
The Conflict Contexts
The three conflict contexts examined are similar in the sense that the dispute over territory plays a major role, and they are characterized by their longevity and intractability. Nevertheless, the conflicts differ in various ways, as do the groups that were examined and their status in the respective conflict.
Kosovo was an autonomous province in the Republic of Serbia until the unilateral declaration of independence in 2008. Kosovo’s independence is generally supported by Kosovo Albanians, who form the majority in Kosovo, and by 97 out of 193 United Nations member states (by March 2020), while it is contested by the Serbian minority in Kosovo and by Serbs in Serbia, whose official politics is that the region is still part of Serbia. The sample of Serbs in Serbia represents a politically powerful majority in their country, but one that has lost the territory of Kosovo and might not be in a position to regain this territory. In contrast, Serbs in Kosovo inhabit the territory in question and are – as compared to Albanians – a less powerful minority group in Kosovo.
The conflict in Cyprus concerns Greek Cypriots (the majority) and Turkish Cypriots (the minority). Since the Turkish Cypriot leadership declared independence of Northern Cyprus in 1983 – only recognized as an independent state by Turkey – the island has been divided into two parts by a UN-patrolled buffer zone. The southern part is governed by Greek Cypriots, who mostly perceive Northern Cyprus as occupied by Turkey and as being part of the Republic of Cyprus. Negotiations are underway between representatives of the two communities to settle the conflict (Psaltis & Cakal, 2016). The group investigated here – Greek Cypriots – is more powerful in numerical terms and because their position is backed up by the international community. However, they have lost territory that they perceive to be rightfully theirs.
The conflict about Israel and the Palestinian territories between Israeli Jews and Palestinians/Arabs is still on-going. In 1947, the United Nations decided to divide the land, followed by Israel’s declaration of independence and outbreak of war between Israel and its Arab neighbors. The borders of the State of Israel, established in 1949, have since been disputed. For decades, the region has faced several wars, and Israeli settlements on Palestinian territories in the West Bank have been established. Despite many peace negotiations, until today Israelis and Palestinians cannot agree on a peaceful solution (Bar-Tal & Salomon, 2006). Jewish Israelis – the group we focus on – form the numerical majority and are politically and economically more powerful in Israel (Heller, 2000). This puts Jewish Israelis in a dominant position regarding the Israeli-Palestinian conflict.
In sum, all groups examined in the article may feel that they have lost territory that rightfully belongs to their group, and compromising would mean giving up certain entitlements. At the same time, there are differences between the groups regarding their status and negotiating position. Testing the theoretical propositions among different groups and contexts allows us to examine the generalizability of these processes.
Study 1
In Study 1 we wanted to distinguish collective psychological ownership from place attachment. Further, we tested among Serbs in Serbia whether higher identifiers had stronger feelings of collective ownership of Kosovo, and felt more attached to it. Finally, we examined whether collective psychological ownership, rather than place attachment, was related to less support for reconciliation with Albanians and whether it mediated the relation between ingroup identification and reconciliation.
Method
Data and Participants
Participants were recruited by researchers from the University of Belgrade predominantly among students, but also via snowball sampling and through online platforms, thereby reaching a more diverse group. Nevertheless, 95% of the sample had a bachelor degree or higher. All participants received an online questionnaire; participation was voluntary, no incentives were provided. Responding to the questionnaire took 15-20 minutes. This study was part of a larger survey that included more scales (e.g. psychological needs, attitudes toward groups, religiosity), and participants reported sociodemographic characteristics. The order in which the constructs of interest were asked matched the order of the theoretical model.
After excluding thirteen participants with two non-Serbian parents, our sample consisted of 264 Serbs aged 19 to 51 (M = 25.00, SD = 6.80); 35% were male. Twenty-six participants were born outside Serbia to at least one Serbian parent, or were born in Serbia, to only one Serbian parent.
Measurements
All questions were measured on 7-point scales ranging from 1 = ‘Strongly disagree’ to 7 = ‘Strongly agree’, unless indicated otherwise.
The main independent construct ingroup identification was assessed by three items adapted from Martinovic and Verkuyten (2013): ‘I feel Serbian’; ‘Being Serbian is an important part of my identity’, and ‘I identify with other Serbs’ (ρ = .87; Raykov, 2004)i.
Due to the scarcity of previous research measuring the mediator collective psychological ownership of territories,ii we designed three items based on the construct definition (Verkuyten & Martinovic, 2017), which formed a reliable scale (ρ = .82). The items were: ‘In your opinion, how much does the territory of Kosovo belong to Serbs’; ‘To what extent do you consider Serbs as the rightful owner of the territory of Kosovo?’, and ‘How much would you say Serbs have the right to claim the territory of Kosovo for themselves?’ (1 = ‘Not at all’ to 7 = ‘Very much’).
The second mediator, place attachment (ρ = .86), was measured with two items (Hernández, Hidalgo, Salazar-Laplace, & Hess, 2007): ‘The territory of Kosovo means a lot to me’ and ‘I am very attached to Kosovo’.
The dependent construct reconciliation was assessed in two ways. Promoting positive relations was measured by three items adapted from Shnabel, Nadler, Ullrich, Dovidio, and Carmi (2009) that formed a reliable scale (ρ = .89): ‘I am willing to promote good relations between Albanians and Serbs’; ‘I am willing to express good will towards Albanians’, and ‘I am willing to learn more about Albanian culture to overcome disputes between Serbs and Albanians’. Willingness to forgive was also assessed by three items (Noor, Brown, & Prentice, 2008) that formed a reliable scale (ρ = .86): ‘I am prepared to forgive Albanians for the bad things they did to Serbs during the war in Kosovo’; ‘I try not to hold a grudge against Albanians for their misdeeds’, and ‘I think I am not able to forgive Albanians for their mistreatment of Serbs’ (reverse coded).
Control variables included age (years), gender (0 = female, 1 = male), educational level (9 levels, treated as continuous; 1 = without any formal education to 9 = doctoral degree), and native (0 = born abroad to at least one Serbian parent, or in Serbia to only one Serbian parent, 1 = born in Serbia to two Serbian parents).
Measurement Model
All constructs were modelled as latent variables to more precisely estimate the effect sizes and standard errors. With a Confirmatory Factor Analysis (CFA) we estimated our measurement model using the Maximum Likelihood Estimator (ML) in Mplus version 8.0 (Muthén & Muthén, 1998-2012). First, we estimated a model with the theorized five latent constructs: ingroup identification, collective psychological ownership, place attachment, promoting positive relations, and forgiveness. These latent variables were allowed to correlate. This model fitted the data well: χ2(67) = 116.68, p < .001;iii CFI = .978; SRMR = .043; RMSEA [90% CI] = .053 [.036, .069] (for information on fit indices, see Hu & Bentler, 1999). There were no considerable cross-loadings of items; standardized factor loadings of all items were .66 or higher.
To further test whether collective psychological ownership and place attachment are distinct, we compared our suggested 5-factor model to a 4-factor model where collective psychological ownership and place attachment were forced on one factor, while the other constructs formed separate factors. Since various different factor models may fit the data well, this method can be used to test whether the theorized model is better (Thompson, 2004, p. 142). The 4-factor model did not have a satisfying fit, χ2(71) = 314.99, p < .001; CFI = .892; SRMR = .087; RMSEA [90% CI] = .11 [.101, .127], and the fit was significantly worse than that of the 5-factor model, ∆χ2(4) = 198.31, p < .001. Thus, collective psychological ownership and place attachment are distinct constructs (see Table S1 in the Supplementary Materials, where we also show that all other combinations of 4-factor models had a significantly worse fit).
Results
Descriptive Findings
Descriptive statistics of the core constructs and their statistical differences from the neutral midpoint of the scale are displayed in Table 1. Participants on average identified with Serbia and perceived that Kosovo belongs to Serbs, but were not strongly attached to Kosovo. Furthermore, they were willing to promote positive relations with the conflicting outgroup but were neutral regarding their willingness to forgive. Mean comparisons across constructs showed that ownership feelings were higher than place attachment, Wald(1) = 133.82, p < .001, and promoting positive relations was higher than the willingness to forgive, Wald(1) = 61.65, p < .001.
Table 1
Means and Standard Deviations of the Manifest Core Constructs, Study 1 (N = 264)
| Construct | Range | M | SD | Wald (df) |
|---|---|---|---|---|
| 1. Ingroup identification | 1-7 | 4.70 | 1.53 | 54.88 (1)*** |
| 2. Collective psychological ownership | 1-7 | 5.29 | 1.53 | 187.30 (1)*** |
| 3. Place attachment | 1-7 | 3.51 | 1.97 | 16.24 (1)*** |
| 4. Promoting positive relations | 1-7 | 5.10 | 1.51 | 139.19 (1)*** |
| 5. Forgiveness | 1-7 | 4.06 | 1.53 | 0.41 (1) |
Note. Means of latent variables are not meaningful. Therefore we present means of manifest variables.
***p < .001.
All latent constructs were significantly correlated with each other in the expected direction (see Table S2 in the Supplementary Materials). Additionally, place attachment was negatively correlated with both reconciliation measures.
Mediation Model
Next, we specified a structural equation model with promoting positive relations and forgiveness as the dependent variables, ingroup identification as the independent variable, and collective psychological ownership and place attachment as mediators, controlling for standard demographics in relation to the mediators and dependent variables.
We estimated bootstrapped standard errors and confidence intervals with 5000 bootstraps drawn to receive more reliable estimations of the mediation paths. Indirect effects are significant when the 95% confidence interval does not include zero (Preacher & Hayes, 2008). Missing values on the manifest variables were dealt with using Full Information Maximum Likelihood (FIML; Graham, 2003). Unstandardized estimates are depicted in Figure 1 (see Table S2 in the Supplementary Materials for standardized results).
Figure 1
Mediation model, Study 1 (N = 264).
Note. Unstandardized direct effects presented, with standard errors in round brackets and total effects in square brackets; the covariance of the two mediators (b = 0.64, SE = 0.17, p < .001) and the two dependent variables (b = 1.21, SE = 0.19, p < .001) was accounted for. We accounted for covariances between control variables and ingroup identification as well as among the control variables. Control variables are not shown. Model fit: χ2(103) = 176.62, p < .001; CFI = .968; SRMR = .040; RMSEA [90% CI] = .052 [.039, .065].
**p < .01. ***p < .001.
Higher Serbian identifiers were less willing to promote positive relations with Albanians or to forgive them, as shown by significant total effects. Furthermore, as expected, higher identifiers expressed stronger feelings of collective ownership of Kosovo and felt more attached to it. Only feelings of ownership however, and not attachment, were related to lower willingness to promote positive relations and to forgive Albanians for their alleged misdeeds.iv The indirect effect via collective psychological ownership was significant: participants who more strongly identified as Serbian were less willing to promote positive relations or to forgive, via their stronger feelings of collective ownership. Place attachment, however, did not operate as a mediator. The remaining direct paths from identification to the two reconciliation measures were not significant and none of the control variables was related with any of the outcomes (see Table S3 in the Supplementary Materials).
Summary
Our findings among Serbs in Serbia indicate that collective psychological ownership and place attachment are distinct constructs. Further, higher ingroup identifiers tended to have a stronger feeling that Kosovo belongs to their ingroup, and this sense of ownership is related to less willingness to reconcile with Albanians. Even though higher identifiers were also more attached to Kosovo, place attachment was no longer associated with reconciliation once collective psychological ownership was taken into account. Thus, the emotional bond with Kosovo seems to matter less for Serbs’ support for reconciliation than the ownership-based bond with Kosovo.
Study 2
We wanted to replicate the findings from Study 1 among a new sample of Serbs in Serbia, and to additionally consider Serbs living in Kosovo, Greek Cypriots in Cyprus, and Jews in Israel. We tested our hypotheses in a multiple-group framework, examining whether the pattern of findings can be replicated in each group. We used somewhat different measures of collective psychological ownership, place attachment, and reconciliation compared to Study 1. While this was partially determined by data availability, such a conceptual replication tests the same pattern of associations, while operationalizing different aspects of the same phenomena. The confidence in the validity of the theoretical propositions is increased when hypotheses are supported using different operationalizations (Stroebe & Strack, 2014). A conceptual replication of our findings from Study 1 would show that the associations found can be generalized across different operationalizations of the same phenomena.
The collective psychological ownership measure captured the agreement with arguments that people might put forward to justify why the ingroup owns the territory more than the outgroup (ownership principles; Verkuyten & Martinovic, 2017). Prominent examples of such principles are first possession (also labelled autochthony; Ceuppens & Geschiere, 2005), and investment (Beggan & Brown, 1994). Since the concept of collective psychological ownership does not have a standing in the literature on intergroup relations, operationalizing it in a different way in Study 2 allows us to compare two possible measurements and to advance our understanding of the content of that concept.
Further, we shifted place attachment to the collective level in Study 2. People can develop an emotional attachment to a place because of its cultural and symbolic importance for one’s ingroup, and especially disputed and lost territory can become important to one’s group (Low, 1992; Scannell & Gifford, 2010). We suggest that when a place is of symbolic importance for a group, individuals may not only feel that they individually belong to a place, but also that their entire group belongs there. When place attachment refers to an affective feeling of the group’s belonging to a place (‘we belong here’), it should not result in controlling behavior and refusal to share, but rather in feelings of belonging and closeness, in contrast to collective psychological ownership with its territorial entitlements (‘the place belongs to us’). Hence, when both place attachment and psychological ownership bind a group to a place, only collective psychological ownership is expected to be related negatively to reconciliation.
Finally, we focused on reconciliation expectations – the perception that the two groups will be able to live harmoniously together in the future. This operationalization differs from the measures in Study 1 (promoting positive relations, being willing to forgive) in that expectations regarding reconciliation may be influenced by external circumstances. Previous research has used such measures (Baysu et al., 2018; Čehajić-Clancy & Bilewicz, 2017; Nadler & Liviatan, 2006; Shnabel et al., 2009), and we do not expect that the relations between the constructs will change. Due to pessimistic external circumstances and long history of conflict, reconciliation expectations may be lower than people’s willingness to promote positive relations or to forgive. Nevertheless, high ingroup identifiers and those with strong feelings of collective ownership of the disputed territory are expected to perceive the possibility of harmonious future relations as less likely than weaker identifiers or those with a weaker sense of collective ownership.
Method
Data and Participants
We used data collected within the framework of COST Action IS1205 ‘Social psychological dynamics of historical representations in the enlarged European Union’.v The study was conducted online among students (voluntary participation) in different countries – also among Dutch in the Netherlands and Croats in Croatia. For our analyses, we used the data collected among Serbs in Serbia, Serbs in Kosovo, Jews in Israel, and Greek Cypriots in Cyprus because of the ongoing conflicts in these regions.
The data set consisted of 546 students: 173 from Serbia, 129 from Kosovo, 135 from Cyprus, and 109 from Israel. Twenty-three percent were male, and the age ranged from 18 to 63 years (M = 22.93, SD = 5.31; for demographic information per sample, see Table 2). Older participants were mostly from Israel. This could be due to their compulsory military service. The sample included 108 participants with a migration background (19.8%), however, they belonged to the respective ethnic ingroup, for instance Jews who migrated to Israel from other parts of the world.
Table 2
Demographic Composition of the Participants per Sample, Study 2 (Total N = 546)
| Sample | N | Age
|
Proportion
|
||
|---|---|---|---|---|---|
| M | SD | Male | Native | ||
| Serbs (Serbia) | 173 | 21.27 | 2.32 | 20% | 73% |
| Serbs (Kosovo) | 129 | 21.91 | 2.93 | 19% | 83% |
| Greek Cypriots | 135 | 21.28 | 2.81 | 23% | 85% |
| Jewish Israelis | 109 | 28.51 | 8.22 | 32% | 80% |
Measurements
All questions were measured on 7-point scales ranging from 1 = ‘Strongly disagree’ to 7 = ‘Strongly agree’, unless indicated otherwise.
Ingroup identification (ρ = .89) was measured by three items with reference to the ethnic group. The first question was ‘How strongly do you feel [ingroup]?’ (similar to Postmes, Haslam, & Jans, 2013), answered on a 10-point scale (1 = ‘Not at all’ to 10 = ‘Very much’). The other two items were similar to Study 1 (7-point scales): ‘Being [ingroup] is an important part of my identity’ and ‘Being [ingroup] is an important way of how I see myself’.vi
Collective psychological ownership was assessed by three items (ρ = .86), representing the autochthony and investment principles on which ownership claims can be based (Verkuyten & Martinovic, 2017). The items were: ‘[Contested territory] belongs more to [ingroup] than to [outgroup] because 1) [ingroup] were there first; 2) [ingroup] have fought for this territory in the past; 3) [ingroup] have invested the most in building [contested territory] into what it is today’ (see Table S4 in the Supplementary Materials for exact formulations per sample).
Place attachment (ρ = .92) was measured by two items, following Hernández and colleagues (2007), but shifting them to the collective level: ‘[Contested territory] feels like our [(ingroup’s)] home’; and ‘I feel that this is the place where we [ingroup] belong’.
Reconciliation expectations were measured by three items (ρ = .88): ‘It is very likely that [outgroup] and [ingroup] will live harmoniously together in the future’; ‘I believe that [outgroup] and [ingroup] can successfully cooperate together’ and ‘I can imagine [ingroup] getting along perfectly well with [outgroup] someday’ (Baysu et al., 2018).
We controlled for age in years (continuous), gender (0 = female, 1 = male), and native (0 = born abroad, 1 = born in respective country).
Measurement Model
We estimated a CFA in Mplus with the theorized four distinct factors ingroup identification, collective psychological ownership, place attachment, and reconciliation expectations, allowing these latent variables to correlate. This model fitted the data well: χ2(38) = 90.43, p < .001; CFI = .987; SRMR = .036; RMSEA [90% CI] = .050 [.037, .064]. Standardized loadings of all items were .69 and higher and there were no considerable cross-loadings.
Next, we tested whether collective psychological ownership was distinct form collective place attachment. We compared the 4-factor model to a 3-factor model where collective psychological ownership and place attachment formed one factor, while ingroup identification and reconciliation loaded on separate factors. This 3-factor model did not have a satisfying model fit, χ2(41) = 701.28, p < .001; CFI = .835; SRMR = .104; RMSEA [90% CI] = .172 [.161, .183], and the fit was significantly worse compared to the 4-factor model, ∆χ2(3) = 610.85, p < .001. Thus, when both psychological ownership and place attachment refer to a group’s bond with a place, our findings show these are still two distinct phenomena (see Table S5 in the Supplementary Materials, where we also show that all other combinations of 3-factor models had a significantly worse fit).
Next, we tested for measurement invariance of the 4-factor model to ensure that participants from the four different contexts interpreted the questions in similar ways (Putnick & Bornstein, 2016). A model with scalar invariance (i.e., with factor loadings and intercepts constrained to be equal across samples) did not fit the data very well, χ2(194) = 444.48, p < .001; CFI = .934; SRMR = .089; RMSEA [90% CI] = .097 [.085, .109]. Following the modification indices, we estimated a partially scalar measurement model, keeping all factor loadings constrained and allowing the intercept of one item measuring reconciliation (’I believe that [outgroup] and [ingroup] can successfully cooperate together’) to vary between groups. From a theoretical perspective, this item differs from the other two by focusing on cooperation, which entails active involvement with the outgroup. Whether involvement can be easily achieved might depend on context specific aspects, such as the degree of spatial segregation. Relaxing this model constraint significantly improved the model fit, ∆χ2(3) = 99.58, p < .001, and resulted in an acceptable model, χ2(191) = 344.91, p < .001; CFI = .960; SRMR = .078; RMSEA [90% CI] = .077 [.064, .090]. A model with partial scalar invariance allows for meaningful comparisons of associations between constructs across conflict contexts (Putnick & Bornstein, 2016).
Results
Descriptive Findings
Descriptive statistics for the total sample and per regional group are displayed in Table 3, including the Wald test differences from the neutral midpoint for all core constructs. Participants on average identified strongly with their ethnic group.
Table 3
Means and Standard Deviations of the Manifest Core Constructs, Study 2 (Total N = 546)
| Construct | Range | Total sample
|
Serbs(Serbia)
|
Serbs(Kosovo)
|
Greek Cypriots
|
Jewish Israelis
|
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| M | SD | Wald (df) | M | SD | Wald (df) | M | SD | Wald (df) | M | SD | Wald (df) | M | SD | Wald (df) | ||
| Ingroup identification | 1-7 | 4.79 | 1.64 | 126.16 (1)*** | 3.80kci | 1.68 | 2.46 (1) | 5.11si | 1.35 | 87.66 (1)*** | 5.10si | 1.35 | 89.24 (1)*** | 5.58skc | 1.45 | 129.81 (1)*** |
| Collective psychological ownership | 1-7 | 5.18 | 1.58 | 297.51 (1)*** | 4.81k | 1.61 | 43.15 (1)*** | 5.94sci | 1.21 | 333.75 (1)** | 5.07k | 1.69 | 52.11 (1)*** | 4.96k | 1.46 | 46.00 (1)*** |
| Place attachment | 1-7 | 4.83 | 1.78 | 63.64 (1)*** | 3.18kci | 1.34 | 63.64 (1)*** | 5.55sci | 1.35 | 169.20 (1)*** | 5.17ski | 1.49 | 80.43 (1)*** | 6.17skc | 1.09 | 431.02 (1)*** |
| Reconciliation expectations | 1-7 | 3.35 | 1.52 | 95.02 (1)*** | 3.30kc | 1.34 | 45.48 (1)*** | 2.91sci | 1.32 | 88.13 (1)*** | 3.96ski | 1.58 | 0.08 (1) | 3.26kc | 1.71 | 20.41 (1)*** |
Note. Subscripts indicate a significant (p < .05) mean difference from the sample of sSerbs(Serbia), kSerbs(Kosovo), cGreek Cypriots, and iIsraeli Jews.
**p < .01. ***p < .001.
Only Serbs in Serbia were neutral about their identification; their level of identification was significantly lower than in the remaining groups. In all samples, participants perceived that the territory in question belonged to their own group rather than to the conflicting outgroup. Serbs in Kosovo scored especially high on collective psychological ownership (significantly higher than others). Furthermore, Serbs in Kosovo, Greek Cypriots, and Israeli Jews scored higher on place attachment than the neutral midpoint, while Serbs in Serbia scored below the midpoint. Israeli Jews were especially attached to the contested territory, and more than all other groups. Greek Cypriots had neutral reconciliation expectations, and significantly higher than participants in other samples whose reconciliation expectations were below the midpoint.
Most correlations in the total sample as well as per group (Table S6 in the Supplementary Materials) were significant and in line with expectations. There were two exceptions. First, in the total sample ingroup identification and reconciliation expectations were not related, and also not among Serbs in Kosovo and Jewish Israelis. Second, we found a significant negative correlation between place attachment and reconciliation in the total sample, which was replicated in all but the Jewish Israeli sample.
Mediation Model
We estimated structural equation models in a multiple-group framework. Reconciliation expectations was the dependent variable, ingroup identification the predictor, and collective psychological ownership and place attachment the mediators. We controlled for standard demographic variables in relation to the mediators and the dependent variable. We applied bootstrapping with 5000 iterations. Missing values on the manifest variables were dealt with using FIML.
To inspect whether there are any meaningful differences in the associations across groups, we first tested an unconstrained model where we allowed all hypothesized paths to vary across Serbs in Serbia, Serbs in Kosovo, Greek Cypriots, and Jewish Israelis against a model where these paths were constrained to be equal across groups. If the fit of the two models is comparable, the more parsimonious constrained model should be interpreted.
The unconstrained model had a satisfying model fit χ2(281) = 491.91, p < .001; CFI = .946; SRMR = .071; RMSEA [90% CI] = .074 [.063, .085]. This was not better than the constrained model, χ2(323) = 541.82, p < .001; CFI = .944; SRMR = .090;vii RMSEA [90% CI] = .070 [.060, .081], as indicated by a non-significant difference test, ∆χ2(42) = 49.91, p = .188. Unstandardized results of the constrained model are depicted in Figure 2. Unstandardized coefficients of the constrained model (including control variables; Model 1), and standardized coefficients of the unconstrained model (Model 2) can be found in Table S7 (see Supplementary Materials).
Figure 2
Mediation model, Study 2 (total N = 546).
Note. Unstandardized direct effects of a constrained multiple-group model presented, with standard errors in round brackets and total effects in square brackets; the covariance of the two mediators was accounted for, b(Serbs(Serbia)) = 0.76, SE = 0.12, p < .001; b(Serbs(Kosovo)) = 0.27, SE = 0.10, p < .001; b(Greek Cypriots) = 0.91, SE = 0.22, p < .001; b(Jewish Israelis) = 0.16, SE = 0.14, p = .28. We accounted for covariances between control variables and ingroup identification as well as among the control variables. Control variables are not shown. Model fit: χ2(335) = 541.82, p < .001; CFI = .944; SRMR = .090; RMSEA [90% CI] = .070 [.060, .081].
**p < .01. ***p < .001.
Higher ingroup identifiers had lower reconciliation expectations – the total effect of ingroup identification was negative and significant. In line with our expectations, higher ingroup identifiers had stronger feelings of collective ownership of the disputed territory, and felt more attached to it. Only collective psychological ownership however related to lower reconciliation expectations. Further, higher identifiers had lower reconciliation expectations via stronger feelings of collective ownership but not via place attachment. In the unconstrained model where we estimated the paths separately per group (see Table S7 in the Supplementary Materials) the mediation by ownership did not reach significance in two samples – among Serbs in Kosovo and Jewish Israelis – but the pattern of associations was the same as in the other two samples and none of the paths differed significantly across groups. The lack of significance regarding the indirect effect is probably due to the small sample size in combination with a relatively complex model.
Older participants had higher reconciliation expectations (b = 0.03, SE = 0.01, p = .044), but no other control variable was related to any of the outcomes (see Table S7 in the Supplementary Materials).
Summary
We have largely replicated the results of Study 1 conceptually and across three conflict areas. Place attachment (this time measured with reference to the group) was different from feelings of collective ownership, and only ownership but not place attachment mediated the path between ingroup identification and reconciliation expectations.
General Discussion
In the present article, we examined in two studies and across the conflict settings of Kosovo, Cyprus, and Israel and the Palestinian territories whether higher ingroup identifiers had stronger feelings of collective ownership of the contested territories and stronger place attachment. Moreover, we proposed that perceptions of ingroup’s ownership of the territory, rather than place attachment, mattered for individuals’ reconciliation intentions and expectations.
Research on the role of individuals’ concerns about territoriality in conflict reconciliation is scarce (see however Rosler et al., 2018). The main contribution of our research is that we showed that individuals’ perception of collective ownership of a territory can indeed be accompanied by lower reconciliation intentions and expectations in territorial conflicts. Our findings resonate with those from previous studies conducted in European countries where the territory is rather uncontested and where ownership claims were shown to relate to anti-immigrant attitudes (Brylka et al., 2015; Martinovic & Verkuyten, 2013). At the same time, we extended the existing research by focusing on areas of territorial conflicts, by using more diverse measures of collective psychological ownership, and by considering reconciliation intentions and expectations between groups in conflicts in relation to collective psychological ownership.
Further, our results indicate that a sense of collective ownership of a territory seems to be distinct from place attachment on an individual as well as collective level. Only the ownership-based connection with the disputed territory was consistently related to less reconciliation intentions and expectations, while place attachment was not, after having accounted for ownership. These findings provide evidence for the claim that ownership perceptions may be unique in that they legitimize the exclusion of other groups from the territory (Merrill, 1998; Verkuyten & Martinovic, 2017), whereas the affective bond with a territory does not have to result in exclusionary attitudes or behaviors (Scannell & Gifford, 2010) – as long as ownership feelings are taking into account.
Another contribution is that we provide novel insights into the relation between ingroup identification and reconciliation by examining the mediating role of collective psychological ownership. The results from our studies suggest that ingroup identification is related to stronger feelings of collective ownership of the disputed territory, and that collective psychological ownership mediates the relation between ingroup identification and reconciliation intentions and expectations. These findings are in line with theoretical reasoning that a high ingroup identification can stand in the way of positive intergroup relations because high ingroup identifiers tend to be more strongly concerned about ingroup interests and rights, such as ingroup entitlements (Jost, Ledgerwood, & Hardin, 2008; Martinovic & Verkuyten, 2013; Verkuyten & Martinovic, 2017). The fact that ingroup identification was mainly negatively related to reconciliation via enhanced feelings of collective ownership highlights the importance of considering collective psychological ownership as one mechanism in the relation between ingroup identification and intergroup relations in contexts where two or more groups can claim ownership of a territory.
Finally, participants in Study 2 came from three different conflict areas. All these groups are involved in conflicts related to ownership of the land. However, the current political contexts, and other aspect, such as the conflict intensity and the specific historical circumstances, differ across groups. The inclusion of such diverse samples confirmed that the obtained results were not specific to one particular social context but could rather generalize across these different conflicts. Further, in Study 2 we were able to conceptually replicate findings of Study 1, which shows that in addition to generalizability across conflict areas, our findings also generalize across different operationalizations of the same phenomena. This increases confidence in the validity of the proposed theoretical mechanism (Stroebe & Strack, 2014).
Limitations and Future Directions
Notwithstanding the contributions of the present research, several limitations provide directions for future research. First, our findings were based on convenience samples consisting mostly of students and highly educated participants. While we do not have reasons to expect differences in the relations between constructs, replications with larger and representative samples would be informative. For instance, students are often more liberal than the average population and the average (dis)agreement with the constructs might be different in a representative sample.
Second, due to the cross-sectional nature of our data, we cannot make directional causal claims regarding the mediating paths. While we have theoretical reasons and empirical evidence to assume that higher identifiers perceive collective ownership more strongly and are, in turn, more resistant to reconciliation (Brylka et al., 2015; Martinovic & Verkuyten, 2013; Wright, 2018), these phenomena might reinforce each other. Experimental and longitudinal research is needed to examine the (mutual) directions of influence.
Third, in both studies, we examined the perspective of only one conflicted party per context. In conflicts with asymmetrical power relations – one prominent example being the Israeli-Palestinian conflict – it is important to consider ingroup’s and outgroup’s perceived entitlements and actual power relations if we want to understand intergroup relations (Newman, 2006; Rouhana & Fiske, 1995). Future research should address this since reconciliation, to be successful, needs to be supported by members of both groups involved (Bar-Tal & Bennink, 2004).
Fourth, reconciliation is a complex process (e.g. Baysu et al., 2018; Noor, Brown, & Prentice, 2008; Shnabel et al., 2009). In our studies, we focused on the willingness to promote positive relations and to forgive (Study 1) and on reconciliation expectations (Study 2). The former measures focus on individuals’ motivation, while the latter focuses on aspects that are outside of the control of individuals. Using different operationalizations demonstrates that our findings generalize to two different aspects of reconciliation. Future studies simultaneously measuring both aspects of reconciliation could try to replicate our findings. More importantly, this research could be extended by investigating the content of reconciliation expectations: What do people have in mind when they (dis)agree with reconciliation expectations? For instance, reconciliation expectations may be mainly driven by acts and negotiations of political leaders, or rather by the behaviour of ingroup and/or outgroup members, or other actors involved.
Finally, it would be important to consider the conditions of reconciliation. Will the conflicting groups have to share the territory and power? Will the territory be split into separate spaces? Will the ingroup gain or lose power (Newman, 2006)? Reconciliation can have different ramifications and hence different expectations for individuals – also depending on whether the status quo benefits their group or not. For powerful groups, reconciliation may mean that the status quo is being legitimized, whereas for groups with little power reconciliation might imply the acceptance of their minority position. Studying peoples’ different expectations of political solutions and decisions regarding the conflict could further illuminate the relation between collective ownership perceptions and reconciliation.
Conclusion
A sense that one’s group has rightful ownership of contested land adds to territorial conflicts being long-lasting and difficult to resolve. The main contribution of the present study was to show that collective psychological ownership is an important phenomenon to consider when studying intergroup reconciliation in territorial conflict areas. Across two studies and three conflict areas, our results clearly show that feelings of collective ownership should not be ignored if we want to better understand why people in territorial conflict contexts differ in their attitudes towards reconciliation. A political discourse that strongly emphasizes exclusive ingroup land ownership can stand in the way of reconciliation and might be an important reason why many territorial conflicts are seemingly intractable.