Can 'we' share the contested territory with 'them'? Shared territorial ownership perceptions and reconciliation intentions in Kosovo.

Territorial ownership claims are central to many interethnic conflicts and can constitute an obstacle to conflict resolution and reconciliation. However, people in conflict areas might also have a perception that the territory simultaneously belongs to one's ingroup and the rival outgroup. We expected such perceptions of shared ownership to be related to higher reconciliation intentions. We examined this expectation in relation to the territory of Kosovo among random national samples of Albanians and Serbs from Kosovo, and Serbs from Serbia (Study 1, total N = 995). In general, participants perceived low levels of shared ownership, however, shared ownership perceptions were positively related to reconciliation intentions in Kosovo. In Study 2 (total N = 375), we experimentally manipulated shared ownership (vs. ingroup ownership) and found that shared ownership elicited stronger reconciliation intentions. It is concluded that fostering a sense of shared ownership can be important for improving intergroup relations in post-conflict settings.

Background

There are many territorial conflicts between ethnic groups in the world whereby two groups believe that the territory rightfully belongs to them (Toft, 2014). Beliefs about territorial ownership are often mutually exclusive – when one group is believed to be entitled to a territory, another group cannot be (Verkuyten & Martinović, 2017). As a result, these beliefs tend to be related to lower intergroup reconciliation intentions in territorial conflicts, including lower forgiveness and less willingness to establish positive relations with outgroup members (e.g., Storz et al., 2020). However, there is also the possibility that people believe that ownership of the territory is shared with the rival outgroup. This perceived outgroup co‐ownership can be expected to be relevant for the process of conflict resolution and reconciliation. Specifically, those who perceive shared ownership might be more inclined to support intergroup reconciliation.

The aim of the present research is to investigate if disputed territory can be perceived to be owned by both the ingroup and outgroup and whether shared ownership perceptions relate to reconciliation intentions – the willingness to promote and establish positive relations with the outgroup. We tested our expectations in the context of the Kosovo conflict and we adopted a two‐sided perspective by relying on random national samples of both Albanians and Serbs – the two groups which disagree about whom the territory of Kosovo belongs to (Judah, 2008). Because Kosovo used to be part of Serbia and the Serbian government has not recognized Kosovo’s independence, we further included the perspective of Serbs from Serbia.

Collective psychological ownership

Psychological ownership refers to an individual’s sense that something is ‘mine’, such as an object, a place, or an idea (Pierce, Kostova, & Dirks, 2001). A sense of ownership can exist independently from institutional and legal ownership in the form of property (Merrill, 1998; Snare, 1972). Feelings of ownership extend beyond the individual level when group members experience a sense of ownership over a target, referred to as collective psychological ownership (Pierce & Jussila, 2010). When there is a feeling of ‘us’, there can also be a feeling of ‘ours’.

A sense of ownership involves not only a subjective connection to what is owned but also shapes relationships between individuals regarding the things that are owned. This is because ownership implies a bundle of rights, such as the right to occupy or use that what is owned, but also to sell it to someone as well as prevent others from using it (Blumenthal, 2010; Snare, 1972). Similarly, a sense of collective ownership structures social relations and shapes how people think, feel, and act towards other groups, since ownership implies a bundle of ingroup entitlements and rights, among others the gatekeeper right to exclude outgroups (Merrill, 1998). Thus, the perception that a particular territory is ‘ours’ is likely to have negative consequences for one’s relations with newcomers or other groups living on that territory (Verkuyten & Martinović, 2017). For example, in Finland, the Netherlands, and the United Kingdom, it has been found that a stronger sense of collective ownership of a country or neighbourhood is related to more negative attitudes towards immigrant‐origin minorities (Brylka, Mähönen, & Jasinskaja‐Lahti, 2015; Nijs, Martinović, Verkuyten, & Sedikides, 2021; Torunczyk‐Ruiz & Martinović, 2020) and stricter anti‐immigration policies (Straver et al., 2021).

However, people might also recognize that another group co‐owns the territory in question, such as in Australia where founder and native ownership correlated positively with each other (Selvanathan, Lickel, & Jetten, 2020). Acknowledging shared ownership does not necessarily imply an equal degree of ownership, but rather any recognition that the outgroup has some justified claims to consider the territory as also belonging to them. We postulate that, even in post‐conflict settings such as Kosovo, there are people who perceive that both groups together own the territory – even if to a different degree – and that such shared ownership perceptions can improve intergroup relations.

Shared ownership perceptions and reconciliation intentions

Reconciliation goes beyond the formal resolution of conflict between political leaders. It involves changing the attitudes and beliefs about the adversary (Bar‐Tal & Bennink, 2004) and the willingness to meet and get to know people from the outgroup, and discuss relevant issues with them (Shnabel, Nadler, Ullrich, Dovidio, & Carmi, 2009). Achieving reconciliation in disputed territories is a challenge (Bar‐Tal & Bennink, 2004) but makes it possible to build trusting intergroup relations and develop future harmonious coexistence between the conflicting groups (Nadler, 2012).

Research in territorial conflict settings (Israel, Kosovo, and Cyprus) shows that stronger perceptions of ingroup ownership of disputed territory relate to lower support for reconciliation with the relevant outgroup (Storz et al., 2020). Israeli Jews, Serbs, and Greek Cypriots who perceived their ingroup to own the contested territory more were less willing to promote positive relations with outgroup members, less ready to forgive, and had lower expectations that the two groups would reconcile. In contrast, the perception of outgroup ownership of the disputed territory has been found to facilitate reconciliation efforts. Israeli Jews who acknowledged that their ingroup had occupied the territory of Judea and Samaria (i.e., acknowledgement of earlier Palestinian ownership of these contested territories) were more supportive of compromise in the Israeli‐Palestinian conflict (Rosler, Sharvit, & Bar‐Tal, 2018). Thus, recognizing that the ingroup has occupied a territory that was originally inhabited by an outgroup can relate positively to reconciliation intentions.

A perception of exclusive outgroup ownership is not very likely in territorial conflict regions and probably also not necessary for reconciliation. In such contexts, it might be easier to acknowledge shared ownership of the contested territory. People who agree more strongly that both groups own the disputed territory might be more willing to overcome the intergroup conflict by reconciling with the outgroup. With shared ownership, positive outcomes such as reconciliation intentions can be expected since one is likely to hold positive attitudes towards co‐owners (Verkuyten & Martinović, 2017), in this case the relevant outgroup. Further, it has been theorized that collective ownership may create a sense of collective responsibility, bind people together, and increase cooperation (Pierce & Jussila, 2010). Reconciliation intentions can be seen as a form of increased readiness to cooperate with the outgroup. Perceptions of shared ownership might thus provide a basis for reconciliation in a territorial conflict setting.

Hence, our main hypothesis is that stronger perceptions that ownership of disputed territory is shared between the ingroup and the conflicting outgroup are related to stronger intentions to reconcile with that outgroup. We tested this hypothesis in two studies – one correlational (Study 1) and one experimental (Study 2) – with regards to the Kosovo conflict.

The context of Kosovo

The conflict between Albanians and Serbs over Kosovo can be tracked down to the Ottoman rule of Kosovo in the 14th century (Daskalovski, 2004). Albanians claim to have been the first to live on the territory of Kosovo as they postulate to descend from the Illyrians and Dardanians who have inhabited the territory before Serbs arrived, whereas Serbs consider Kosovo the cradle of their civilization (Judah, 2008). Both groups use these historical narratives to claim Kosovo for themselves and the conflict is in large parts about the ownership of Kosovo by Albanians and/or Serbs (Daskalovski, 2004).

Politically, Kosovo was an autonomous province of the Socialist Federal Republic of Yugoslavia since 1945, located in the Republic of Serbia. Between February 1998 and June 1999, tensions between Kosovo’s Albanian population and Serbs resulted in an armed conflict. Kosovo unilaterally declared independence from Serbia in 2008 and is now recognized as an independent country by 98 of 193 UN member states, and 22 of 27 EU member states. The independence was supported by Albanians in Kosovo but mostly disputed by the Serbian minority. Additionally, the Serbian government officially views Kosovo as part of Serbia, a standpoint supported by many Serbs from Serbia (Center for Insights in Survey Research, 2015). Thus, since our research is about perceptions of territorial ownership, next to Albanians and Serbs from Kosovo, Serbs from Serbia are a relevant group to consider in this particular case. It is estimated that around 87% of the Kosovo population are Albanians and around 8% are Serbs (European Centre for Minority Issues Kosovo, 2013).

Study 1

In Study 1, we measured shared ownership in two ways: First, we separately assessed perceptions of ingroup and outgroup ownership and computed indirect shared ownership perceptions. Second, we directly asked about shared ownership perceptions. This might give participants from a conflict region more room for agreement compared with confronting them with separate questions on ingroup and outgroup ownership, whereby they might feel they have to choose one or the other. We inspected how common it is to perceive ingroup, outgroup, and shared ownership of Kosovo, and we tested our hypothesis that shared ownership (both measures) relates to more reconciliation intentions.

Method

Data and participants

We collected data among Albanians and Serbs in Kosovo and Serbs in Serbia. Participants were recruited by research agency Kantar in collaboration with TNS Bulgaria in 2019 and interviewed by interviewers from their own ethnic group. Following informed consent, they completed a questionnaire in Serbian or Albanian on a tablet (computer‐assisted personal interviewing). Multistage random probability sampling was applied to recruit participants. More information about the sampling method can be found in Online Supplement 1.

The sample consists of 390 Albanians in Kosovo (46.2% male), 200 Serbs in Kosovo (53.5% male), and 405 Serbs in Serbia (50.4% male).1 Albanians were on average 43 (SD = 16.75), Serbs in Kosovo 48 (SD = 16.31), and Serbs in Serbia 45 years old (SD = 15.37). Participants with primary, secondary and tertiary education were all represented (see Table S2.1 in Online Supplement 2 for more information).

Measures

The items were administered such that the ingroup was always mentioned first and the outgroup second. The reason is the politically sensitive context in which mentioning the outgroup first can be experienced as biased and insulting. We present the items as asked to Albanian participants but Serbs responded to the same items (with reversed mentioning of the groups). The reliability of the scales is indicated with the recommended composite reliability measure rho (ρ; Raykov, 2004).

Ingroup (ρAlbanians = .81; ρSerbs(Kosovo) = .82; ρSerbs(Serbia) = .90) and outgroup ownership perceptions (ρAlbanians = .64; ρSerbs(Kosovo) = .84; ρSerbs(Serbia) = .89) were measured by three items: ‘How much does the territory of Kosovo belong to the following group? Albanians; Serbs’, ‘To what extent can each of the following groups be seen as the rightful owner of the territory of Kosovo? Albanians; Serbs’, and ‘How much can each of these groups claim the territory of Kosovo for themselves? Albanians; Serbs’. The answer options ranged from (1) = ‘Not at all’ to (7) = ‘Very much’. These items were previously used by Storz et al. (2020) to measure ingroup ownership but here we use them to capture outgroup ownership as well.

Based on these two measures, we calculated a score representing indirect shared ownership perceptions using a formula developed in the literature on dual identity and constructive disruption (see Shuman, Saguy, van Zomeren, & Halperin, 2020). Our adapted formula is:

By subtracting the absolute value of the difference between ingroup and outgroup ownership perceptions from the sum of the two, this formula allocates higher values to participants who strongly agreed with both ingroup and outgroup ownership (shared ownership perceptions, e.g., (7 + 7) − |7 − 7| = 14 – 0 = 14), but lower values to participants who either agreed with no group ownership (e.g., (1 + 1) − |1 − 1| = 2 – 0 = 2), or only with one of the two (ingroup ownership perceptions, e.g., (7 + 1) − |7 − 1| = 8 − 6 = 2, or outgroup ownership perceptions, e.g., (1 + 7) − |1 − 7| = 8 − 6 = 2). Thus, our new measure had values from 2 to 14 in steps of two, which we rescaled to 1 to 7. As we had three items for ingroup ownership and the same three for outgroup ownership, we applied this formula three times and then used the indirect shared ownership perceptions as a latent factor (ρAlbanians = .66; ρSerbs(Kosovo) = .85; ρSerbs(Serbia) = .91).

Direct shared ownership perceptions were measured by three items designed specifically for this study that assessed the (dis‐)agreement with shared ownership of Kosovo on a 7‐point scale (1 = ‘Strongly disagree’ to 7 = ‘Strongly agree’): ‘I think that Albanians and Serbs own Kosovo together’, ‘I feel that Kosovo belongs to both Albanians and Serbs’, and ‘Albanians and Serbs share the ownership of Kosovo’ (ρAlbanians = .82; ρSerbs(Kosovo) = .78; ρSerbs(Serbia) = .93).

Reconciliation intentions were assessed by seven items (inspired by Shnabel et al., 2009) on 7‐point scales (1 = ‘Strongly disagree’ to 7 = ‘Strongly agree’): ‘I am willing to promote good relations between Albanians and Serbs’, ‘I am willing to get to know more Serbs’, ‘I am willing to participate in a workshop that brings Albanians and Serbs together’, ‘I am willing to convince my Albanian friends of the importance of establishing positive relations with Serbs’, ‘I am willing to publicly criticize other Albanians whose attitudes towards Serbs are prejudicial’, ‘I would easily accept if someone in my family married a Serb’, and ‘I would let my children play with Serbian children’ (ρAlbanians(Kosovo) = .89; ρSerbs(Kosovo) = .90; ρSerbs(Serbia) = .93).

We controlled for religiosity with the question ‘How religious are you?’ (1 = ‘Not at all’ to 7 = ‘Very much’), as there is evidence that in intergroup conflicts between religiously distinct groups, religiosity can stand in the way of support for political solutions (see e.g., Maoz & Eidelson, 2007, on the Israeli‐Palestinian conflict). Moreover, political orientation on a 5‐point left to right scale was included as a control variable (Shamir & Shikaki, 2002). Finally, we controlled for age measured in years (continuous), gender (0 = female; 1 = male), and educational level 2 (primary, secondary, and tertiary education, with secondary as reference category).

Results

Measurement invariance

We first examined whether ingroup ownership, outgroup ownership, and directly assessed shared ownership are empirically distinct constructs and whether they differ from reconciliation intentions. Furthermore, we determined whether these four latent constructs have the same meaning across Albanians from Kosovo, Serbs from Kosovo, and Serbs from Serbia (Putnick & Bornstein, 2016). Thus, we fitted a series of multi‐group measurement models and tested for measurement invariance in Mplus version 8.0. We used the Maximum Likelihood robust estimator to account for the skewness in ingroup and outgroup ownership perceptions (Lei & Shiverdecker, 2020).

A model with scalar invariance (i.e., with factor loadings and intercepts constrained to be equal across the three subsamples) did not fit the data well (see Table S3.1 in Online Supplement 3). The metric model in which factor loadings were constrained to be equal and intercepts were allowed to vary across subsamples, had a reasonable model fit. Apart from a significant Chi‐square test, all other fit indices show that the fit was not significantly worse compared with that of the configural model where also loadings varied between subsamples.3 We accepted the metric model, which does not allow us to statistically compare latent means across samples, but allows for a comparison of regression coefficients (Putnick & Bornstein, 2016).

Descriptive findings

We first inspected the latent measures for ingroup and outgroup ownership perceptions. In all three groups, ingroup ownership perceptions were exceptionally high, whereas outgroup ownership perceptions were below the neutral midpoint of the scale (see Table 1, with Wald tests assessing differences from the scale midpoint). The correlation between ingroup and outgroup ownership was negative among all three subsamples (r (Albanians from Kosovo) = −.21, p = .008; r (Serbs from Kosovo) = −.46, p < .001; r (Serbs from Serbia) = −.60, p < .001).

Table 1

Means, standard deviations and Wald test of the latent variables; Study 1 (total N = 995)

	M	SD	Wald(df)
Total sample (N = 995)
Ingroup ownership perceptions	6.36	1.04	5,127.33(1)***
Outgroup ownership perceptions	2.31	1.32	1,640.84(1)***
Indirect shared ownership perceptions	2.34	1.17	2,047.63(1)***
Direct shared ownership perceptions	3.30	1.55	201.37(1)***
Reconciliation intentions	3.19	1.36	363.36(1)***
Albanians (Kosovo; N = 390)
Ingroup ownership perceptions	6.62	.89	3,471.93(1)***
Outgroup ownership perceptions	1.96	1.11	1,307.96(1)***
Indirect shared ownership perceptions	1.93	1.09	1,398.99(.1)***
Direct shared ownership perceptions	3.05	1.58	140.39(1)***
Reconciliation intentions	3.29	1.32	104.06(1)***
Serbs (Kosovo; N = 200)
Ingroup ownership perceptions	6.60	.79	2,184.51(1)***
Outgroup ownership perceptions	2.15	1.10	565.83(1)***
Indirect shared ownership perceptions	2.11	1.05	653.30(1)***
Direct shared ownership perceptions	2.93	1.13	176.45(1)***
Reconciliation intentions	3.07	1.16	123.09(1)***
Serbs (Serbia; N = 405)
Ingroup ownership perceptions	5.99	1.17	1,165.33(1)***
Outgroup ownership perceptions	2.74	1.67	303.78(1)***
Indirect shared ownership perceptions	2.59	1.35	445.61(.1)***
Direct shared ownership perceptions	3.71	1.63	12.51(1)***
Reconciliation intentions	3.14	1.32	154.33(1)***

***p < .001. All items were assessed on scales from 1 to 7. Wald tests testing for differences from the scale midpoint.

To get a sense of how many people scored high on both ingroup and outgroup ownership and thus held shared ownership perceptions, we calculated a mean score of the three variables assessing ingroup ownership perceptions, and a mean score of the three variables assessing outgroup ownership perceptions. We then recoded these observed mean scores so that all values below the midpoint 4 were treated as 0 = ‘low ownership’, and from 4 onwards as 1 = ‘high ownership’.4 Cross‐tabulating the two dummies resulted in four types of ownership perceptions: ‘no ownership’ of Kosovo by either group, ‘ingroup ownership only’, ‘outgroup ownership only’, and ‘shared ownership’ (see Table 2). While most participants were in the category of ingroup ownership (ranging from 71% of Serbs from Serbia to 93% of Kosovo Albanians), there was a portion of participants in the category of shared ownership perceptions, ranging from only 5% of Kosovo Albanians, to 9% of Serbs from Kosovo, and about a quarter of Serbs in Serbia. However, most respondents were more nuanced than Table 2 suggests. Only 29% of Albanians, 23% of Serbs from Kosovo, and 19% of Serbs from Serbia scored 7 (‘very much’) on ingroup ownership and 1 (‘not at all’) on outgroup ownership. This means that at least 70% of participants in each sample acknowledged outgroup ownership to some (even if small) degree.

Table 2

Distribution of ingroup and outgroup ownership perceptions of Kosovo, Study 1 (total N = 995)

	Low ingroup ownership perceptions (1–3)		High ingroup ownership perceptions (4–7)
	No ownership		Ingroup ownership
Low outgroup ownership perceptions (1–3)	Albanians from Kosovo	2.3%	Albanians from Kosovo	92.6%
	Serbs from Kosovo	1.0%	Serbs from Kosovo	89.5%
	Serbs from Serbia	1.2%	Serbs from Serbia	70.9%

We then inspected the latent scales for shared ownership perceptions both for the indirect and the direct measure. In all three groups, both means were below the midpoint (Table 1). The mean of the direct measure was, however, higher than that of the indirect measure (Albanians: Wald(1) = 184,61, p < .001; Serbs(Kosovo): Wald(1) = 111.24, p < .001; Serbs(Serbia): Wald(1) = 248.21, p < .001). Thus, it seems to be easier for participants to agree with items that explicitly ask about shared ownership than with items on outgroup ownership alone, which was used to assess the indirect shared ownership perceptions. The two measures correlated positively in all three groups (Table 3).

Table 3

Correlations between the main latent variables, Study 1 (total N = 995)

	2.	3.
Total sample (N = 995)
1. Indirect shared ownership perceptions	.58***	.28***
2. Direct shared ownership perceptions		.49***
3. Reconciliation intentions
Albanians (Kosovo; N = 390)
1. Indirect shared ownership perceptions	.51***	.22***
2. Direct shared ownership perceptions		.55***
3. Reconciliation intentions
Serbs (Kosovo; N = 200)
1. Indirect shared ownership perceptions	.61***	.30***
2. Direct shared ownership perceptions		.51***
3. Reconciliation intentions
Serbs (Serbia; N = 405)
1. Indirect shared ownership perceptions	.60***	.39***
2. Direct shared ownership perceptions		.52***
3. Reconciliation intentions

***p < .001.

Reconciliation intentions were on average below the neutral midpoint in all three groups (Table 1). Further, both the indirect and direct shared ownership perceptions were positively correlated with reconciliation intentions (see Table 3).

Explanatory findings

To test our hypothesis that shared ownership is related to stronger reconciliation intentions, we used structural equation modelling (SEM) with latent variables in Mplus version 8.0. For the sake of simplicity, we report the results from an analysis on a pooled sample (but see Online Supplement 5 for similar results from a multiple‐group SEM model). We estimated two models. In the first model, the willingness to reconcile was regressed on indirect shared ownership and in the second model on direct shared ownership. In both models, we controlled for ethnic group (with Serbs from Serbia as the reference category), religiosity, political orientation, age, gender, and the participants’ educational level.

As Model 1 in Table 4 shows, indirect shared ownership perceptions were positively related to reconciliation intentions (b = .19, SE = .02, p < .001).5 Furthermore, directly assessed shared ownership perceptions (Table 4, Model 2) were also significantly related to higher reconciliation intentions (b = .46, SE = .03, p < .001).6 To test for differences in the associations across groups, we additionally estimated models where we specified an interaction with ethnic group (contrasting Albanians (1) and Serbs from Kosovo (1) with Serbs from Serbia (0)) and shared ownership perceptions. The interaction terms with ethnic group were not significant both in the case of indirect and direct shared ownership, which confirms that the pattern of findings is similar for all three groups (see Online Supplement 4, Table S4.1 for the results, and Online Supplement 5, Tables S5.1 and S5.2 for multiple‐group analyses).

Table 4

Structural equation model explaining reconciliation intentions by an indirect (Model 1) and direct (Model 2) measure of shared ownership perceptions, unstandardized results, Study 1 (N = 995)

	Model 1		Model 2
	b	SE	b	SE
Indirect shared ownership perceptions	.38***	.05
Direct shared ownership perceptions			.46***	.03
Control variables
Religiosity	−.03	.03	−.03	.02
Political orientation (left‐right)	−.05	.05	−.02	.04
Age	.00	.00	.00	.00
Gender	.18*	.09	.15	.08
Educational level (ref.: Secondary)
Primary	−.25*	.12	−.19	.11
Tertiary	.42***	.11	.35***	.10
Ethnic group (ref. Serbs (Serbia))
Albanians	.59***	.11	.61***	.09
Serbs (Kosovo)	.21	.12	.39***	.11
Explained variance (R 2)	.17		.32

*p < .05; **p < .01; ***p < .001. Model fit Model 1: χ2(112) = 573.84, p < .001, RMSEA [90% CI] = 0.064 [0.059, 0.070], CFI = .902, SRMR = .054; Model fit Model 2: χ2(112) = 532.52, p < .001, RMSEA [90% CI] = 0.061 [0.056, 0.067], CFI = .919, SRMR = .053

Discussion

The results of Study 1 show that perceptions of ingroup ownership are very high and rather uniformly endorsed in Kosovo, whereas perceptions of outgroup ownership are much lower and there is more variation in its endorsement. Importantly, in all three subsamples, there are people who perceive ownership of Kosovo to be shared between Albanians and Serbs. About three quarters of participants see the outgroup as being entitled to the territory of Kosovo at least to some degree, and at least one in 20 participants thinks that both groups are rather entitled. These proportions seem to differ across ethnic groups, with Albanians having the least and Serbs in Serbia the most inclusive perceptions of ownership, while Serbs in Kosovo are falling in‐between. Further, the two measures of shared ownership perceptions were both associated with stronger reconciliation intentions. Altogether, these findings are in line with our central hypothesis that perceptions of shared ownership are related to stronger reconciliation intentions.

Study 2

In Study 2, we experimentally manipulated shared ownership of Kosovo in order to estimate the causal effect of shared ownership (as opposed to ingroup ownership) on reconciliation intentions.

Data for Study 2 were collected via research agency Kantar in collaboration with TNS Bulgaria in 2019. This was done simultaneously with the data collected for Study 1, and within the same sampling and data collection procedure. As the agencies could not guarantee a large enough sample of Serbs in Kosovo (the minority group) for both studies, we assigned Serbs in Kosovo to the correlational study only. Consequently, in the experimental study, we had to focus on the two larger and more accessible groups of Albanians from Kosovo and Serbs from Serbia. As in Study 1, we obtained random national samples of these two groups (see Online Supplement 1).

We aimed at comparing participants in two different experimental conditions (shared vs. ingroup ownership condition) using a t‐test. Power analyses in G*Power (Faul, Erdefelder, Lang, & Buchner, 2007) for an experimental study assuming a 2‐tailed p‐value and an equal number of participants in these two conditions indicated a minimum required sample size of 176 per condition in order to have sufficient power (.80) to detect a small to medium effect size (d = .3), with an alpha of .05. Thus we needed at least 352 Albanians and 352 Serbs to be able to test the hypothesis for each ethnic group separately. As we had no prior information on the effect size given the novelty of the study, we increased our target sample to 400 per ethnic group so that we would be able to detect even smaller effects and take into account possible dropouts.

Our sample consisted of 410 Albanians in Kosovo (49% men) and 399 Serbs in Serbia (54.6% men). Albanians were on average 43 (SD = 16.53) and Serbs were 45 years old (SD = 14.74). There were 417 participants in the shared ownership condition and 392 participants in the ingroup ownership condition. The two conditions did not differ from each other in terms of demographics, and Albanians and Serbs were equally distributed across the conditions (see Online Supplement 2, Table S2.2).

Procedure

We experimentally manipulated shared ownership perceptions of Kosovo by presenting half of the participants with a short narrative of Kosovo being owned together by the ingroup and the outgroup. The other half of the participants read a text in which ingroup ownership of Kosovo was highlighted. Both texts were created for the purpose of the experiment and no source was mentioned. The words in brackets were added when manipulating shared ownership:

Historians have researched the history of Albanians (and Serbs) in Kosovo in depth. There is clear evidence that (both) Albanians (and Serbs) have been living on the territory of Kosovo for centuries. Further, Albanians (both groups) have contributed to the development of the territory, and the Albanian (as well as Serbian) culture and identity are closely connected to Kosovo. For these and other reasons, one could say that Kosovo belongs (both) to Albanians (and Serbs).7

To ensure that participants read the text carefully and understood the conclusion about ownership, we included a comprehension check: ‘What was the main conclusion in the text above?’ with the answer options of 1 = ‘Kosovo has played an important role in the European history’; 2 = ‘Kosovo belongs to both Albanians and Serbs/Serbs and Albanians’ (in the shared ownership condition only) or 2 = ‘Kosovo belongs to Albanians/Serbs’ (in the ingroup ownership condition only); and 3 = ‘Kosovo is an ethnically mixed country’. Only 40% of the Albanian and 53% of the Serbian participants passed the comprehension check and we selected these for our main analysis (N = 375; but see Online Supplement 7 for additional analyses that includes also participants who responded with ‘Kosovo is an ethnically mixed country’ in the shared ownership condition).8 Thus, we retained 162 Albanian participants, (ingroup ownership condition N = 127; shared ownership condition N = 35) and 213 Serbian participants (ingroup ownership condition N = 144; shared ownership condition N = 69).9

Given this reduced sample size, we analysed participants in a pooled sample (N = 375; but see Online Supplement 6 for mean differences per ethnic group), meaning that we had 104 participants in the shared ownership condition and 271 participants in the ingroup ownership condition.

As a manipulation check and after having read and responded to the experiment, participants were asked how much they agreed with shared ownership of Kosovo by Albanians and Serbs, operationalized as in Study 1. The three items formed a reliable scale across ethic groups (ρAlbanians = .96; ρSerbs = .90) and experimental conditions (ρShared ownership = .87; ρIngroup ownership = .86), and we computed a mean score.

Reconciliation intentions were measured with the same seven items as in Study 1, rescaled into a mean score. The construct was reliable across ethnic groups (ρ Albanians = .91; ρSerbs = .92) and conditions (ρShared ownership = .90; ρIngroup ownership = .92).

In SPSS version 24, with an independent sample t‐test, we checked the manipulation by examining whether shared ownership perceptions differed between participants in the two experimental conditions. Participants who read about shared ownership agreed more with shared ownership (M = 4.00, SD = 1.64) than those who read about ingroup ownership (M = 3.30, SD = 1.57; t(373) = 3.80, p < .001, 95% CI: [0.34, 1.06]).10 We additionally tested whether the results across the two ethnic groups were comparable by regressing shared ownership on the experimental condition, ethnic group, and their interaction. The interaction term was not significant (b = −.23, SE = .38, p = .544, 95% CI: [−0.97, 0.51]). Thus, the manipulation worked similarly among Albanians and Serbs who responded to the comprehension check as intended. Importantly, participants who responded with ‘Kosovo is an ethnically mixed country’ to the comprehension check in the shared ownership condition did not differ in their perception of shared ownership from participants in the ingroup ownership condition (see Online Supplement 7).

Next, with an independent sample t‐test, we found confirmation for our hypothesis that participants in the shared ownership condition showed higher reconciliation intentions (M = 3.71, SD = 1.38) than those in the ingroup ownership condition (M = 3.15, SD = 1.44; see Figure 1), t(373) = 3.38, p = .001, 95% CI: [0.23, 0.88]. Additionally, in a regression analysis, the interaction term between ethnicity and condition was not significant (b = .07, SE = .34, p = .833, 95% CI: [−0.60, 0.75]), meaning that the manipulation had a similar effect on reconciliation intentions for Serbs and Albanians.

Figure 1

Mean differences in reconciliation intentions in the shared and ingroup ownership condition, Study 2 (N = 375). Note: **p < .01; Error bars indicate the standard errors of the means.

In Study 2, among both Albanians from Kosovo and Serbs from Serbia, we found that reading about shared ownership lead to stronger reconciliation intentions than reading about ingroup ownership. However, a substantial portion of the participants who were presented with the shared ownership narrative responded differently to the comprehension check than intended, and additional analyses (Online Supplement 7) showed that their sense of shared ownership was comparable to that of participants who were presented with the ingroup ownership narrative. It could be the case that some participants simply did not take the time to read the text of the manipulation. Still, the relatively high proportion of non‐intended answers to the comprehension check in the shared‐ compared to ingroup‐ownership condition suggests that in the former, the participants might have focused more on the mention of both ethnic groups and overlooked the mention of shared ownership. Lastly, they might have chosen that other reading check as a plausible alternative because they rejected the idea of shared ownership. The latter would then mean that it is difficult to move some people in the direction of reconciliation by using the shared ownership rhetoric.

General discussion

Perceiving ownership of disputed territory to be shared between two rival groups might be an important prerequisite for solving territorial conflicts. However, we do not know to what extent such shared ownership perceptions exist among inhabitants of conflict regions, and the question whether shared ownership indeed relates to more reconciliation intentions has to our knowledge not received attention so far. We addressed this question in two studies – one correlational and one experimental – in post‐conflict Kosovo, using random national samples of Albanians and Serbs from Kosovo as well as Serbs from Serbia, who have a vested interest in the territory of Kosovo.

In Study 1, we found that a sense of ingroup ownership is strongly present and that outgroup ownership is much less recognized. Still, at least 5% of participants from each ethnic group clearly saw both Albanians and Serbs as owners of Kosovo, and even 70% perceived that the outgroup was at least somewhat entitled to the territory. Using an indirect as well as a direct measure of shared ownership, we further found rather low recognition of shared ownership, and particularly among the two groups living in Kosovo. Serbs in Serbia might be more open to the idea of shared ownership because they do not live on the disputed territory and might have a rather symbolic interest in the ownership of Kosovo. Previous research has shown that Serbs who live in areas of recent conflict have a stronger ethos of conflict than Serbs who live outside such areas (Stankov, Knežević, Petrović, Međedović, & Lazarević, 2019), and a nationally representative survey showed that Serbs in Serbia worry more about issues such as corruption or health rather than the Kosovo conflict (Zivanovic, 2018).

Even though the idea of shared ownership was generally not very much endorsed, in Study 1, greater perceptions of shared ownership were associated with stronger intentions to reconcile relations between Albanians and Serbs in Kosovo. This was found among all three groups of participants, and using both measures of shared ownership. Furthermore, in the experimental Study 2, we provided first evidence for the causal effect of shared ownership on reconciliation intentions. Reading about shared ownership of Kosovo elicited stronger reconciliation intentions among Albanians in Kosovo and Serbs in Serbia alike. These findings are in line with our expectations derived from theories on collective psychological ownership (Pierce & Jussila, 2010; Verkuyten & Martinović, 2017), and they resonate with the findings from Rosler et al. (2018) about the importance of acknowledging outgroup ownership for intergroup relations.

However, about half of the participants in Study 2 who read the narrative about shared ownership responded to the comprehension check other than intended, and we managed to trigger a sense of shared ownership only among those who passed this check. It is possible that our story about shared ownership for a substantial portion of participants primarily made the ethnic diversity of Kosovo salient rather than shared ownership, which is why they chose that alternative reading check. However, it can also be difficult to convince people of something they do not believe in (Shermer, 2017), and it could just as well be the case that a portion of our participants simply disagreed with the shared ownership narrative and refused to accept the conclusion that Kosovo belongs to both Serbs and Albanians. Instead, they chose the other meaningful response ‘Kosovo is an ethnically mixed country’ to the comprehension check, disproportionately so compared with those in the ingroup ownership condition. Perhaps it was more acceptable to acknowledge that both groups live in Kosovo than that both groups own the territory. Future research is needed to evaluate further how participants react to shared ownership narratives in contexts where ownership is disputed.

Our work has several limitations. First, there was very little variation in ingroup ownership perceptions (Study 1), which is why our indirect measure of shared ownership primarily reflected agreement with outgroup ownership. Future studies could consider the role of shared ownership in relation to territories that are less contested and where ingroup ownership is not necessarily the default. This would help discover whether perceptions of sharedness have different consequences than perceptions of outgroup ownership.

Second, we did not conduct Study 2 among Serbs from Kosovo, who might be more concerned with the Kosovo conflict than Serbs from Serbia and react differently to the shared ownership narrative. Moreover, Serbs in Kosovo represent a numerical minority for whom a shared ownership narrative might have a different effect than for the majority of Albanians in Kosovo. However, Serbs in Kosovo were included in the correlational Study 1 and we found no differences across groups. We would therefore expect to detect a similar effect of the experiment in Study 2 for Serbs in Kosovo. Future research could try to replicate the experiment among them to discover whether shared ownership can be manipulated in a similar way among a minority group. Additionally, one could vary the source of the narrative and explore reactions to the story that comes from an ingroup versus outgroup historian.

Third, in the ingroup ownership narrative (Study 2) it was not mentioned explicitly that the outgroup does not own Kosovo, only that the ingroup owns it. For ethical reasons, we could not deny outgroup ownership in the text. Our ingroup ownership manipulation might therefore also have evoked perceptions of outgroup ownership. However, in such contested contexts, as we have shown in Study 1, few people have pronounced outgroup ownership perceptions in the first place. If anything, it is likely that the effects of the shared ownership narrative that we found would have been even stronger had we, in the contrasting narrative, along with emphasizing ingroup ownership, denied outgroup ownership.

Finally, our experiment did not include a control group. We therefore do not know whether we decreased shared ownership perceptions in the ingroup ownership condition or increased them in the shared ownership condition. However, looking at the means across our two studies, treating the correlational Study 1 as a baseline, it seems that average agreement with shared ownership in Study 1 is comparable to that in the ingroup ownership condition of Study 2. Thus, it is likely that the shared ownership narrative increased shared ownership perceptions while the ingroup ownership narrative did not move participants’ perceptions of shared ownership. An experimental study with a control condition could more formally confirm these conclusions.

Conclusion

We have shown that in a territorial conflict setting people can perceive that the territory is owned together by the two rival groups and that such shared ownership perceptions are positively related to reconciliation intentions. Importantly, we have also provided first evidence that it is possible to increase people’s reconciliation intentions by presenting them with a narrative of shared ownership of the contested land. While previous studies have considered factors such as victimhood perceptions (e.g., Vollhardt & Bilali, 2015), emotions (e.g., Halperin, 2011) and contact interventions (e.g., Maoz, 2011), the findings presented in this paper demonstrate that it is relevant to also acknowledge land ownership perceptions when studying reconciliation intentions among group members in territorial conflict areas. Furthermore, while previous research on collective psychological ownership showed the relevance of ingroup ownership in understanding intergroup relations in undisputed contexts (Nijs et al., 2021; Torunczyk‐Ruiz & Martinović, 2020), we have provided first empirical evidence that in territorial conflicts, it is important to consider in how far group members recognize territorial ownership of each group involved in the conflict, not just of their own group. The present research gives a first indication that it is possible to promote the recognition of shared ownership, even when territorial ownership is contested. While further research is needed, the findings give rise to opportunities for promoting positive intergroup relations. For societies in conflict, it might be worth to teach the history of conflict in schools in more nuanced ways, so that it is conveyed that, next to the ingroup, the outgroup also has (legitimate) reasons to claim ownership of the territory. This could be achieved even in ethnically separated classrooms, which are common in conflict regions. The same can be done in the public discourse on territorial ownership.

Funding information

This research has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 715842) awarded to the second author.

Conflicts of interest

All authors report no conflict of interest.

Author contributions

Nora Storz (Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Writing – original draft), Borja Martinović (Conceptualization; Data curation; Funding acquisition; Supervision; Writing – review and editing), Edona Maloku (Writing – review and editing), Iris Žeželj (Writing – review and editing).

Supinfo S1. Sampling method, Study 1 and Study 2.

Click here for additional data file.