Mixed and worried? Examining the association between locality type and worrying about social phenomena in Arab citizens of Israel

Abstract

This study examined the association between the type of locality ethnic minorities reside in and their worrying about social phenomena through the lens of the social capital theory. The study aimed at showing that worry profiles, as well the extent of being concerned about various social phenomena, are not the same for the minority residents of ethnically homogeneous and of ethnically heterogeneous localities. Moreover, the main assumption was that residing in heterogeneous locality corresponds to an increase in concerns about social phenomena. The data were obtained from the Personal and Community Security Index Survey which was conducted among the Arab citizens of Israel in 2020 (N = 947). The data were analyzed using multinomial and linear regression models. The results show that residing in ethnically mixed localities was associated with having a highly worried profile. Residing in mixed localities also related to a greater worry about the general social and the violence-related phenomena. The results supported the social capital theory. The results suggest that residing in ethnically homogeneous localities plays a protective role with respect to being concerned about social phenomena. The results imply that public decision makers should develop long-term plans to reduce worry levels in minority residents of ethnically mixed localities which will include provision of tailored services, fighting against violence, and informing about the proposed solutions for each phenomenon of concern.

Keywords:

• Arab society
• ethnic minority
• mixed cities
• worry

Introduction

Worry is an “emotional reaction, an unpleasant state of mind, preoccupation with thoughts about uncertain and unpleasant events ranging from personal issues such as health or finance to broader issues such as social or technological change and environmental pollution” (Akinbobola 2012, p. 167). People tend to worry (sometimes simultaneously) about numerous issues: chronic health condition (Scherr, Jensen, and Christy 2017), environment (Akinbobola 2012), victimization (Marzbali et al. 2012; Hart 2017), crime (Jackson 2011), terrorism (Rashid and Olofsson 2021) and more. Recently, worrying about COVID-19 (Schulz et al. 2021) or its various consequences (Taylor et al. 2020) has joined this list. In these and other contexts, or regardless of them, one can worry in response to an immediate situation (proximal threat) and/or a situation that is yet to take place (distal threat) (Jackson and Kuha 2014). Similarly, one can worry about oneself and/or others (Taylor et al. 2020; Schulz et al. 2021).

Worry is functional for every individual in any socieity as it can improve cognitive planning necessary for coping with threats and upgrading task performance (Castillo et al. 2010), and stimulate action (Jackson and Gray 2010). Experience of worrying may assist people in their coping with uncertainties (Jackson and Gray 2010). Empirical research has shown that individuals with greater COVID-19 concern, both personal and vicarious, were more likely to practice social distancing compared to those who were less concerned about COVID-19 (Vaterlaus, Spruance, and Patten 2021; Bouman and Steg 2023). The extent of worry was found positively relating to the intention to vaccinate against H1N1 (Scherr, Jensen, and Christy 2017) and COVID-19 (Faasse and Newby 2020). However, worry can also be high and sometimes even go out of control therefore becoming pathological (Castillo et al. 2010). It was found that the more frequent the worry, the greater is the negative impact on quality of life (Jackson and Kuha 2014). Similarly, cancer worry was found associating with reports of doctor avoidance (Persoskie, Ferrer, and Klein 2014). Greater extent of modern health worries was associated with poorer perceptions of health status and reduced frequency of medication use (Filipkowski et al. 2010). Finally, COVID-19 worry was found positively relating to a severity of mental health issues (Barzilay et al. 2020; Bergman et al. 2021; Boateng et al. 2021).

Numerous factors can explain the extent of worrying about particular phenomena. One factor that was rarely studied in the current regard is the type of residential locality. Ethnic minority populations typically reside in either homogeneous localities, where they constitute the (absolute) majority or the total population, or in mixed localities where they constitute a minority share of the population. Members of ethnic minority residing in these types of localities may have varying levels of social capital, of ability to mobilize it for individual goals, including those aimed at minimizing risks and coping with uncertainty, and perhaps even need different types of social capital to address worries and concerns. Hence, members of ethnic minority residing in localities of these two types may exhibit different worry profiles and varying levels of worrying about social issues.

Israel is a good setting for studying the abovementioned link. First, its population is comprised, inter alia, of a notable Arab minority, representing about 18% of the country’s population (‘Ali, 2019). Second, members of Arab minority in this country reside in both ethnically homogeneous and ethnically heterogeneous (mixed) localities (Shdema and Martin 2022).

Understanding worries as a social phenomenon is important because of their potential long-term consequences, particularly for the mental health and wellbeing of a minority population. In general, Israeli studies show that Arab population consistently reports lower scores than Jewish population on various mental health and wellbeing indicators: perceived happiness (Baron-Epel et al. 2008), depression and anxiety (Braun-Lewensohn, Abu-Kaf, and Kalagy 2021), social capital (Paldi et al. 2021), and resilience (Kimhi et al. 2020). In addition, findings of the international research suggest that worry is a risk factor for mental health problems (Mayorga et al. 2022). COVID-19 worry was significantly and positively related to depression and/or anxiety (Barzilay et al. 2020; Bergman et al. 2021; Boateng et al. 2021; Gilbar et al. 2022; Mayorga et al. 2022), as well as stress (Mayorga et al. 2022), loneliness (Horesh, Kapel Lev‐Ari, and Hasson‐Ohayon 2020), decreased quality of sleep (Gao and Scullin 2020) and more.

Studying the link between type of locality and worries is also of high importance from the practical aspect. The findings of this study will map the worry outcomes according to a type of locality and propose an intervention in order to improve wellbeing and reduce the levels of concerns in an ethnic minority population.

Homogeneous and mixed localities

Before approaching the study’s main theoretical framework, we briefly discuss the notion of a locality type. About 90% of Arab population of Israel resides in homogeneous localities (Shdema and Martin 2022). There are about 120 of such localities over the country, of which only about 10% are midsize cities—rest are small towns and villages (Alfasi 2014). Arab localities are marked by a relatively poor socio-economic situation (Birenbaum-Carmeli and Chassida 2020) while many of them are suffering from a lack of basic infrastructure such as public spaces or paved streets, as well as a slow but a steady invasion of industry and commerce into previously residential-only areas (Alfasi 2014). State authorities do not invest much efforts into a proper planning of Arab localities. Therefore, urbanization and other transitional processes taking place in them are rather latent than planned (Kallus 2013). Consequently, Arab towns have preserved many elements of family-based structure, with notable reproduction of rural orders (Totry-Fakhoury and Alfasi 2017). Arab villages consist of several clans (called hamulahs), each residing in their own limited but well-defined areas (Alfasi 2014).

About 8% of Israel’s Arab population resides in mixed cities (Shdema and Martin 2022). This term refers to localities with predominantly Jewish residents and a notable share of Arab residents residing alongside each other (Goldblatt and Omer 2014). To date, the State of Israel has officially eight localities as mixed cities: Akko, Haifa, Lod, Maalot-Tarshiha, Nof HaGalil, Jerusalem, Ramle, and Tel-Aviv-Jaffa (‘Ali, Lewin-Chen, and Najami Yousef 2021; Inter‐Agency Task Force on Israeli Arab Issues 2021). Jerusalem, although officially defined as mixed city, is better viewed as divided city (Diab, Shdema, and Schnell 2022) since the segregation between Jewish and Arab population there is almost complete (Inter‐Agency Task Force on Israeli Arab Issues 2021). Some of the mixed cities are veteran, like Haifa, Akko, Tel-Aviv-Jaffa, Lod and Ramle, as they were established before 1948 (Blatman-Thomas 2017). Another part of them are newly established, like Nof haGalil (Diab, Shdema, and Schnell 2022). These were initially Jewish localities which have started to be populated by Arab families since 1980s (Blatman-Thomas 2017) because of improvements in standard of living, scarcity of land available for construction in the nearby Arab localities, and a willingness to escape from the traditionalism dominating in home localities (Diab, Shdema, and Schnell 2022).

Mixed cities differ between them in terms of the share of their Arab residents: from about 11% in Haifa (Diab, Shdema, and Schnell 2022) to about a third in Tel-Aviv-Jaffa (Goldblatt and Omer 2014). Also, types of spatial location of the two ethnic groups: from highly segregated Maalot-Tarshiha (Falah 1996) to relatively lowly segregated Nof Ha-Galil (Blatman-Thomas 2017). Finally, the nature of the Jewish-Arab relationship in these localities varies: from cooperative interethnic relations in Haifa (Kallus 2013) through a low tolerance of the Arab presence in Nof HaGalil (Diab, Shdema, and Schnell 2022) to the negative interethnic relations in Ramle, Lod, and Tel-Aviv-Jaffa (Abraham Initiatives 2020).

Although Israeli research on mixed cities in various scientific domains is growing (Goldblatt and Omer 2014; Shdema, Haj-Yahya, and Schnell 2018; Diab, Shdema, and Schnell 2022), no studies have so far investigated locality differences in worrying about various social phenomena. The studies on worry conducted in Israel (Horesh, Kapel Lev‐Ari, and Hasson‐Ohayon 2020; Bergman et al. 2021; Gilbar et al. 2022) also did not address the difference between types of localities. Nor did the international research studying worries or concerns (Jackson 2011; Hart 2017; Rashid and Olofsson 2021; Schulz et al. 2021). Finally, rare are the studies which differentiated between various types of worries or concerns (Hart 2017; Taylor et al. 2020) and examined typologies of individuals in accordance with their levels of worry (Jackson and Gray 2010; Tsouros et al. 2021). For all these reasons, the current study can be seen as a novelty in the field.

Literature review

Social capital

Social capital is a widely studied concept in social sciences. One of its definitions, coined by Bourdieu, sees social capital as “the total resources, feasible or potential, that an individual or a group accumulates by means of the constant maintenance of social networks or reciprocal social interactions” (Amit and Riss 2007, pp. 290–291). Nan Lin defined social capital as “resources embedded in a social structure that are accessed and/or mobilized in purposive actions” (Kindler and Wójcikowska-Baniak 2019, p. 97).

Social capital is seen as a heterogeneous entity. The reason for such an approach to social capital is that it is accessed and mobilized via different types of social networks, which involve actors situated at different positions in the stratification system of a given society. Bonding social capital refers to resources embedded in the in-group networks (Gong, Xu, and Wang 2021). In such networks individuals are connected with each other on the basis of common ascribed characteristics, values and norms (family, close friends, and people believing in or valuing the same things). In contrast, bridging social capital refers to connections across groups, therefore being essential for establishing integroup tolerance and acceptance (Chu and Yang 2019). Finally, linking social capital refers to resources that can be acquired by means of active participation in group activities or interaction with people or institutions that can assist in leveraging these resources (Daoud et al. 2017). Bridging social capital is leveraged from the connections between people or groups holding similar social status, whereas linking social capital is leveraged from the relationship with people or groups of different social status or influence (Elgar et al. 2011).

Social capital and locality

Potential amount of social capital differs between localities of various types mainly because the nature of social relations is different in each type of locality. For example, in more rural areas, social ties are based on knit bonds and relatively close connections. In more urban areas, the nature of social ties is transitory and less personal (Lancee and Seibel 2014). This issue is critical in the ethnically mixed localities in Israel which tend to be residentially segregated (Shdema, Haj-Yahya, and Schnell 2018). Spatial segregation can result in unequal access to social capital mainly because of the tendency of a higher social status people (for instance, ethnic majority) to preserve homogeneity in their neighbourhood and distant themselves from people of lower social status (e.g., ethnic minority) (Otero, Volker, and Rozer 2022).

Social capital, mental health and wellbeing

Social capital is correlated with various health indicators (Baron-Epel et al. 2008). In numerous studies, social capital exhibited positive relation to psychological wellbeing (Tegegne and Glanville 2019; Gong, Xu, and Wang 2021), life satisfaction/happiness (Arpino and de Valk 2018; Jiang and Renema 2021), and mental health (Paldi et al. 2021). The explanation for this correlation is typically attributed to the extent of engagement in social life. Social relationships tend to provide access to resources. Social relationships tend to contribute to increased sense of belonging, sense of social control, and self-esteem. They are able as well to assist in coping with the everyday life uncertainties and threats by providing instrumental and/or expressive support (Tegegne and Glanville 2019).

Minorities, localities, social capital, and worry

Why would members of ethnic minority residing in localities of different types exhibit varying extents of worrying about various social phenomena? As previously mentioned, various localities represent different type of social capital. Bonding social capital, which typically characterizes ethnically homogeneous localities (Amit and Riss 2014), provides sense of belonging, predictability, security, and stability. All these can benefit mental health (Johnson et al. 2017). Bridging social capital, typically attributed to ethnically heterogeneous localities (Amit and Riss 2014), positively impacts mental health through enabling information and resource exchange between social groups and facilitating social mobility (Johnson et al. 2017). Linking social capital may also positively impact mental health as this capital reflects confidence in public/state institutions. It was found that linking social capital is positively associated with both self-rated health and life satisfaction (Elgar et al. 2011).

Despite the advantages of all three types of social capital, bonding social capital seems to be more benefitting than bridging/linking ones for the wellbeing of the members of ethnic minorities. One study found that ethnically homogeneous neighborhoods reported a higher level of satisfaction with their lives than those living in ethnically heterogeneous neighborhoods (Neto 2001). Mixed localities in Israel are all defined as cities (Diab, Shdema, and Schnell 2022), whereas many of homogeneous localities are defined as small urban or rural. Some other studies found that rural residence, which is typically characterized by bonding social capital (Lancee and Seibel 2014), was associated with greater subjective wellbeing (Tegegne and Glanville 2019; Rosenberg and Taipale 2022).

Based on this review, the following hypotheses are proposed:

• H1: Ethnic minority residents of ethnically mixed localities will have poorer worry profile than their counterparts residing in ethnically homogeneous localities.

• H2: Ethnic minority residents of ethnically mixed localities will be more worried about social phenomena than their counterparts residing in ethnically homogeneous localities.

Additional factors

Evaluation of violence

In accordance with social disorganization theory, people in disorganized localities typically have little desire to contribute to the establishment of the informal social control inside them (Silver and Miller 2004). Localities characterized by perceived increase in violence are likely to be characterized also by provision of low institutional support for conventional values (Corsaro, Frank, and Ozer 2015). Greater perceived likelihood of crime was found corresponding to higher frequency of worrying about crime (Jackson 2011). In a similar vein, we expect that perception of increase in the level of violence will be positively associated with the worry outcomes.

Prior victimization

Victimization experience generally makes people feel vulnerable (Zhao et al. 2014). In one study, victimization experience significantly predicted dysfunctional worry (Jackson and Gray 2010). Hence, we expect positive association between prior victimization and worrying in the studied context.

The study also controlled for socio-demographic background. As for gender, research shows that women scored higher than men on general worry (Castillo et al. 2010), COVID-19 worries (Barzilay et al. 2020; Horesh, Kapel Lev‐Ari, and Hasson‐Ohayon 2020), and worry about crime (Hart 2017). As for age, individuals who were older endorsed greater COVID-19 worry (Barzilay et al. 2020). Finally, another study found that people with higher education and those unmarried were less worried about terrorism than people with lower education and married (Rashid and Olofsson 2021). The current study expects to find similar differences and associations with respect to the worry outcomes.

Methods

Data and sample

The data for the current study were obtained from the 2020 Personal and Community Security Survey conducted in Arab population of Israel. The survey instrument was developed by the research team headed by the principal author of the manuscript. It contained approximately 110 items (including socio-demographic data) on various aspects of violence, institutional performance evaluations, relationship with the police, and several COVID-19-related items.

The sample included Arab citizens residing inside the Green Line in Israel. The survey sampling was performed in two stages. First, 28 localities, a representative cross-section of the Arab localities in terms of geographical area, size, municipal status, socio-economic cluster, and religious and ethnic composition, were sampled. Each locality was assigned a quota of interviewees according to its size. Second, respondents were randomly sampled in all the localities included in the sample from the Ministry of the Interior’s Registry of Residents.

In 2020, as opposed to the previous years of the survey, two separate surveys, using identical instruments, were conducted. One was a nationwide survey conducted using the abovedescribed sampling strategy. Another one was conducted in six mixed localities only (Akko, Haifa, Lod, Maalot-Tarshiha, Ramla, Tel-Aviv-Jaffa). In total, about 1230 people were contacted for participation in the survey. Of those, 1127 participated (723—in nationwide survey and 304—in the mixed cities survey). Most of the non-response in the nationwide survey stemmed from technical reasons (i.e. incorrect home address provided, potential respondent worked outside her/his locality of residence, potential respondent was infected with the coronavirus) and only a small part—due to a refusal to participate in the survey. Those who provided their consent for participation were interviewed face-to-face by Arab speaking interviewers at the participants’ homes.

Of the participants, 1004 (700—from the nationwide, and 304—from the mixed localities survey) provided valid responses to the items asking on the extent of worry about 13 social phenomena (see Supplementary Table 1). However, since the nationwide sample included not only people who resided in ethnically homogeneous localities but also residents of ethnically mixed localities (n = 57), we had to exclude the latter in order to ensure the independency of the samples. Following this procedure, nationwide data included only the entries of the respondents residing in ethnically homogeneous localities. The analytical sample included entries of 947 survey participants.

Measures

Dependent variables

Worry profile was assessed as a categorical variable consisting of three categories: highly worried, moderately worried, and lowly worried. This typology is similar to that employed in another study (Tsouros et al. 2021).

Worry about the general social phenomena was measured continuously by taking the mean of responses to the four items asking about being worried (from ‘1′ – to a very low extent to ‘5′ – to a very high extent) of unemployment, road accidents, property crimes, and COVID-19 spread in Arab society (α = 0.758). The five-option scale measuring worries/concerns was also employed in other studies (Faasse and Newby 2020; Gilbar et al. 2022).

Worry about violence-related phenomena was measured continuously by taking the mean of responses to the six items asking about being worried (from ‘1′ – to a very low extent to ‘5′ – to a very high extent) of violence, crime, drugs, protection fees, shooting and weapon use, and quarrels (α = 0.899).

Worry about political phenomena - was measured continuously by taking the mean of responses to the two items asking about being worried (from ‘1′ – to a very low extent to ‘5′ – to a very high extent) about the current state of democracy in Israel, inequality between Jews and Arabs within the country, and the stagnation in the peace process between Israel and Palestinians (α = 0.855).

Main independent variable

Locality type – a dichotomous variable, with respondents who were surveyed in the mixed cities survey as the study (“1”) category, and respondents who participated in the nationwide survey as the reference (“0”) category.

Additional independent variables

Perceived locality violence change. The original item, which assessed the perception of change in the level of violence in the respondents’ localities compared to the year prior to the survey (‘sharp increase’, ‘increase’, ‘decrease’, ‘sharp decrease’, ‘no change’), was dichotomized. Respondents, who reported any extent of decline in the level of violence, as well as the lack of change, represented the reference category.

Prior victimization was measured as the mean of scores on three items assessing the extent (from 1 – ‘very low’, to 5 – ‘very high’) to which respondents, their friends or relatives, were offended or fallen victims of the following violent phenomena during the year prior to the survey: (i) curses, spits, threats, blackmail, (ii) use of weapons, and (iii) mass rampage (α = 0.861). Higher scores represented a greater extent of victimization.

Socio-demographic background: Gender (0 = Women, 1 = Men); Age (in years); Level of education (0 = Less than tertiary education, 1 = Tertiary education). Marital status (0 = Unmarried for any reason (single, divorced, or widowed), 1 = Married).

Data analysis

To create the worry profile variable, k-means cluster analysis was performed using the data on all the 14 survey items asking about being worried of various social phenomena. The results of the analysis and the ANOVA test suggested that the three-group clustering was the best solution for the data as only in this situation the clusters significantly differed between them with respect to each item (see Supplementary Table 1).

To create the rest of the dependent variables, factor analysis using principal component procedure and Varimax rotation was performed (see Supplementary Table 2). The values of Kaiser-Meyer-Olkin measure of sampling adequacy and Bartlett’s test of sphericity indicate that it was appropriate to perform this analysis (KMO = 0.912, χ²(78) = 6924.15, p < 0.001). The loading threshold was set at 0.5 (Lemire, Sicotte, and Paré 2008). The analysis has yielded three factors explaining 67.06% of the variance. One item—worry about the violence against women—was not loaded on any factor. Another (and a final) round of the analysis, performed with the remaining 13 items, has yielded again three factors explaining 68.79% of the variance. Each factor, as described in the Measures, reflected a separate outcome variable.

Following the trichotomous nature of the worry profile variable, multinomial regression analysis was performed. The category of lowly worried served as a reference category. Following the continuous nature of the rest of the outcome variables, series of linear regression models were analysed. Multicollinearity diagnostics were performed in each one of them. In all models, the values of variance inflation factor and tolerance were well below the critical values mentioned in the literature (Tabachnik and Fidell 2007). Specifically, all variance inflation factors (VIF) values were lower than 1.5, and tolerance values were higher than 0.7. Therefore, multicollinearity was not an issue in the analysis. All the analyses were performed using SPSS V.23.

Results

Descriptive statistics

Sample statistics are provided in Table 1.

Table 1. Sample statistics – independent variables and socio-demographic background.

	n	% or M	SD
Locality type	947
Ethnically mixed	304	32.1
Ethnically homogeneous	643	67.9
Perceived locality violence change	917
Increased	647	70.6
Otherwise (decreased/did not change)	270	29.4
Prior victimization (1–5)	947	1.7	.9
Gender	935
Male	485	51.9
Female	450	48.1
Age (18–79)	945	40.3	14.9
Level of education	942
Tertiary	320	34.0
Non-tertiary	622	66.0
Marital status	946
Married	630	66.6
Unmarried	316	33.4

Note. % = percentage of respondents in each category; M = Mean; n = Number of cases in each category; SD = Standard deviation.

The sample was predominantly male (51.9%). The average age was 40.3 years (SD = 14.9). Most of the respondents (66%) had non-tertiary education and were married (66.6%). The sample was characterized by a relatively low level of victimization (M = 1.7, SD = 0.9). Most of the respondents (70.6%) reported that compared to the previous year, (any level of) increase in the level of violence took place in their localities.

The outcome variable statistics are shown in Table 2.

Table 2. Sample statistics – dependent variables.

	n	% or M	SD
Worry profile	921
Highly worried	429	46.6
Moderately worried	402	43.6
Lowly worried	90	9.8
Worry about the general social phenomena (1–5)	947	3.8	.8
Worry about the violence-related phenomena (1–5)	947	3.9	.8
Worry about the political phenomena (1–5)	947	3.2	.9

Note. % = percentage of respondents in each category; M = Mean; n = Number of cases in each category; SD = Standard deviation.

Of the sample, 46.6% were clustered in the highly worried, 43.6% - in the moderately worried, and 9.8% - in the lowly worried group. The sample was characterised by moderate-up-to high extent of worrying about the violence-related (M = 3.9, SD = 0.8) and the general social phenomena (M = 3.8, SD = 0.8), and somewhat lower extent of worrying about the political phenomena (M = 3.2, SD = 0.9).

Predicting the worry profile

Table 3 shows the results of the analysis predicting the worry profile. The model was significant (χ²(14) = 84.91, p < 0.001), exhibited a good fit to the data (Pearson χ² (1502) = 1504.53, p = 0.477), and explained 10.9% of the dependent variable’s variance.

Table 3. Results of the multinomial regression analysis predicting the likelihood of belonging to worry profile.

Variable	Estimate	SE	RRR	95% CI for RRR		p
				LL	UL
High worry
Intercept	.38	.49				.433
Mixed localities^a	.78	.29	2.17	1.23	3.85	.008
Locality violence increased^b	.81	.26	2.25	1.36	3.72	.002
Prior victimization	.54	.18	1.71	1.22	2.41	.002
Male^c	−0.68	.26	.51	.31	.84	.008
Age	−0.01	.01	.98	.97	1.004	.123
Tertiary education^d	.25	.27	1.29	.76	2.18	.344
Married^e	.54	.30	1.72	.96	3.07	.069
Moderate worry
Intercept	1.65	.48				.000
Mixed localities^a	.10	.30	1.11	.62	1.97	.736
Locality violence increased^b	.39	.25	1.47	.90	2.39	.121
Prior victimization	.37	.18	1.44	1.02	2.04	.037
Male^c	−0.72	.25	.49	.30	.80	.004
Age	−0.03	.009	.97	.96	.99	.003
Tertiary education^d	−0.16	.27	.85	.51	1.44	.556
Married^e	.86	.30	2.35	1.32	4.21	.044
−2log likelihood	1479.65
Nagelkerke R	.109
N	873

Note. Reference categories:

^aHomogeneous localities.

^bDecreased or remained unchanged.

^cFemale.

^dNon-tertiary education.

^eUnmarried.

CI = Confidence Interval, LL = Lower Limit, RRR = Relative Risk Ratio, SE = Standard Error, UL = Upper Limit.

The results show that residing in ethnically mixed locality was positively associated with belonging to a highly worried cluster (RRR = 2.17, p = 0.008). This means that residents of mixed localities were 117% more likely than residents of homogeneous localities to belong to this cluster. However, no association was found between locality type and belonging to a moderately worried cluster (RRR = 1.11, p = 0.736). Therefore, Hypothesis 1 was partially confirmed.

As to the rest of the associations, perception of increase in locality violence was positively associated with belonging to the highly worried cluster (RRR = 2.25, p = 0.002). Victimization was positively associated with belonging to both highly (RRR = 1.71, p = 0.002) and moderately (RRR = 1.44, p = 0.037) worried cluster. In contrast, male gender was negatively associated with belonging to each of these clusters (highly worried: RRR = 0.51, p = 0.008; moderately worried: RRR = 0.49, p = 0.004). Finally, age was negatively (RRR = 0.97, p = 0.003), but being married was positively (RRR = 2.35, p = 0.044), associated with belonging to the moderately worried cluster.

Predicting the worry about social phenomena

Table 4 shows the results of the analysis predicting the extent of being worried about each of the three types of social phenomena. The model predicting the extent of being worried about the general social phenomena was significant (F(7;891) = 7.39, p < 0.001) and explained 5.5% of the dependent variable’s variance. The model predicting the extent of being worried about the violence-related phenomena was also significant (F(7;891) = 14.48, p < 0.001) and explained 10.2% in the variance of the dependent variable. The model predicting the extent of being worried about the political phenomena was also significant (F(7;891) = 5.44, p < 0.001) and explained 4.1% in the variance of the dependent variable.

Table 4. Results of the linear regression analysis predicting the extent of worrying about the social phenomena (N = 899).

	General phenomena		Violence-related phenomena		Political phenomena
Predictors	b (SE)	p	b (SE)	p	b (SE)	p
Intercept	3.5 (.10)	.000	3.33 (.10)	.000	2.99 (.12)	.000
Mixed localities^a	.32 (.06)	000	.24 (.06)	.000	−0.06 (.07)	.349
Locality violence increased^b	.11 (.06)	.043	.34 (.06)	.000	.10 (.07)	.148
Prior victimization	.03 (.03)	.264	.13 (.03)	.000	.001 (.03)	.966
Male^c	−0.10 (.05)	.059	−0.12 (.05)	.031	.01 (.06)	.876
Age	.001 (.002)	.630	.000 (.002)	.923	.01 (.002)	.024
Tertiary education^d	.09 (.05)	.109	.12 (.06)	.033	.35 (.06)	.000
Married^e	.03 (.06)	.601	.002 (.07)	.978	−0.21 (.07)	.006
Fisher exact ratio	7.39	.000	14.48	.000	5.44	.000
R^2 (non-adjusted)	.055		.102		.041

Note. Reference categories:

^aHomogeneous localities.

^bDecreased or remained unchanged.

^cFemale.

^dNon-tertiary education.

^eUnmarried.

b = Unstandardized coefficient; p = Significance value; SE = Standard error.

Significant findings are marked with bold.

The results show significant association between locality type and the extent of worrying about general social (b = 0.32, p < 0.001) and violence-related (b = 0.24, p < 0.001) phenomena. Residents of mixed localities exhibited greater concern about each one of these two types of social phenomena than residents of homogeneous localities. In contrast, locality type was unrelated to the extent of worrying about the political phenomena (b = −0.06, p = 0.349). This means that residents of ethnically mixed localities did not differ in their extent of being worried about the political phenomena from residents of ethnically homogeneous localities. Therefore, also Hypothesis 2 was partially confirmed.

As to the rest of the significant associations, perception of increase in the locality violence was positively associated with the extent of worrying about the general social (b = 0.11, p = 0.043) and the violence-related (b = 0.34, p < 0.001) phenomena. Prior victimization was positively (b = 0.13, p < 0.001) and male gender was negatively (b = −0.12, p = 0.031) associated with the extent of worrying about the violence-related phenomena. Age was positively (b = 0.01, p = 0.024) and being married was negatively (b = −0.21, p = 0.006) related to the extent of worrying about the political phenomena. Finally, having tertiary education was positively associated with the extent of worrying about the violence-related (b = 0.12, p = 0.033) and the political (b = 0.35, p < 0.001) phenomena.

Discussion

The goal of the current study was to assess the relationship between type of residential locality and various worry-related outcomes in an ethnic minority population. As the findings show, residing in locality of different ethnic composition plays a significant role in explanation of the extent of concern in this population. Three groups of findings are outlined below.

First, three worry profiles were yielded following cluster analysis, similar to Tsouros et al. (2021): highly (cautious), moderately (worried), and lowly (relaxed). In addition, social phenomena differed in the mean extent of concern about them. All this suggests that Arab citizens of Israel are generally concerned about social phenomena to a varying extent, and that particular groups of phenomena cause them greater worry than others. This, however, does not suggest a harmony between worry profile and the extent of concern about particular phenomena. For example, it is likely that not everyone who situates in the cautious cluster is highly worried about all social phenomena. Future studies should refer to this more extensively.

Second, residential locality was found partially corresponding to worry profile. Residents of mixed cities were more likely to belong to the highly worried cluster. Bonding social capital, which characterizes ethnically homogeneous localities (Amit and Riss 2014), may provide better social support (Tegegne and Glanville 2019) which ultimately decreases the extent of being concerned about social phenomena. This seems to apply to ethnically homogeneous Arab localities where common cultural ground allows for establishing bonding social capital. Future studies should examine the moderating and or mediating effect of social support and other indicators of social capital, like trust, social and political participation, and reciprocity (Paldi et al. 2021) on worries.

Third, locality type was found associating with the extent of being worried about social phenomena. The explanation to this finding is similar to the previous one. Bonding social capital, associated with social similarity (Tegegne and Glanville 2019), provides greater sense of stability, reciprocity and predictability (Johnson et al. 2017), therefore playing a protective role for the wellbeing of individuals (Gong, Xu, and Wang 2021). Therefore, unsurprisingly, Arab residents of ethnically homogeneous localities were less likely to worry than their counterparts residing in ethnically mixed localities. Nevertheless, this seems to apply to the general social and the violence-related phenomena only. The lack of difference in the extent of worrying about the political phenomena may signal political indifference prevailing in the Arab society of Israel. The turnout on nationwide elections in this sector is traditionally relatively low (Hitman 2021), and only following the 2021 elections, for the first time since 1948, Arab party has become part of the governing coalition (Ali 2022). Future qualitative studies may shed additional light on the link between political participation of ethnic minorities and their extent of being worried about the political phenomena.

Limitations

This study has numerous limitations. First, it is cross-sectional. Therefore, no causality can be concluded from the findings. For the same reason, it was impossible to assess changes in the extent of worrying about social phenomena and to link these changes to the type of residential locality. Note that worries do not remain the same throughout time and tend to change in their frequency and/or severity (Rashid and Olofsson 2021; Schulz et al. 2021). Second, the study was conducted in Israel. Therefore, generalization of the findings on the other minority populations in the country, or members of Arab minorities residing in other countries, should be done with care. Third, although social capital approach was employed to explain the relationship between locality type and worry, no direct measurement of social capital, be it by type (bonding, bridging, or linking) or using indicators (trust, social participation, and more), was applied. Future studies should consider this issue.

Conclusions

The study has found that for the most part, locality of residence explains differences in worry. Specifically, high worry profile, as well as worry about social and violence-related phenomena, are more likely to be associated with residence in ethnically mixed than in ethnically homogeneous locality. Social capital approach seems to be a good explanatory framework in this regard.

The findings of the study provide a reason for greater intervention on the policy making level. Specifically, ethnic minority residents of mixed localities should receive a greater attention from the policy makers, especially with respect to the general social and the violence-related phenomena. This can be done by a greater involvement of police in prevention of crime and ensuring a secure life of Arab residents in these localities. In addition, social workers and psychologists should work with members of Arab minority residing in mixed locality for reaching a greater wellbeing. Finally, greater amount of an information from the official sources regarding the scope of the abovementioned phenomena, as well as the proposed solutions for each one them, should be provided.

Disclosure statement

No potential conflict of interest was reported by the author(s).