The Long Arm of the Family: Family Role in Partner Selection Among Male Refugees in Germany

Abstract

Partner selection is a complex process shaped by structural, social, and personal factors. For refugees and forced migrants, displacement and adaptation add further challenges, with families—even from far away—playing a crucial role in shaping expectations and influencing partnership decisions. This influence can be direct—through arranged marriages or introductions—or indirect, by setting expectations based on ethnicity, religion, or social status. Family endorsement is often crucial in reinforcing ethnic, religious, or language group continuity and community acceptance. This study examines the role of family in partner selection among refugees who arrived in Germany largely during the 2015–2016 period, a group characterized by a high proportion of unmarried men with limited family support. This group is compared to the German resident population, both with and without a migration background. For the descriptive analyses, we draw on two data sources: the PARFORM dataset, collected in 2022–2023, which covers male Syrian and Afghan refugees who arrived between 2014 and 2018; and the ninth wave of the CILS4EU-DE, collected in 2022. The multivariate analyses rely exclusively on the PARFORM dataset and utilize linear probability models with selection correction. We contribute to the literature in three key ways: identifying mechanisms of family influence despite physical distance; introducing a multidimensional framework to measure partnership outcomes based on ethnic, religious, and linguistic endogamy; and providing new insights into the underexplored partnership formation of this refugee cohort.

1. Introduction

Partner selection is a complex and multifaceted stage of human relationships, shaped by structural, social, and personal factors. In this process, parents and family play a central role in shaping expectations, setting boundaries, and guiding their offspring’s key decisions. They function as “third parties”—social groups or members of such groups—who can significantly influence or even control the partner choice process [1]. Through the transmission of values and preferences within the socialization process, parents and family members crucially shape their offspring’s own preferences [2,3]. Parents (and other couples within the family) can also serve as an example or model with their own relationship, be it homogamous or mixed with regard to various aspects such as religion, ethnic background, or language. Beyond socialization, parents may also employ more direct means to steer their children’s partner selection, thereby imposing their own preferences.

Previous research in European countries suggests that immigrants and their children—particularly those from countries with more conservative religious norms—are more likely to experience strong parental and familial influence in partner selection. This is attributed to the families’ cultural backgrounds—broadly defined to include religious beliefs, value orientations, social norms, customs, and traditions [4]—which differ markedly from those of the host country, as well as to strong transnational networks and local communities that enforce partnership norms [3,5,6].

A group of particularly vulnerable migrants are refugees. Refugees and forced migrants face challenges of displacement from their country of origin and of having to adapt to new social environments, commonly experiencing the disruption of established social networks and support systems. In such diaspora contexts, family involvement in the partner search process can gain a particular relevance as it can help with protecting cultural continuity as well as maintaining traditions and values in unfamiliar settings. Such involvement can manifest through direct involvement in partner selection, such as in arranged marriages or when introducing potential partners, or through more subtle forms of influence, like setting expectations about acceptable partners based on ethnicity, religion, or social status. For many refugees, family endorsement of a partner is often crucial, as it is tied to broader concerns of cultural preservation and community acceptance. However, these dynamics can shift significantly when parents and kin remain in the country of origin, while forced migrants must often navigate entirely different social environments in Western Europe.

This study focuses on such a contextual situation, examining the experiences of forced migrants who arrived primarily during the 2015–2016 period—a time when Germany and other European countries witnessed a substantial influx of refugees from the Middle East and Afghanistan [7]. At arrival in Germany, a significant proportion of Syrian and Afghan refugees—our analyzed population—were unmarried men of prime marriageable age, indicating that partner selection may be a pertinent issue for them. A further distinctive aspect of the 2015/16 refugee migration is that most of these young men arrived in Germany without their parents or families [8]. Additionally, these refugees arrived in a country with no substantial pre-existing Syrian or Afghan ethnic communities, making fellow refugees their only readily accessible co-ethnic social group. These circumstances created a markedly different context for partner selection compared to that in the countries of origin or among other immigrant groups, such as in cases of voluntary migration accompanied by family members or embedded within larger co-ethnic communities.

Given the traditionally strong involvement of parents and extended family in the partner selection process in Afghanistan and Syria [9], an open question remains as to whether such third parties can continue to exert comparable influence over their offspring’s partnership choices from abroad or under the significantly altered context conditions in Germany. The intersection of family influence with the special circumstances faced by refugees—such as legal uncertainties, economic hardship, and social isolation—creates a distinctive context for partner search that differs from the experiences of other migrant or local populations. These dynamics make the investigation of—often remote—parental involvement in the partner search of offspring particularly compelling. While personal preferences, individual characteristics, and structural opportunities undoubtedly shape partner selection, this paper focuses on the critical role of parents and families in the partner search among male refugees. Specifically, it examines how several key mechanisms through which parents influence their children’s partner choices—such as the transmission of norms, parental control and approval, and active matchmaking efforts—operate within the unique context of forced displacement. The study’s contribution is threefold. First, it identifies and tests the specific mechanisms through which parents and families influence refugees’ partner choices, despite the challenges posed by physical distance and cultural dissonance with partnership practices in Germany. The second key contribution lies in the multidimensional measurement of group endogamy, expanding the analysis to differentiate partnerships based on nationality, religion, religious denomination, and language, thereby providing a comprehensive understanding of how similarity shapes partner selection among refugees. In additional analyses, we further differentiate between partnership types, distinguishing between marital and non-marital unions. Third, our study sheds light on the 2015/16 refugee cohort, a group for which little is known with regard to partnership formation.

The paper begins by outlining a theoretical framework, with particular attention to the concept of endogamy and the potential direct and indirect forms of parental involvement in offspring’s partnership formation. Building on this framework, we formulate a set of hypotheses. We then introduce the data, highlighting the value of the newly collected PARFORM dataset, which focuses on male refugees from Syria and Afghanistan who arrived in Germany between 2014 and 2018. This dataset enables a detailed analysis of the mechanisms underlying their partner selection processes. The paper concludes with a presentation of the results and a discussion of their implications for the social integration of forcibly displaced populations in Germany as well as the transferability of our results to female refugees as well as other populations.

2. Theoretical Frame and Expectations

2.1. Endogamy in Partnerships

The dynamics of partner search are influenced not only by individual preferences and structural factors but also by the expectations and values maintained by parents and extended family members. A key aspect of this influence is the frequent promotion of endogamy, i.e., encouragement of unions within ethnic or religious boundaries to preserve social cohesion, integrity, and homogeneity within an ethnic or religious group, thereby reinforcing group solidarity [1,10]. Parents, in particular, tend to play an active role in their children’s partner selection process, often advocating for endogamous unions to maintain the family’s cultural heritage and protect its integrity [11]. By guiding their offspring toward partners who share similar cultural backgrounds, parents aim to uphold familial and community values, reduce the potential for conflict, and ensure the stability of their family within the broader context of displacement and resettlement.

Parents and families enforce endogamy norms through several mechanisms that emphasize the importance of cultural and social continuity within their group. First, children are socialized within their own social group and are taught to identify closely with it. Social Identity Theory [12,13] explains that individuals maintain a positive self-concept and sense of self-worth through their affiliation with their group. To reinforce this self-concept, individuals tend to view their own group (e.g., ethnic or religious) more positively than others, often rating it as superior in comparison to out-groups. This strong sense of belonging fosters in-group favoritism, leading to more positive views towards in-group members and less favorable, sometimes even negative, attitudes toward those outside the group. This bias extends to partner choice, where ethnic identification and a strong sense of group belonging lead individuals to prefer partners from their own group [14]. Empirical evidence supports this phenomenon, showing that individuals with strong in-group identification are more likely to choose partners from the same group, reinforcing endogamy norms [15,16].

Second, parents actively seek endogamous unions for their children, believing that such partnerships are more harmonious and resilient due to shared cultural resources, values, and social norms. According to Kalmijn et al. [17], partnerships between individuals of similar ethnic backgrounds are less likely to experience conflict because shared tastes, values, and mutual understanding create a foundation for effective communication and cooperation. This similarity fosters mutual support and a stronger bond, which are essential for the long-term stability of the relationship. Similarly, the preference for endogamous unions is rooted in Becker’s New Home Economics [18] and Kalmijn’s Cultural Matching Framework [19]. Becker’s theory views marriages as economic units that produce valuable “commodities” such as emotional support, marital harmony, and effective childrearing, which in turn generate utility. Endogamous unions are seen as more efficient at producing these commodities and in turn yield higher utility because shared values and mutual understanding minimize conflict and promote collaboration within the household. Kalmijn’s Cultural Matching Framework builds on this idea by suggesting that partners who share similar backgrounds are better equipped to meet each other’s expectations, negotiate their roles, and maintain a cohesive family environment.

These theoretical perspectives highlight why parents often prioritize endogamous unions for their children, viewing them as a way to preserve cultural heritage and promote marital stability and success. Among many migrants originating from more collectivistic countries, parents tend to be particularly involved in their children’s decision-making related to the partnership formation (e.g., [10,20]). This is likely to be similar for refugees from Syria and Afghanistan. The following sections will examine the direct and indirect mechanisms through which parents and families steer their children’s mate selection towards endogamy [2].

2.2. Direct Means of Parental Involvement

Direct involvement entails parents actively participating in the mate selection process to guide it in a direction they find favorable. A key example of such direct involvement is matchmaking, where parents or family members take an active role in finding suitable partners and arranging introductions. These introductions often occur in controlled settings, such as family gatherings, cultural events, or even during visits to the parents’ home country [1,2,21]. Despite the decline in traditional arranged marriages, fully independent partner selection remains rare among the descendants of immigrants from Muslim countries living in Europe [22,23,24,25,26,27,28]. Most second-generation individuals from these backgrounds meet their partners through family connections [20,29]. Traditional engagement procedures tend to involve parental participation such as in visits between the families and formal agreement to the marriage [30]. As a result, unions initiated by couples with parental approval or those arranged by families with the offspring’s consent are common [31,32]. In contrast, parental involvement in mate selection is significantly less prevalent among native Europeans and second-generation migrants from other European countries [20,29,33,34].

The strongest instrument of parental matchmaking is marriage arrangement [27]. Arranged marriages tend to follow a well-defined and standardized process [27,35,36], with regulations in place to safeguard against coercion. Although parents select a potential spouse for their child, the couple is typically given an opportunity to become acquainted and can either consent to or reject the proposed match [35,36]. This practice is particularly prevalent among specific immigrant communities in Europe, such as Turkish [37,38,39] and Pakistani groups [40,41]. Arranged marriages are also a common method for forming transnational unions, where one partner originates from the parents’ country of origin [42]. In migrant communities that rely heavily on arranged marriages, there has been a shift from parental dominance to increased autonomy over time, allowing the younger generation greater input and independence in their partner selection [21,32,33,37,43,44,45].

Another direct way in which parents influence their children’s mate selection is through control. Behavioral control is a strategy used by parents to regulate their offspring’s dating and partner search behavior, involving the establishment and enforcement of rules and the monitoring of their activities [2,34,46,47]. Dating rules can take the form of supervision, restriction, or prescription [33]. Supervision rules require young people to keep parents informed about their dating activities, such as introducing their dates to them. Restrictive rules place limits on dating, for example, prohibiting sexual activity. Prescriptive rules outline parental expectations for potential partners and for behavior, such as adhering to traditional gender norms or conforming to group-specific norms [48]. Beyond setting supervisory or prescriptive rules or restricting dating behavior, parents may also engage in general monitoring of their children’s dating behavior. This parental monitoring has been found to be negatively correlated with romantic involvement [49] and sexual experiences [50].

However, adolescents and young adults might avoid parental control by not telling their parents about their romantic relationships [34]. Relating to this, parents’ ability to control their children’s mate selection process is significantly influenced by the degree of their offspring’s dependence on them [43,51,52,53]. One of the most notable forms of this dependence is close physical proximity or co-residence with parents which comes along with more frequent contacts but also a higher prevalence of conflict [53,54]. Living nearby or with parents can greatly affect partner selection and relationship formation, as parents are more likely to be aware of the partnership and actively intervene. This proximity increases parents’ capacity to monitor and control their children’s behavior in general [55,56]. Co-residence often comes with other dependencies, such as financial reliance.

2.3. Indirect Means of Parental Involvement

Parents indirectly influence offspring’s partner selection by transferring their culture and preferences for homophily to their children via the mechanism of cultural transmission. In this process, culture-related knowledge and values are passed down from parents or other family members to the child [57,58]. Endogamy can be both actively pursued by parents within this process as well as an unconscious result of cultural transmission. This transmission happens within the broader processes of enculturation—the natural immersion of an individual within their culture—and socialization, which involves intentional teaching and guidance [59]. Cultural groups, characterized by shared elements such as language, values, norms, social practices, and religious beliefs, therefore transmit these elements to younger generations with the aim of ensuring the group’s continuity. When minority groups perceive the dominant culture as a threat to their group’s preservation, they often adopt additional measures to strengthen their cultural transmission processes [60,61]. Immigrant parents, for instance, strive to recreate aspects of their country of origin’s cultural and social environment that are typically integral to the socialization process [62,63].

Parental approval of a partner and of the union is crucial for many immigrant descendants [21,30,64]. Relationships that are socially marginalized, such as to some extent mixed unions, often encounter higher levels of disapproval [65]. To secure their parents’ approval, some young adults strategically select partners whom they know their parents will accept [44] or adjust their partner choices to align with their parents’ preferences and expectations [66,67]. This adaptation can be a conscious or subconscious effort driven by the desire for parental approval [21]. Additionally, children may comply with parental wishes not only to avoid conflict with parents or to gain their approval but also to evade potential social sanctions from their wider family or community [11,47]. However, parental (dis)approval requires parents’ knowledge. Yet, some children choose not disclose their relationship to their parents [34].

Sanctions are employed to penalize behavior that deviates from group norms [1,58] In the context of ethnic partner selection, sanctions are, for example, used to reprimand individuals who violate norms of ethnic or religious endogamy [1,11]. Negative sanctions are often imposed when norms are not fully internalized or followed and can vary in severity. Mild social sanctions may include gossip or disparaging remarks [68], while more severe sanctions might involve temporary exclusion from the family or community, withdrawal of support, or even severing family ties [11,55,69]. Positive sanctions or rewards for adhering to norms, on the other hand, can be used to reinforce and encourage norm-conforming behavior [68]. When norms are deeply internalized, individuals may self-impose internal sanctions, such as feelings of guilt or shame, if they deviate from expected behaviors [70].

2.4. Expectations

The theoretical perspectives discussed above highlight the crucial role of family in partnership formation, where group endogamy functions both as a means of preserving cultural heritage and as a protective measure that minimizes the risks of marital discord and instability, thereby enhancing the offspring’s chances for a fulfilling and stable family life. Both direct and indirect processes can promote endogamous partnerships among offspring. This paper evaluates the effectiveness of these processes by testing hypotheses based on existing theoretical frameworks.

H1: 

Parental involvement in the mate selection process is associated with a higher propensity for endogamous partnerships.

H2: 

A closer residential arrangement (and consequently a greater potential of parental control over the offspring) is associated with a higher propensity for endogamous partnerships.

H3: 

The more strongly parents endorse endogamy in their children’s relationships, the more likely the children are to choose endogamous partnerships.

H4: 

The more strongly young people adhere to their parents’ norms of partnership endogamy, the higher the likelihood of gaining parental approval for their choice of partner.

3. Data and Methodology

For our analysis of young refugees’ partnerships and the role of their families in partner selection, we draw on data from the first wave of the PARFORM study (pre-publication version v2.0, for technical reports see Kogan et al. [71] and Gerber et al. [72]), which focuses on the partnership formation of young male refugees from Syria and Afghanistan in Germany. The data, collected throughout 2022 and 2023, provide insights into refugees’ current and past partnerships, socio-demographic characteristics, and their physical and emotional closeness to their parents. The target groups consisted of Syrian and Afghan men born between 1984 and 2002, who arrived in Germany between 2014 and 2018, unmarried. To our knowledge, no comparable data exist for young female refugees of the same origins, which prevents us from making gender-based comparisons.

To ensure a sufficient number of respondents, we selected 67 municipalities in Germany with a significant population of Syrian and Afghan refugees and approached all individuals of the target population registered. The municipalities included both large and medium-sized cities and towns, as well as neighboring suburban areas. Lists of Syrian and Afghan citizens were acquired directly from the municipalities and from the German Central Register of Foreigners Living in Germany (AZR), provided by the Federal Office for Migration and Refugees (Bundesamt für Migration und Flüchtlinge, BAMF).

The data were collected for a total of 3412 participants, of whom 701 did not fully complete the survey. Participants were surveyed using two methods: computer-assisted face-to-face interviews (CAPI) with 1139 respondents and web-based interviews (CAWI) with 2273 respondents. The data collection was carried out by the survey institute infas.

The data was weighted for univariate und multivariate analyses. The main goal of weighting is to correct for biases caused by the sampling design and by selective participation or non-participation. The following characteristics were used by infas for creating weights: origin (Syria, Afghanistan), sampling tranche, federal state, size class of the political municipality (only large cities), age group, and marital status. By using the weights, population estimates can be made from the sample in an unbiased manner.

We use a sample of young adults with and without migration background residing in Germany to serve as a comparison group for the refugees for the question on how they met their partners. For this purpose we use data from the 9th wave of the German part of the “Children of Immigrants Longitudinal Survey in Four European Countries” (CILS4EU-DE [73]). The cohort study first surveyed participants at the age of 14/15 in 2010 when they were high school students in Germany. In wave 6 in 2016, the survey included a panel refreshment sample for which a sample was drawn from population registers of the same cohort as the initial survey sample (by then ~20 years old). Respondents were between the ages of 26 and 28 at the time of participation in wave 9, which was collected in 2022. Children with a migration background were oversampled in the study. The survey was conducted in three different modes, starting with a CAWI mode, and additional modes of PAPI and phone interviews if the respondents did not answer in the CAWI survey.

In the first part, we compare the conditions of meeting their current partner between male refugees and young adults with and without a migration background residing in Germany. For this analysis we restrict the sample to respondents who are currently in a partnership. This yields an analytical sample of N = 1283 male refugees (33%) from the PARFORM data and N = 257 men and women with a migration background from predominantly Muslim countries (7%), N = 638 men and women with a migration background from other countries (16%), and N = 1730 German majority men and women, i.e., those with no migration background at least up to the third generation (44%) from the CILS4EU data. Using the term ‘migration background’, we describe individuals who either themselves, one of their parents, or one of their grandparents was born outside of Germany [74]. Overall, the sample for this analysis includes N = 3908 respondents.

In the second part, we conduct an in-depth analysis of male refugees’ partner selection based solely on the PARFORM data. These main analyses focus on refugee men only. We impute missing values in the control variables by means of a multiple imputation procedure. For the multiple imputation, we apply a predictive mean matching and k nearest neighbor matching procedure with a total of 25 imputations per respondent. Hence, all analyses account for the multiple observations used per person by clustering standard errors on the individual level, keeping them robust. The analytical sample for the refugee analyses includes only refugees who reported to currently be in a partnership or who reported having a partnership since they arrived in Germany if they are not currently in one. N varies throughout the analyses, as cases with missing values for the dependent variables are excluded from the analyses after the completion of the imputation procedure, and the numbers of missing values on the dependent variables differ. The analyses include between N = 1894 and N = 1666 individuals. Descriptive statistics of all used variables can be found in the Supplementary Materials (Tables S1 and S4) and are based on the unimputed dataset.

3.1. Dependent Variables

The focal outcomes in the main analyses, including only refugee men, are the characteristics of the current or previous partner, in particular whether the partner has or had the same nationality, religion, religious denomination, and mother tongue as the respondent. Refugees were asked about the characteristics of their current partner if they were in a partnership at the time of the interview. If they were not currently in a partnership, they were asked to report the characteristics of the partner from their last partnership in Germany. Both questions allow for the partnership to have been ongoing since before the arrival in Germany. The outcomes are captured in a dichotomous form and pertain to several dimensions of endogamy: (a) by national origin (co-national vs. international partnership), (b) by religion (co-religious, i.e., partners have the same religion vs. interreligious partnership), (c) by religious denomination (co-denominational, i.e., partners not only belong to the same broad religious group, e.g., Islam, but also share the same branch of Islam, e.g., Sunni, vs. interdenominational partnership), and (d) by mother tongue (co-linguistic partnership vs. partnership between two individuals with a different mother tongue).

Parents’ approval of the partnership is measured via the respondents’ perceptions on whether their parents accept their partnership (1—they do not accept it at all to 5—they completely accept it). Respondents who reported that their parents do not know about the partnership, were asked to indicate their hypothetical parental acceptance. This is converted to a dichotomous variable, indicating high parental approval if the acceptance is of value 4 or higher compared to low parental approval if it is lower.

3.2. Independent Variables

We assessed the direct influence of parents via matchmaking by asking whether respondents met their partner through the family (yes vs. no) and whether the parents took part in the decision about entering into a relationship with the partner (respondent’s family was involved in the decision on the partnership vs. the respondent decided alone). Parents’ direct involvement via control is measured using a variable about the physical proximity of parents to the respondents. In this variable, we measure if at least one of respondents’ parents lives in the same household as the respondents, if at least one of respondents’ parents lives in Germany (but none live in the same household as the respondents), or if the parents live elsewhere (this last category includes those whose parents are both deceased).

Parents’ indirect involvement includes norm transmission and positive sanctioning—measured via parental approval (see above). Norm transmission is measured by respondents’ report of how important it is to their parents that their partnership is endogamous on the four dimensions corresponding to the dependent variables mentioned above, i.e., nationality, religion, religious denomination, and mother tongue (1—not at all important to 5—very important). This information is coded dichotomously in the four variables, indicating that respondents think their parents prefer them to have a co-ethnic partner vs. not, a co-religious partner vs. not, a co-denominational partner vs. not, and a partner speaking the same language vs. not. Preference for endogamy is assumed if parents’ importance is indicated as a value of 4 or higher, meaning important or very important.

We include several control variables in the models. These are survey mode (CAPI vs. CAWI), respondents’ religion (no religion, Christian, Muslim, other religion) and education (university degree vs. no university education), age at participation (continuous), education of the parent with the higher education (no degree, primary school degree, secondary school degree, university degree). We also control for respondents’ religiosity (religion is important to them vs. not), the partnership type (marriage, committed partnership, casual partnership), and their subjective socio-economic status in childhood (1—much lower than average to 5—much higher than average). We additionally include if refugees had dating experience before migration (yes vs. no) and the years since they had arrived and started dating in Germany (for those who arrived in Germany before the age of 16, we used the time since age 16). We also control for respondents’ country of origin (Syria vs. Afghanistan). We include dummies of having a restricted vs. non-restricted legal status as well as of being employed vs. not and we include an indicator for whether the respondents reported a current or a previous partnership. Lastly, we include a dummy indicator of whether the reported partnership started before arrival in Germany vs. after. In the analyses looking at parental approval of the partnership as an outcome, we additionally include a continuous measure of the quality of the relationship with the parents (1–5), with higher values indicating a closer relationship.

3.3. Analytical Steps

Since our analyses focus on currently or previously partnered individuals, it is important to also account for those who did not report any relationships during their time in Germany. In the original sample, 23% did not report having had any partnership in Germany, and 36% did not report any information on either a current or previous partner (see Table S3 in the Supplementary Materials). We address this sample selection bias, which occurs when the sample used for analysis is non-random and potentially biased, by including inverse Mills ratios as a control variable. To this end, we estimate a probit model of being included in the sample of those who reported a partnership (selection equation; see Tables S4 and S32 in the Supplementary Materials for descriptive statistics of additional control variables used), and include this probability in the analytical sample as a control variable in the main analyses. This approach helps to obtain unbiased and consistent estimates for the outcome variable, correcting for potential selection bias.

For the main analyses, we employ linear probability models (LPMs) and apply sampling weights.

4. Results

4.1. Descriptive Results

In a first step, we compare the share of refugees who found their partners with the help of their family with that of other groups, namely members of the German majority population and individuals with a migration background.

Table 1. Meeting partner via family.

	Met Current Partner Not via Family	Met Current Partner via Family	Met Current Partner Exclusively via Family
Refugees 1	58.75%	41.25%	29.13%
Migration background from Muslim majority origin country 2	92.44%	7.56%	7.00%
Other migration background 2	94.28%	5.72%	3.63%
Natives (without migration background) 2	95.92%	4.08%	2.12%

Source: ¹ PARFORM wave 1 (pre-publication version v2.0), ² CILS4EU-DE (waves 3, 6, 9), data weighted. Note: The category of Muslim majority origin countries subsumes the following countries and country groups: Afghanistan, Albania, Azerbaijan, Bangladesh, Bosnia and Herzegovina, Cote d’Ivoire, Ethiopia, Gambia, Guinea, Indonesia, Iran, Kazakhstan, Kyrgyzstan, MENA (Lebanon, Northern Africa, Western Asia (for more detailed information on these country group, see [74])), Nigeria, Pakistan, Eritrea, Senegal, Sierra Leone, Somalia, Tajikistan, Tanzania, Turkey, Turkmenistan.

displays the shares of having met the current partner via family members by population group, showing that refugees are much more likely to report this. About 41% of refugees in the sample reported having met their current partner via their family, while this share is only about 4% for members of the native German majority. Individuals with a migration background are only slightly more likely than the native-born majority to have met their current partner via their family. Even among the individuals with a migration background from Muslim majority countries—who are presumably more similar to the refugees in our sample in terms of religion, customs and, for some, also language—only around 8% have met their partners via their family. The latter group was constructed to enable a comparison between refugees and migrants, who originate from countries which have a comparable prevalence of traditional gender roles, strong family orientation, and family cohesion. When looking at the shares of those who reported having found their partner only through the family, refugees are still much more likely to have reported only the family as the way of having met their partner compared to the other groups. It is, however, also important to keep in mind that the sample of refugees only includes men, while the shares reported for the other three groups include both men and women.

In the following main analyses, we focus on male refugees’ partnerships, which include the current partnership or the previous partnership for those who are not currently in a partnership (but reported having had one in Germany previously). Of the analytical sample, 32% reported information on their previous partner (meaning they are not currently in a partnership) and 68% reported information on their current partner (see Table S3 in the Supplementary Materials). The endogamy within these reported partnerships is highest for the dimensions of mother tongue (in 69% of cases, partners speak the same mother tongue) and religion (69% of the partnerships are co-religious), while endogamy is slightly lower for religious denomination (64% of the partnerships are co-denominational) and nationality (57% of the partnerships are co-national). In addition, over 60% of respondents state that their parents approve of their reported partnership (see Table S1 in the Supplementary Materials).

4.2. Direct Means of Parental Involvement: Multivariate Results

We begin by examining how parents and family are directly involved in their children’s process of finding a partner. Figure 1 shows how parental involvement affects the chances that the child is in a relationship with someone who shares a similar background —such as the same religion, denomination, language, or nationality. The results are shown in percentage points, based on a linear probability model.

The findings demonstrate that when children meet their partner through their family or when parents are involved in deciding on the relationship, the couple is more likely to be similar across all four dimensions: nationality, religion, denomination, and language. All associations are highly statistically significant. Meeting a partner through the family seems to have a slightly stronger impact than just having parents involved in the decision. For example, children who met their partner through their family are up to 25 percentage points more likely to end up in a relationship with someone who shares the same religion, denomination, or language as compared to refugees, whose parents did not play any role in the partner selection process. When parents were involved in deciding on the relationship, the probability of having a partner with similar characteristics was a bit lower. One of the most striking findings is related to nationality: if the partner was introduced by the family, the likelihood of the relationship being with someone of the same nationality was almost 40 percentage points higher than among couples that met without family involvement. If parents were involved in the decision-making process, the likelihood was about 25 percentage points higher. These results support hypothesis H1, which predicted stronger endogamy in partnerships where parents and family are involved through some kind of matchmaking.

Next, we look at the role of parental control—in this case, measured by how physically close the parents live to their children (see Figure 2). When parents live in Germany (compared to living in another country) their children are more likely to be in relationships with partners who share a similar background—and this pattern is statistically significant across all four dimensions: nationality, religion, denomination, and language.

Whether parents live in the same household as their sons or outside of Germany does not seem to make a big difference in relation to their sons ending up with a partner who shares a similar background. The only exception is nationality: male refugees living with their parents are more likely to have a partner from the same country compared to refugees whose parents live abroad. Overall, the limited effect of living together with parents is a bit surprising. One possible reason is that male refugees who co-reside with their parents may be less likely to form a partnership at all, which is supported by additional analyses which suggest that young men living with their parents have the lowest likelihood of being in a partnership (see Table S33 in the Supplementary Materials). All in all, these results lead us to conclude that parental control in terms of physical closeness plays a nuanced role in whether refugees’ partnerships are endogamous or not. Apart from parents living in Germany, there is little impact of parents living in the same household as the refugees. Thus, we find partial support for hypothesis H2.

4.3. Indirect Means of Parental Involvement: Multivariate Results

Turning to the indirect involvement of parents in their offspring’s partner choice, our analyses on the influence of the cultural transmission process are displayed in Figure 3. Parental preference for endogamy in the respective dimension is positively associated with refugees’ endogamous partnerships in all four dimensions. However, the coefficients are only significant on conventional levels for co-national and co-linguistic partnerships. Refugees whose parents endorse endogamy in terms of national origin are more likely to have a co-national partner and those whose parents endorse linguistic endogamy are more likely to have a co-linguistic partner. The coefficient pertaining to parental preference for religious endogamy is of a comparable magnitude, but is significant solely at the 10% level. These results largely support hypothesis H3, which suggested that parental preferences for endogamy increase the actual endogamy in partnerships of male refugees.

The second way that parents can indirectly influence their children’s partner choices is through approval or acceptance of the relationship. This is what we refer to as positive sanctioning. Figure 4 shows how much parental approval refugees reported depending on whether their partnership matches their parents’ preferences for similarity in background.

The graph shows that when refugees are in an endogamous relationship (e.g., with a partner of the same nationality, religion, denomination, or language) but their parents do not specifically prefer that kind of match, parental approval is not higher than for those whose relationship is not endogamous and whose parents have no clear preference for endogamy. What really stands out is the following: if parents prefer an endogamous relationship but their son’s partnership is not endogamous, parental approval turns out significantly lower. In these cases, approval is 15 to 22 percentage points lower, depending on the dimension. On the other hand, when refugees’ relationships match what their parents prefer, approval is significantly higher across all four dimensions in comparison to those who are not in an endogamous union and whose parents have no clear preference with regard to endogamy. This supports hypothesis H4, which stated that parents are more likely to approve of partnerships that align with their preferences for similarity.

4.4. Sensitivity Analyses

We ran several sensitivity analyses in order to make sure our results are robust to alternative specifications and definitions. First, we sought to ensure that family’s involvement in the decision on the partnership was defined adequately as a dichotomous variable. To this end, we ran the analysis for H1 with the original five-point scale indicator of family’s involvement (see Table S4 in the Supplementary Materials for descriptive statistics of this variable). For all dimensions of endogamy, the results show that the most striking effect is that of the category, “Only the respondent decided on the partnership” (see Table S10 in the Supplementary Materials), which supports our decision to code the dichotomous variable in this way: the family was involved in the decision in some way vs. the respondent decided entirely on his own to engage in the partnership.

In order to make sure that the analyses for H4 on the association between met/unmet parental endogamy preference and parental approval were not driven by misrepresentation of respondents’ perceived approval of their parents who are no longer alive (and therefore could not express approval or disapproval), we estimated the analyses for H4 again for a subsample of refugee respondents whose parents were both still alive at the time of survey participation (60% of the sample). The results show the same patterns as in the main analyses for H4 (see Table S11 in the Supplementary Materials).

Our analyses included respondents who currently reported a partnership and its characteristics as well as those who were currently not in a partnership but had previously had a partnership in Germany. Since the previously reported partnerships no longer exist, they might differ in terms of their composition compared to the current partnerships. Hence, we re-estimated all analyses only for respondents who reported a current partnership. Coefficients in the analyses on this subsample are slightly smaller than in the main analyses but reflect the same general patterns as reported in the main analyses (see Tables S12–S16 in the Supplementary Materials).

We additionally estimated all analyses of the main models with a dataset that used a different imputation strategy (imputation of missing values for all variables instead of only for control variables and no deletion of missing values on the dependent variables) in order to make sure that the imputation of missing values did not bias the results. The results did not change (see Tables S17–S21 in the Supplementary Materials).

The reported partnerships include both marriages and non-marital partnerships such as committed and casual partnerships. We re-ran the analyses separately for current marriages and non-marital partnerships in order to evaluate whether the results are driven by any of those partnership types. In the main analysis sample, 38% of reported partnerships are marriages, while 36% are committed and only 26% are casual partnerships. The analyses including only marital partnerships mostly do not yield any significant differences across the categories of the independent variables in the analyses pertaining to hypotheses H2–H4 (see Tables S22–S26 in the Supplementary Materials). The coefficients in the analyses for non-marital partnerships on the other hand show the same patterns as the main analyses with, however, consistently larger effect sizes for most predictors and dimensions. A reason for this discrepancy might be related to the lack of variance in the dependent variable when analyzing marital unions. For instance, 91% of marriages in the data are co-linguistic, whereas only 52% of committed or casual partnerships are co-linguistic, with this discrepancy between endogamy in marital and non-marital partnerships being similar for the other endogamy dimensions. As a consequence to this very high share of endogamous marriages, direct parental involvement in the children meeting their partners, physical proximity of parents to their offspring, as well as parental preferences do not have much explanatory power with regard to marital unions. The high shares of endogamous marital unions can, nonetheless, be seen as supportive evidence for indirect parental influence, i.e., that parents pass on their culture as well as a preference for and/or norm of marital endogamy to their children within the culture transmission process. The pattern of parental influence is far more differentiated in non-marital unions.

Lastly, we added a control variable to the main analysis models differentiating between the following parental constellations: (a) parents have different religions, (b) parents belong to different religious denominations within the same religion, (c) parents belong to the same religious denomination (see Table S4 in the Supplementary Materials for descriptive statistics of this variable). Through this variable, we assessed whether potential role modeling within the family confounded the results. Results remained almost identical to those found in the main analyses (see Tables S27–S31 in the Supplementary Materials).

5. Summary and Conclusions

This paper examined the role of parental and family involvement in partner selection among young male refugees from Syria and Afghanistan in Germany. Particular attention was given to the mechanisms through which these third parties promote endogamy in their offspring’s partnerships. The context of refugee migration to Germany in the mid-2010s posed unique challenges for the many young, unmarried men, many of whom arrived without their families and in the absence of established Syrian or Afghan ethnic communities. While personal preferences, individual characteristics, and structural opportunities undoubtedly influence partner selection, this study focused on the central role of parents and families—key actors in the partner choice processes in the origin countries. However, in a diasporic context, their involvement is shaped, and at times constrained, by physical separation and the markedly different social and cultural environment surrounding partnership formation in the host country.

The results indicate that, despite—or perhaps precisely because of—the physical and other obstacles to parental involvement, 41% of refugee respondents reported having met their current partner through family connections. This finding underscores the significant role that families continue to play in partner selection among refugees, standing in marked contrast to patterns observed among other young people in Germany, including the children of immigrants from Muslim-majority countries.

Using the PARFORM dataset on the partnership formation experiences of refugees who arrived between 2014 and 2018, the study further explores both the direct and indirect mechanisms of parental involvement. Additionally, it highlights the significance of endogamy in terms of nationality, religion, religious denomination, and language. By emphasizing the multidimensional nature of group endogamy and extending theoretical insights on endogamy and cultural preservation into the specific context of forced migration, the study advances the body of research on partner selection among migrants, particularly those from Muslim-majority countries [5,6].

We draw on Social Identity Theory [12] and the Cultural Matching Framework [19], which suggest that endogamous unions are valued for their potential to preserve cultural identity and enhance marital harmony. Parents play a central role in promoting endogamy. This study empirically tests the mechanisms through which parents and families influence their offspring’s partner selection—both directly and indirectly. Notably, the present research extends these frameworks to the refugee context, exploring how these dynamics unfold in the absence of immediate family support.

Direct influence typically involves parents or family members introducing potential partners or arranging unions, often during family gatherings or visits to the country of origin. While parental control has declined in some migrant groups, matchmaking and behavioral control—such as setting rules or monitoring dating—still remain common in more traditional communities. However, parental matchmaking efforts may also have diminished due to the unique displacement situation—specifically, the absence of extended social networks typically relied upon for partner selection in origin countries and the physical separation of some families.

Our results for young male refugees indicate, however, that direct parental involvement, particularly through matchmaking, is significantly associated with an increase in endogamy across all dimensions. The strongest effects are observed for nationality. Refugees who met their partners through family networks were considerably more likely to form endogamous unions, highlighting the enduring influence of family connections even in contexts of displacement. This finding resonates with earlier research showing the prevalence of family-mediated partner selection among immigrants from Muslim-majority countries in Europe [1,3]. While the broader trend among migrant populations has shown increasing autonomy in partner selection over time [37], the results suggest that in refugee contexts, marked by disrupted social networks and displacement, family matchmaking continues to play a relevant role, potentially functioning as a stabilizing force.

Parental proximity further shapes partnership outcomes, although its effects are more nuanced. Having parents present in Germany was associated with greater endogamy in terms of nationality, religion, denomination, and language, but co-residence did not yield a consistent significant influence compared to living abroad. This finding partially echoes previous work suggesting that physical proximity amplifies parental control and cultural transmission [54,55], but it also underscores the adaptive strategies families employ in cross-border contexts. The ability of families to influence partner selection even without co-residence highlights the resilience of cultural norms in immigrant contexts.

Indirect influence operates primarily through cultural transmission and sanctions. Parents also pass on norms and preferences for endogamy, encouraging children to choose partners with similar national, linguistic, or religious backgrounds. Social Identity Theory frames this as a way to preserve in-group cohesion and maintain a positive group identity. Approval and sanctions further reinforce these norms: endogamous unions receive parental acceptance, while nonconforming partnerships face disapproval or social penalties, such as pressure or withdrawal of support.

Our results suggest that indirect parental influence through cultural transmission and sanctions also played a critical role in refugees’ partner selection processes. Refugees whose parents endorse endogamy were more likely to form endogamous partnerships, particularly with regard to nationality and language, reflecting the effectiveness of intergenerational transmission of cultural preferences [58]. Additionally, parental approval of partnerships was strongly contingent on alignment with endogamy preferences. Relationships that met these preferences were associated with significantly higher levels of parental acceptance, while deviations were met with disapproval. These findings align with earlier studies on the importance of parental endorsement in maintaining group cohesion [11,64].

Interestingly, the impact of parental involvement varied across partnership types. While endogamy with regard to all aspects under study was overwhelmingly prevalent in marital unions, parental influence was more pronounced in non-marital partnerships, where greater variability in endogamy allowed for stronger differentiation in outcomes. This pattern is in line with the winnowing thesis [75] according to which endogamy increases with the level of commitment within a union, i.e., from dating, over cohabitation, to marriage. Thereby, non-marital partnerships may serve as a testing ground, providing young people with more freedom of choice with regard to the characteristics of a potential partner. Nonetheless, they are not entirely free of parental involvement and cultural expectations. Non-marital unions may also serve as a testing ground for the negotiation of cultural norms, whereas marital unions remain firmly anchored in traditional expectations.

All in all, this study demonstrates the centrality of parents and families in shaping partner selection among refugees, reinforcing their role as custodians of cultural norms. By fostering endogamy through both direct and indirect means—often across borders—families serve as mediators between displaced individuals and their cultural heritage, balancing the demands of integration with the preservation of identity. These findings contribute to the understanding of refugees’ integration processes, highlighting the importance of familial dynamics in navigating the complex interplay between group continuity and social adaptation. In this way, this study contributes to the broader discourse on refugee migration and integration.

However, an open question remains: are the patterns observed among Syrian and Afghan refugees similar to those among the native German population or among other migrant groups? The current data do not allow us to answer this question directly, but they point to promising directions for future research. While the process of cultural transmission is universal, the content and intensity of what is transmitted likely vary between groups. According to the Theory of Cultural Transmission in Minorities, the motivation to preserve and pass on cultural heritage becomes particularly salient in migration contexts where cultural continuity is perceived to be under threat. Central to this transmission are shared cultural core elements—such as norms, values, and language—that are essential for group cohesion and mutual recognition. As a result, families may undertake actions aimed at preserving these elements, including the encouragement of endogamous partner choices [60]. Conversely, the cultural transmission within the majority is far less compromised, making the emphasis on choosing a culturally similar partner less salient for native Germans [76]. While individual preferences for culturally similar partners—such as shared values or lifestyles—may operate across all groups [1], the cultural transmission motive is likely less pronounced among natives. Regarding direct parental involvement, young adult natives in Germany are far less likely to meet their partners through family networks. Instead, partnerships typically emerge through friends, educational institutions, workplaces, or other public spaces [20]. Parents, moreover, generally refrain from actively participating in the partner selection process [76], a pattern that our own results also reflect (see Table 1).

Another key distinction in partner selection among refugees, compared to native majority or minority groups, lies in the opportunity structure that shapes the available partner pool. For refugees, significantly skewed sex ratios [8] constrain partner availability, whereas the native population typically has access to a much broader and more balanced pool of similar partners with regard to religion or language. Furthermore, transnational strategies for partner choice—such as traveling to the country of origin or facilitating family reunification—are often legally restricted or financially burdensome for refugees [77]. This makes forming transnational unions far less feasible for refugees—at least until they have acquired a stable legal status in the host country—than for other immigrant groups with greater international mobility and legal security (e.g., [5]).

Another question that remains unanswered by this study is how patterns of parental involvement in partner selection vary by gender. Existing research suggests that parental expectations and involvement tend to be more pronounced for daughters while sons are generally granted greater autonomy, particularly in matters such as dating (e.g., [47,78]). In the Afghan context, for instance, strong cultural norms around female chastity and gender segregation are likely to influence both parental control and the opportunities for meeting potential partners [9,79]. The extent to which such norms are upheld in the host country may further affect these dynamics. Additionally, the conditions of arrival and demographic composition differ between male and female refugees from these regions [8], suggesting divergent starting positions, opportunity structures, and possibilities for parental involvement. For example, female refugees more often arrived in Germany accompanied by family members, potentially facilitating stronger family influence in partner selection. The inability to generalize findings from male respondents to the broader female refugee population remains a key limitation of the current study.

Another limitation is the study’s reliance on self-reported data on part of the young men which may be subject to recall bias or social desirability bias, particularly in sensitive areas such as parental approval. Additionally, the analyses capture partnerships formed within a relatively short timeframe after migration, potentially overlooking longer-term shifts in partnership patterns or the evolving influence of parents as refugees integrate into the host society. Furthermore, the focus on endogamy along the lines of nationality, religion, religious denomination, and language may also neglect other relevant factors, such as education or socioeconomic status, that could impact partner selection. These limitations suggest avenues for further research to provide a more comprehensive understanding of partnership formation among refugees.