“Escape” from home? The moderating role of sexual orientation on the association between social origin and educational attainment Open Access

Families are not only a place of emotional support; families also transmit privileges or disadvantages (Jæger and Breen, 2016). This is especially true for the intergenerational transmission of education. It is familial cultural capital that is enhanced and, thus, more decisive for educational success than the knowledge acquired through schools (Jæger and Breen, 2016). Social origin is still the yardstick of all factors when it comes to one's own educational success (Bernardi and Gil-Hernández, 2021). Germany, in particular, with its deeply rooted ideology of “innate talent” (Powell and Solga, 2011), has a decades-spanning documented link between social origin and educational achievements (Betthäuser, 2020; Neugebauer, 2015).

Lesbian, gay and bisexual (LGB) people, however, generally have weaker ties to their families (Dewaele et al., 2011; Fischer, 2022). LGB individuals build closer ties to their friends (Dewaele et al., 2011; Fischer, 2022), their “family-of-choice” (Weston, 1991), in order to replace ambivalent relationships with their “family-of-origin.” The weaker ties to the LGB individuals’ family-of-origin may impact the educational pathways that LGB people take as they are less influenced by parental cultural capital, instead gaining access through their family-of-choice. Accordingly, we want to shed light on the question of whether the intergenerational transmission of social class to educational attainment differs between LGB people and heterosexuals. We refer to LGB people because existing quantitative research on educational attainment, including this study, has not yet been able to look at other non-heterosexual identities (nor at gender identity) owing to data limitations. However, we believe that many of the arguments made in this article could be applicable to other non-heterosexual identities too (as well as transgender and non-binary identities).

A large number of studies in diverse contexts show generally higher educational attainment among LGB individuals than among straight individuals in the United States (Badgett, Carpenter and Sansone, 2021), Canada (Waite and Denier, 2015), the United Kingdom (Arabsheibani, Marin and Wadsworth, 2004), Sweden (Ahmed and Hammarstedt, 2010), France (Laurent and Mihoubi, 2012) and Germany (Humpert, 2016; Kroh et al., 2017). At the same time, only a few studies have aimed to understand why this educational advantage arises and whether this advantage might originate from specific subgroups within the LGB population.

Some studies have focused on the moderating role of gender and find that the educational advantage might be concentrated among gay men, and that it is less applicable to lesbian women and bisexual people (Fine, 2015; Mittleman, 2022; Mollborn and Everett, 2015; Walsemann et al., 2014). However, little attention has been paid to the possible role of social origin. It may be that one explanation of the educational differences of sexual minorities compared with the straight population lies in the heterogenous effect of social class origins on educational attainment. One previous study has found that upward educational mobility is higher among LGB individuals across a variety of countries (Boertien, Perales and Pessin, 2023). This provides a first indication that social origin might moderate the relationship between social origin and educational attainment. However, this previous study did not dive deeper into social origin, both empirically and theoretically, where it relied on ad hoc explanations rather than theoretical considerations rooted in social stratification research. The current paper fills this gap by using core notions from social class research to construct a theoretical foundation for explaining why social origin could moderate the effect of sexual orientation on socioeconomic outcomes.

In his memoir, Returning to Reims, Didier Eribon (2016) describes his feelings of rejection and shame as a gay man with working-class origins and the urge to escape from his working-class background—especially because of the masculinity ideals that were marked with physical appearance of steeled muscles and rough physicality surrounding him at home. Precisely because of his sexual orientation as a gay man, he distanced himself from “home,” not only geographically but also emotionally, as he entered the “gay world.” Eribon (2016, p. 26) asks himself to what extent his sexual orientation had given him the “opportunity” or the need to escape the socially disadvantaged working class: "I was a gay child, a gay adolescent, not a working-class child.” Eribon's memoir focuses on his experience as a gay man, but the challenges imposed by heteronormativity are shared by non-heterosexual people in general.

The experience of same-sex attraction and the related social tensions within the family, school and other environments can lead to situations where LGB people adjust their values, attitudes and actions. This adjustment of the habitus can result in LGB people making different educational choices as compared with straight people (McDermott, 2011). Some LGB people might drop out of education to enter employment and deal with challenges that arise during adolescence or young adulthood (McDermott, 2011). Other LGB people might see education as a way out of possibly discriminatory environments, as a resource to deal with future discrimination (Barrett, Pollack and Tilden, 2002) or as necessary to access queer spaces (Barrett and Pollack, 2005; Rooke, 2007). Since people from higher social backgrounds generally have high educational expectations, a positive adjustment of the perceived benefits of education can be particularly relevant for LGB people from lower social backgrounds. Therefore, we propose the “queer habitus” hypothesis, which states that the formal higher education of minoritized sexual orientations can be primarily observed among those of lower social backgrounds because of a disrupted class habitus initiated by the development of their sexuality.

Empirically, the paper contributes by using a broader operationalization of social origin relying on occupational status, and by studying educational classifications that are more tailored to the German context. The analysis is based on the nationally representative and largest household panel survey in Germany, the Socio-Economic Panel study (SOEP), and its 2019 nationwide oversample of LGB households (sample Q; Fischer et al., 2022). This study has two empirical advantages over existing research in this field. First, because of the rich biographical information, including parental information on occupational status, we make social class origin central to this study. Second, because the data contain both single and partnered individuals, we can account for the effect of partnerships and we can estimate the main effect of sexual orientation regardless of partnership status.

We find that both LGB men and LGB women have higher educational attainment than their straight counterparts in Germany. At the same time, the educational premium of LGB people is concentrated among those with a disadvantaged social origin compared with straight individuals from similar social backgrounds. In addition, our results suggest there could even be an educational penalty among LGB people of advantaged social origin. Overall, this lends support to the queer habitus hypothesis, which argues that the disruption of the heterosexual habitus by an LGB identity causes a weaker association of social origin with educational attainment.

Whereas qualitative research has addressed issues at the intersection of class, gender and sexuality (Barrett and Pollack, 2005; Rooke, 2007), quantitative research has only recently started to look at the interaction between sexuality and gender when studying the educational outcomes of LGB people (Fine, 2015; Mollborn and Everett, 2015). Lesbian and bisexual women show lower educational attainment than heterosexual women, while gay men, but not bisexual men, have higher educational attainment than heterosexual men (Mittleman, 2022). It is possible that men who identify as gay are higher educated because the college experience itself increases the willingness to disclose a non-heterosexual orientation in surveys (Barrett and Pollack, 2005). However, Mittleman (2022) challenges this assumption by showing that higher grades and completion rates are also visible within the group of gay men who attend college. Arguably, this shows that sexual orientation—rather than an effect of college attendance on being “out”—exerts an impact on educational attainment (Mittleman, 2022).

These findings are in contrast to some analyses suggesting that a later age of coming out is associated with greater educational success among gay men (Barrett, Pollack and Tilden, 2002; Pearson and Wilkinson, 2017, 2018). Gay men who come out during puberty have a greater risk of being disadvantaged in their school careers because they have to deal with (covert) homophobia much earlier (Pearson and Wilkinson, 2017): “The only way for gays overall to be overachievers would be for gays who delay [their coming-out] to be very strong overachievers” (Barrett, Pollack and Tilden, 2002, p. 178). However, McQuillan (2017) provides evidence that individuals with a “stable” gay identity (e.g., gay teenagers who are already out in adolescence and continue to identify as gay later in life) do not perform worse than heterosexuals.

In short, it is plausible that an advantage in educational success of gay men exists net of any selection processes that may occur. For lesbian women, an educational advantage is found only among older cohorts who presumably came out later in adulthood. This reverses for younger cohorts in the United States, with both bisexual and lesbian women being less likely to graduate from college than heterosexual women, a conclusion that also holds for bisexual men (Mittleman, 2022). We know much less about the intersection of gender and sexuality in countries beyond the United States.

One possible explanation for educational differences between men and women may lie in hegemonic gender norms that boys and girls are socialized into (Morris, 2011). Girls are socialized according to a “good girl” stereotype, which demands them to be quieter and interested in reading, compared with boys for whom academic disinterest and physical activities are linked to normative masculinity (Morris, 2011). This “bad boy” norm is particularly prevalent in working-class cultures (Barrett and Pollack, 2005). Competitiveness as hegemonic masculinity behavior also applies to boys from middle and upper classes, however in a more nuanced way than the more physical demonstrations of “strength” found in the working class (Barrett and Pollack, 2005; Morris, 2011).

Sexual minorities differ in the pressure that hegemonic gender norms place on heterosexual boys and girls in their socialization, as other stereotypes related to gender are prevalent owing to sexuality (Klysing, Lindqvist and Björklund, 2021). Because gay boys are confronted with marginalized masculinities, they fail to conform to those hegemonic masculinity behaviors in order to “prove” their masculine gender to others. Therefore, heterosexual boys are likely to achieve lower educational levels than gay boys because high achieving at school would violate heterosexual hegemonic gender norms. Distance from hegemonic masculinity may further encourage gay boys to achieve high academic success, which may be a way to (over)compensate for lack in self-worth due to stigma and threats, a phenomenon that Pachankis and Hatzenbuehler (2013) call “the best little boy in the world” (McQuillan, 2017; Pachankis et al., 2017). As other masculinities feel unattainable, gay men may be more likely to try to achieve high educational degrees because this competitive self-mastery and, thus, a “masculine proof” may seem more manageable in order to later attain prestigious careers (Mittleman, 2022; Pachankis et al., 2017).

In contrast, lesbian and bisexual girls may feel less compelled to “do femininity” than heterosexual girls (Pearson and Wilkinson, 2017, pp. 563–564). This distance from the so-called good girl norm potentially leads to lower academic achievement among lesbian and bisexual girls than among heterosexual girls (Fine, 2015; Mollborn and Everett, 2015; Walsemann et al., 2014). Academic achievement among sexual minority girls might be suppressed by the disparate treatment of school authorities, especially for sexual minority girls “who enact ‘masculine’ styles of speech, dress, and behavior,” and thus are sanctioned by their “bad girl” performance (Mittleman, 2022, p. 330). Girls who violate “good girl” norms are more likely to face sanctions, like suspension, that may have a direct negative impact on their subsequent academic success (Pearson and Wilkinson, 2018). Against this background, we propose the first hypothesis.

Education has always been at the center of the association between social origin and destination to achieve social mobility (Breen, 2010). Nevertheless, a number of studies point out that social origin not only remains the most relevant factor for educational inequality (Dumont, Klinge and Maaz, 2019), but also has a substantial direct effect on income and social class destination (Bernardi and Gil-Hernández, 2021).

Cultural reproduction literature argues that it is not only financial and formal access to educational institutions, but also the accumulated capital mediated by habitus that determines which individuals pursue academic careers and which do not (Jæger and Breen, 2016). Bourdieu (1990) argued that, through experiences in day-to-day life, people interiorize a habitus that determines their taste, values and actions (Dumont, Klinge and Maaz, 2019). Social origin shapes people's habitus through the limits that material factors pose to a person's environment (field), and also by a habitus that can emerge at a collective level (Reay, 2004), such as a habitus related to different social classes.

The middle- and upper-class habitus consists of a higher sense of entitlement, while people from the working class do not have as much confidence in themselves in later socialization (Calarco, 2011). Thus, the working-class habitus evaluates educational trajectories far from the familiar environment as too difficult, while financial burdens appear too high, which is why working-class individuals might exclude themselves from where they are already excluded (Swartz, 1997).

An individual's habitus is in constant interplay with the environment. Bourdieu (1990) argued that when an environment is encountered that is not familiar and clashes with a person's habitus, a disjuncture is created that can lead to change or transformation of a person's habitus (Reay, 2004). Such a situation is likely to arise when environments change rapidly or when people encounter new environments (with a different “temporality”), which has been labeled hysteresis (Bourdieu, 1990). Hysteresis can result in a state of reflexiveness where individuals aim to realign their habitus to the new situation (Strand and Lizardo, 2017).

Some scholars have argued that there is also a form of heterosexual habitus (Degele, 2005; McDermott, 2011; Rooke, 2007). Therefore, non-heterosexual people are more likely to encounter situations of reflexiveness as they challenge the heterosexual habitus through their sexual orientation. The reflexivity that results from these disruptions could also challenge aspects of people's habitus that have been structured by social origin. Breaking the heterosexual habitus may lead to a deliberate search for LGB safe spaces where the disrupted habitus can be realigned: “In other words, ‘gay’ subjective dispositions can align with objective conditions through safe relationships and spaces” (McDermott, 2011, p. 75). As LGB individuals seek out safe spaces where their habitus can be realigned, the resulting shifts in values, attitudes and actions may challenge aspects of their class habitus.

There are different ways in which this realignment of the habitus can affect the educational attainment of non-heterosexual people. On the one hand, the challenges related to having a minoritized sexual identity can lead people to drop out of education or focus on employment to be more independent. On the other hand, the need to find LGB peers and “safe spaces” can increase the incentives to invest in education. Given that individuals from middle- and upper-class families already hold positive attitudes towards education, a re-evaluation of the benefits of education are likely to be primarily relevant for LGB people from lower social backgrounds.

There are various mechanisms through which LGB people from working-class origins could obtain more positive attitudes towards education. We suggest that one central mechanism breaking the class habitus of LGB people is greater ambivalence with respect to relationships to LGB people's family-of-origin (Dewaele et al., 2011; Fischer and Kalmijn, 2021), which might suppress the reproduction of cultural capital (Dumont, Klinge and Maaz, 2019; Jæger and Breen, 2016). In fact, compared with straight individuals, LGB individuals are at greater risk of homelessness (Tierney and Ward, 2017), are more likely to move away from their home region and move out earlier (Fischer and Kalmijn, 2021), and have weaker ties and less frequent contact to their family-of-origin (Fischer, 2022; Kasprowski et al., 2021). Weaker family ties might not only affect social support structures for resilience, as gay and bisexual men rely less on their family-of-origin and more on their family-of-choice (Frost, Meyer and Schwartz, 2016). Moreover, differences in family and friendship ties might devalue the role of the parental cultural capital in the intergenerational transmission of education, perhaps enhancing the role of social capital gained through their family-of-choice. This would be one channel through which the class habitus is disrupted by the sexual habitus. Yet, ambivalent relationships to the family-of-origin might also translate into a loss of financial support for educational trajectories. Consequently, for middle- and upper-class gay individuals, this might translate into a loss of some privileges and possibly even relatively lower levels of education compared with their heterosexual peers from high social origins.

An alternative mechanism that could increase the educational aspirations of LGB people from working-class families includes the attractiveness of environments that can be accessed with educational degrees and the resources those provide. For instance, educational attainment can be a way of investing in resources to deal with future discrimination or to access jobs and neighborhoods that have lower levels of discrimination (Barrett, Pollack and Tilden, 2002). This can be particularly relevant for LGB people with a lower social origin as their environments are characterized by higher levels of stigma and discrimination (Mollborn and Everett, 2015). Education, and the connected social and economic resources, can also be regarded as necessary to access queer spaces and middle-class lifestyles that are central to dominant representations of queer culture (Barrett and Pollack, 2005; Rooke, 2007). For instance, queer spaces are often based in relatively expensive urban areas, and some studies have shown how working-class LGB people encounter challenges integrating within LGB spaces (Rooke, 2007).

In short, because LGB people challenge the heterosexual habitus, they encounter situations where a realignment of their habitus can lead to change or transformation (McDermott, 2011). In this process LGB individuals could be more likely to adjust their educational path according to their need to find safe spaces in more educated environments than the loss of financial support might affect their sense of distinction (Calarco, 2011) already internalized through earlier socialization. As individuals from higher social origins already have positive educational attitudes, this mechanism might be relatively less relevant for them. However, if weaker family ties are the main mechanism relating sexual identity to educational attainment, we might even observe an educational penalty among LGB people from advantaged backgrounds. In light of this we formulate the following hypothesis:

Germany is an ideal case to test the outlined relationships for several reasons. Germany has a strong relationship between social class origin and educational attainment (Betthäuser, 2020; Breen, 2010; Dumont, Klinge and Maaz, 2019; Neugebauer, 2015). Although there are differences in the education system from state to state, Germany is characterized by early stratification in the transition to the secondary education system (Roth and Siegert, 2016). The early selection after grade 4 to 6 is legitimated by the ideology of “innate talent,” ascribing pupils either practical or theoretical talents to recommend a specific school type (Powell and Solga, 2011). The highest secondary school level ends after grade 12 or 13 with a general or a specialized higher education entry certificate. For people with university entrance qualifications, the vocational education system is highly standardized and similarly attractive (Neugebauer, 2015). On the upper level, the tertiary system can be described as a binary system. The tertiary education system splits into classical (state) universities on one side and (public but also private) universities of applied sciences on the other, while the latter are also accessible with a 12-year domain-specific higher education certificate (Blossfeld, 2018).

This early and highly stratified school system with a high emphasis on affordable public schools and universities with no tuition fees makes Germany an ideal case to test differences in the cultural reproduction of straight individuals and LGB people. Even though economic considerations are likely to still play a role in the German context too, it might be more feasible for LGB people from disadvantaged backgrounds to aspire to a university education than in contexts where tuition fees form additional obstacles to break the heterosexual habitus.

The school system enables individuals from lower-tracked schools to attend Berufskollegs thereafter to gain high-school entrance certificates or to start studying after completing vocational training. To earn these certificates, there is fully funded financial aid available that makes individuals more independent from parents’ financial support. Thus, in Germany, LGB people may also be more likely to adjust their educational goals in later stages of life compared with their straight peers.

The data come from the German Socio-Economic Panel study (SOEP) 2019 (De Vries et al., 2021). The SOEP is an annual household panel survey, conducted since 1984, including around 20,000 households and up to 30,000 individuals. In more recent SOEP surveys, there are several improvements regarding sexual orientation data compared with earlier versions. First, in 2016, a direct question about the respondent's sexual orientation was introduced (Kroh et al., 2017), which allows researchers to study self-identified LGB people beyond partnered individuals. Second, in 2019, SOEP surveyed an oversample of LGB(T) individuals. For this oversample (sample Q), LGB individuals were identified by a random telephone screening of the entire non-institutionalized adult population of Germany. Respondents were selected into the boost sample by means of a self-reported sexual orientation question. The screening resulted in a gross sample of 835 individuals, with a high response rate of 58.3% for the personal interviews (CAPI; De Vries et al., 2021). The final sample size of 1,077 respondents and their inclusion into a high-quality probability-based household survey allows, for the first time, for conducting differentiated and robust comparative analyses among LGB people and heterosexual people in Germany. The existing panel provides a large control group of straight individuals. Design and survey weights are included separately for LGB and straight individuals to take into account the probability of inclusion to the CATI screenings first by mobile phone and landline contact (design weights), followed by non-response weights that consider the selective patterns of participating in the survey (De Vries et al., 2021).

Sexual orientation is measured by the question, “Would you describe yourself as …?” Respondents could choose between heterosexual (that is, attracted to the opposite sex); homosexual (gay or lesbian, that is, attracted to the same sex); bisexual (that is, attracted to both sexes); other, namely …¹; and no answer / prefer not to say (De Vries et al., 2021). We focus on heterosexual and LGB respondents, as previous research has shown that the other categories include considerable shares of heterosexual people who do not understand the question or answer options (Elliott et al., 2019). Gender is self-reported but limited to binary answer options only (woman or man). The boost sample from 2019 includes a non-binary instrument. Since this is not the case for the existing panel in 2019, the sample is overall restricted to men and women.

Taking into account the sample size of LGB people, we make use of the parsimonious specification of the parental International Socio-Economic Index of occupational status (ISEI) to measure social origin (Ganzeboom, De Graaf and Treiman, 1992). This can be regarded as a more encompassing measure of social origin than parental education, which has been used in previous research (Boertien, Perales and Pessin, 2023), because ISEI scores are based on both the average education and the average earnings of people in a given occupation. Social origin is calculated as the average value of the combined parental ISEI scores, because this approach is more accurate than the conventional dominance coding in continuous social origin measures while it also takes the other parent's social status and, thus, their educational achievements into account (Thaning and Hällsten, 2020). The ISEI ranges from 11 (low) to 89 (high). For example, 89 includes judges, whereas most working-class occupations reach up to roughly 40.

Educational attainment is measured by the Comparative Analysis of Social Mobility in Industrial Nations (CASMIN) classification. Relative to the International Standard Classification of Education (ISCED), CASMIN has a better and more pragmatic fit to estimate the heterogeneity of education in Germany and is social class theory–driven (Schneider, 2016).² To preserve statistical power, CASMIN is divided into three major groups: (1) basic (CASMIN 1abc) educational attainment, that is, no educational certificate up to basic vocational qualification; (2) middle (CASMIN 2abc) education, that is, intermediate vocational qualification up to general or vocational maturity certificate; and (3) high (CASMIN 3ab) educational attainment, that is, at least a tertiary education degree (bachelor's or higher).

Age and birth cohort are included as control variables, because both are predictors for educational attainment (Breen, 2010) and both affect the educational advantage of sexual minorities from older cohorts (Mittleman, 2022; Pearson and Wilkinson, 2017). The cohort variable is separated into five groups by ten years, except for the oldest generation, which is the 1949 or older generations. The reference category is the youngest cohort, 1980–1989. The design of cohort groups is geared towards when most individuals have completed their educational attainment. Therefore, the lower bound of the sample is restricted to those who are aged 30 or older. Next, we consider urbanization, because individuals from urban areas are less likely to be college educated than those from suburban or rural areas (Fine, 2015). Urbanization level is coded into four categories derived from the question or information from the main residence during childhood at the age of 15. It ranges from a rural area (fewer than 20,000 citizens), small town (20,001 to 100,000), urban (100,001 to 500,000), up to a metropolitan area (more than 500,000). We control for partnership status (binary coded), as it could be an important determinant of disclosure of LGB status in the survey.

The analytical sample comprises all observations that had no missing information on the dependent, independent and control variables. Furthermore, we exclude the refugee oversamples of the household panel study to limit other sources of educational disparities. Finally, we restrict the sample to individuals who do not have any missing values for all covariates that are considered. This leaves us with a sample of N = 15,746; LGB individuals = 589. The proportion of women and men identifying as LGB people can be found in Table 1. All analyses use SOEP-provided survey weights that account for design selectivity and non-response outlined above (De Vries et al., 2021). A replication package for all analyses can be found at OSF (https://osf.io/3274j/).

The first important finding is that 33% of LGB individuals, but only 27% of heterosexuals, attain tertiary education (t-test: P-value < 0.05; Table 2). We also see that only 21% of LGB individuals has at most basic education, compared with 32% of heterosexual individuals. Therefore, without considering covariates, the overall LGB population is higher educated than heterosexuals in Germany. However, important covariates that are associated with higher education also differ: The LGB population is six years younger and a smaller proportion, compared with heterosexuals, is in a relationship (Table 2). Both are relevant indicators associated with higher educational attainment (Tables 3 and 4). Older individuals are less educated than younger ones and partnered individuals are higher educated. This finding is in line with the educational expansion that took place in all European countries in the second half of the 20th century (Breen, 2010) and that partnered individuals are higher educated (Wilson and Stuchbury, 2010).

Table 3 explains educational attainment by an interaction effect of gender and sexual identity. Surprisingly, there is no substantial gender interaction effect. Therefore, we do not find support for the hegemonic gender norms hypothesis (H1), which expects higher educational attainment for sexual minority men than for heterosexual men, but lower educational attainment for sexual minority women compared with heterosexual women. Robustness checks applying gender-separated models show the same results.⁴ As a consequence, we restrict the rest of the analysis to pooled models with gender as a control variable.

In testing whether social origin might have a stronger effect on educational attainment for heterosexuals than for sexual minorities, we find that an educational advantage is observed only for LGB individuals from lower ISEI families. First, Table 4 shows that there is a statistically significant interaction effect of sexual identity with ISEI: OR = 0.98, and P-value < 0.001. Figure 1 visualizes the interaction effect and is based on the ordered logit model that includes the interaction effect, as well as all control variables in Model 2 (Table 4). The figure shows that, for heterosexuals, there is an almost exponential relationship between social origin and obtaining higher education (but the relationship tends towards a plateau at the top of the distribution). Among LGB individuals, there is an almost linear relationship between social origin and educational attainment. As a consequence, we observe that the educational premium related to having an LGB identity is observed only for individuals from lower-class backgrounds, concretely, for LGB people with a parental average ISEI score below 42 (β = -0.02, P-value < 0.001). The effect size at lower levels of ISEI is considerable. At an ISEI level of 23, for instance, 9.6% of straight individuals reach tertiary education, compared with 18.5% of LGB individuals. Hence, the effect size of 9% corresponds to almost double the probability of straight individuals.

Figure 2 presents a Johnson-Neyman plot that calculates for which parts of the social origin distribution the interaction effect is significant. The plot shows that the moderation is significant among working-class individuals up to parental ISEI scores of 51.34 and (negative) for individuals stemming from upper classes (ISEI scores between 82.25 and 89). This supports H2, the queer habitus hypothesis, which expects a weaker association of social origin on educational attainment among sexual minorities compared with the heterosexual population.⁵

In addition, although the confidence intervals of sexual minorities and heterosexuals overlap at the upper ends of the ISEI social origin, it cannot be ruled out that LGB individuals from higher ISEI families may have a slightly lower probability of achieving high educational attainment relative to heterosexuals from more upper-class backgrounds (Bernardi, Chakhaia and Leopold, 2016). Because the Johnson-Neyman plot shows that LGB individuals with privileged family backgrounds, beginning from ISEI scores of 82.25, likely receive lower educational attainment, we argue that there could also be a smaller transmission of advantage among LGB individuals from higher social origins. Thus, in general, we find support for the notion that the association of social origin and educational attainment is stronger among heterosexuals than among LGB individuals. In particular, the educational advantage found among LGB people is concentrated among those sexual minorities with low ISEI family backgrounds.

Our data are well suited to study educational attainment, but cohort studies that follow people through their educational careers might be better suited to document the mechanisms at play. Nonetheless, we ran additional analysis with available variables to gain some insight into mechanisms. These numbers, presented in Appendix Tables 1–4, show that LGB individuals are more likely to move away from their region of birth (especially among LGB individuals who grew up in more rural parts) and visit their parents less as adults. These factors can explain 21% of the differences in educational attainment between LGB and heterosexual people with lower levels of parental ISEI. Even though these numbers support to some extent the idea that links to family-of-origin are weakened among LGB individuals, one complication is that these outcomes are measured during adulthood and might therefore be the consequences of higher educational attainment, rather than vice versa. For a subsample, we had information on how often respondents indicated they had fights with parents until age 15. Half of LGB people reported fighting often with their parents, compared with 29% of heterosexual people, suggesting that weaker relations with families-of-origin are likely to have been present before transitioning towards university.

The appendix also includes additional robustness checks where different versions of the outcome variable were considered (lower educational attainment, years of education); these analyses show substantively very similar results.

This study set out to investigate to what extent the relationship between sexual identity and education is moderated by gender and social origin. Previous studies have argued that hegemonic gender norms (H1) cause LGB men to be higher educated, but LGB women to be lower educated, than their heterosexual counterparts. Further, this study introduced the queer habitus hypothesis (H2), which focuses on the intersection of social origin and sexuality, explaining the higher educational attainment of sexual minorities based on the reflexivity resulting from a challenged heterosexual habitus. Using novel SOEP data, we make the following central findings.

With regard to the first hypothesis of hegemonic gender norms (H1), we do not find support for gendered differences in educational attainment in Germany. This finding differs from that of previous research on the United States (Mittleman, 2022). Why do we find a different result in Germany? One possibility is that the German school system sanctions gender-nonconforming behavior by sexual minority girls less. Sanctioning has been argued to be a mechanism underlying the lower educational attainment of lesbian and bisexual women in the US (Mittleman, 2022). Another possibility is that gender differences in education are smaller in Germany in general. As observed in Table 1, there were barely any differences in educational attainment between heterosexual men and women in our sample. This absence of gender differences in educational attainment might also imply that other processes related to gender play out less strongly in the German context. Future research can investigate how the interaction between sexual identity and gender develops in younger cohorts where the gender gap in education favors women more strongly.

With regard to the second hypothesis, queer habitus (H2), we find evidence for a stronger association of social origin on educational attainment among heterosexuals. One explanation for this might be that LGB individuals from lower-class backgrounds who are socially immobile “remain in the closet” (Barrett and Pollack, 2005; Mollborn and Everett, 2015). This would mean that our result of a higher likelihood for high education among LGB individuals from the working class is slightly overestimated. Previous research has addressed this possibility and concluded that selection is unlikely to explain the educational premiums observed for sexual minorities (Fine, 2015; McQuillan, 2017; Mittleman, 2022; Mollborn and Everett, 2015). Appendix Table 5 shows that in our sample there are also no notable differences in social origin between LGB and heterosexual people.

Instead, we argue for an alternative explanation for the higher educational attainment among sexual minorities: The disruption of the heterosexual habitus can lead to a re-evaluation of attitudes towards education and increase the perceived benefits of education among LGB people from lower social backgrounds (McDermott, 2011; Rooke, 2007). Because sexual minorities experience the need to realign their habitus to the need to find “LGB safe spaces,” they might be more likely to pursue higher education.

It could be unsurprising that the educational “advantage” is concentrated among LGB individuals from working-class backgrounds for three reasons. First, all sexual minorities are in need of finding LGB safe spaces. If such safe spaces are generally easier to access through gaining high levels of education, and if gaining education is part of the class habitus for individuals from advantaged social backgrounds, gaining education might not contest the class habitus for them. Hence, the disruption of the heterosexual habitus might more powerfully affect the class habitus of LGB individuals from working-class origins (McDermott, 2011).

Second, a similar process might be put into motion if LGB individuals from working-class backgrounds experience a greater need to “escape” (Mollborn and Everett, 2015) intolerant families and environments (Barrett and Pollack, 2005), which would result in a greater motivation compared with heterosexual working-class people to “invest” in their education.

Third, given their upper-class social origins, LGB people from such social origins are more likely to pursue higher education regardless (Jæger and Breen, 2016). However, we found some evidence that LGB individuals from the most advantaged social origins are lower educated than their heterosexual counterparts. This would support the argument that weakened ties to the family-of-origin are an important mechanism at play, as LGB individuals from higher social origins might have more to lose from weakened ties. If upper-class families normally protect their children from downward mobility, also labeled as “compensatory advantage” (Bernardi and Gil-Hernández, 2021), this “parachuting” could be disabled when ties to families-of-origin are contested. We indeed observe that LGB individuals from the most advantaged social origins are lower educated than their heterosexual counterparts. Hence, the evidence presented in this paper shows that the link between social origin and educational attainment is weaker among LGB individuals than among heterosexual individuals.

Finally, LGB people might also attain education to accumulate resources to deal with future discrimination (Barrett, Pollack and Tilden, 2002). Indeed, there are similarities between LGB people and other disadvantaged groups in society who obtain high levels of education but nonetheless have relatively low earnings such as women and migrants. Previous studies have shown how migrants and ethnic minorities have high educational aspirations (Feliciano and Lanuza, 2017; Kao and Thompson, 2003) and how girls outperform boys in school, particularly among children with a lower social origin (Entwisle, Alexander and Olson, 2007). Indeed, this observation underlines that high levels of educational attainment could be interpreted as a reaction to experienced disadvantage, rather than reflecting an advantaged position in society more generally. However, the lower educational attainment among LGB individuals of advantaged social origins suggests that increased incentives to attain education are not sufficient to explain the educational outcomes of LGB people. Future research can investigate to what extent we also observe premiums among LGB people from lower socioeconomic backgrounds when it comes to outcomes such as income or status, and whether these premiums are also less visible among LGB people from more advantaged backgrounds.

The evidence we found must be contextualized within the German educational system. LGB people are less dependent on their parents’ economic capital compared with other countries that have been studied such as the US. In Germany, the educational system, which struggles with traditionally high intergenerational transmission (Breen, 2010), is comparatively favorable financially compared with the US. LGB people may have more educational opportunities because of ambivalent family-of-origin relations (Dewaele et al., 2011; Fischer and Kalmijn, 2021), as major financial support from parents is less crucial for studying at state universities without tuition fees. Hence, the higher educational chances for LGB individuals from lower ISEI origins might be specific to Germany or contexts with no or low fees to access university. Further cross-country comparisons could test this possibility in future research. Initial evidence from a comparative study also found higher upward educational mobility among LGB individuals in Australia, the UK and the US (Boertien, Perales and Pessin, 2023). But more cross-nationally comparable data are needed to come to firmer claims in that regard.

As with every study, some limitations remain. First, while there has never been such a large and representative sample of LGB individuals in Germany, the sample is still too limited to estimate differences between bisexual and gay individuals without a loss of statistical confidence (see appendix). On the one hand, bisexual individuals might be considered to be more gender typical and less visibly identified as part of a minoritized group. However, previous research has shown how bisexual individuals deal with particular stereotypes and have relatively low levels of earnings and well-being (Mize, 2016). How these disadvantages play a role in educational careers is less clear. In other words, the specific challenges faced by lesbian, gay and bisexual individuals, as well as other sexual identities that were not part of this study (e.g., queer, pansexual), should be an objective for future research. Second, because the focus of the study was on the intersections of sexual orientation, gender and social origin, the call to consider the timing of outing (Fine, 2015; Mollborn and Everett, 2015; Pearson and Wilkinson, 2017) could not be answered here.

With regard to future research in this area, a priority would be to study the proposed mechanisms of weaker family-of-origin ties and stronger family-of-choice ties that might mediate the weaker association among sexual minorities through a disrupted heterosexual habitus. Thus, if research finds support for underlying mechanisms, the proposed theoretical framework might help to understand the higher educational attainment within the context of (covert) homophobia and would provide further evidence against a “pure” selection explanation for the higher levels of education among LGB individuals (Barrett and Pollack, 2005; Mishel, 2019).

Even though qualitative research has already investigated the intersection between class and sexuality (McDermott, 2011; Rooke, 2007), these mostly shed light on the challenges that working-class LGB people experience in the educational system, such as the need to abandon education to gain greater independence. Qualitative studies of upwardly mobile LGB people could greatly help in understanding the patterns we observe here, where LGB persons from lower social origins attain high levels of education.

This study provides robust evidence of a weaker “origin-to-education” association among sexual minorities. As it is argued that education is the major mediator to foster social equality (Betthäuser, 2020; Breen, 2010), sexual minorities should be better off in labor market outcomes. Yet, sexual minorities face income penalties compared with heterosexual men (Drydakis, 2022) and face a “gay glass ceiling” in terms of higher social status positions (Aksoy et al., 2019). Therefore, closer investigations of the other paths within the O-E-D (origin-education-destination) triangle should be investigated. Because the educational advantage of LGB individuals is concentrated among those from lower working-class backgrounds, next to discrimination on the labor market (De Vries, 2021), it might be that the “long shadow of social class” (Bernardi and Gil-Hernández, 2021; Laurison and Friedman, 2016) overshadows the gay educational achievements and leads to relatively low earnings in the labor market.

Finally, in light of ongoing debates about diversity, this study indicates that an abbreviated understanding of diversity policies designed to help abolish discrimination based on sexual orientation only, but overlooking intersectionality, might not directly tackle the hidden mechanism of upper-classes cultural matching and class discrimination processes. Although the relationship between social origin and educational attainment is weaker for sexual minorities, straight upper-class individuals remain privileged in educational attainment whereas for sexual minorities, their sexuality seems to pose disadvantages compared with their upper-class peers. Therefore, working-class LGB individuals, just like other working-class individuals, would benefit greatly from policies aimed at abolishing class privileges and fostering class equality. At the same time anti-discrimination policies and funding for LGBT education programs would help LGB individuals to overcome the challenges to find safe spaces.

AI was not used for this article.

A replication package for all analyses can be found at OSF (https://osf.io/3274j/).

The authors do not have any conflicts of interest.

No ethical concerns were identified in the production of this article, which relies on anonymous secondary data sources.

Supplemental data for this article can be accessed online at https://doi.org/10.1162/euso_a_00019.