To ensure the validity and reliability of the measures used in this study, all three instruments -BCASSFS, Beliefs about the Etiology of Sexual Orientation (BESO), Opinions About the Rights of Individuals with a Homosexual Sexual Orientation (ORHSO) were translated and adapted into Greek following the structured process for cross-cultural adaptation of psychometric tools (Sousa & Rojjanasrirat, 2011). Prior to translation, we obtained formal permission from the original developers of the BCASSFS to translate and distribute the tools for research purposes. Each scale was independently translated into Greek by two bilingual experts familiar with LGBTQIA+ issues and psychometric assessment. The Greek translations were then back translated into the original language by two additional bilingual experts who were blinded to the original items. Each instrument was then piloted on a small sample (n = 5) of Greek participants from diverse demographic backgrounds. Feedback was collected regarding the clarity and comprehensibility of the items, and minor revisions were made to address any issues identified. For the BESO, and ORHSO, although no previous formal Greek validation existed, we performed reliability analyses and ensured internal consistency for both scales within the current sample. Additionally, convergent and discriminant validity analyses were conducted to evaluate their construct validity in relation to the BCASSFS.

A total of 378 people were included in the study. The inclusion criteria were: (a) being adults aged 18–65 years living in Greece, and (b) self-identifying as exclusively heterosexual based on their responses to the Kinsey scale (value 0: exclusively heterosexual; Kinsey et al., 2003). Of the 378 participants, 17 were excluded because they indicated a sexual orientation other than exclusively heterosexual.

Participants were recruited from February to April 2022. The virtual snowball sampling technique was employed (Baltar & Brunet, 2012) via Google Forms platform. Participants were initially selected from researchers’ professional and academic networks, including colleagues, university students, and contacts from LGBTQIA+ advocacy groups. These individuals were asked to share the study link with their personal networks, promoting participation across various demographic groups. No monetary incentives were offered for participation. Instead, participants were informed of the study’s academic and social importance, emphasizing their contribution to understanding attitudes toward same-sex parenting in Greece. Participants were then asked to fill out the informed consent form. They were assured about the anonymity and confidentiality of their responses and the right to withdraw in accordance with the Declaration of Helsinki (World Medical Association, 2013). Subsequently, they provided information on their demographic characteristics and responded to the study measures in the following order: (a) the BCASSFS, (b) the BESO scale, (c) the ORHSO scale, (d) the ATLG scale, (e) support for gay rights items, (f) political ideology items and (g) religiosity items. The average participation time was 20 minutes.

Of the initial 378 participants, 17 were excluded as they had indicated a sexual orientation other than exclusively heterosexual. Of the remaining 361 participants, 173 were males (47.92%) and 188 females (52.08%), aged 18–65 (M = 26.16, SD = 8.12). No gender differences regarding demographic characteristics were observed, apart from educational level (Fisher’s exact test < .001), frequency of contact with LGBTQIA+ people, t(180) = 2.38, p < .05, and contact with LGBTQIA+ couples, χ²(1) = 4.88, p < .05. Table 1 shows participant characteristics.

Absolute skewness values were found to be less than 3.0 (Range: 0.05 to 0.83), while absolute kurtosis values were found to be less than 8.0 (Range: 0.13 to 1.03), suggesting normal distribution for all items (Kline, 2005). No missing values or outliers were detected. Outliers were assessed using the z-scores technique (-3 < z scores < 3).

One of the main goals of this study was to establish the factorial validity of the BCASSFS for the general Greek population. To achieve this, four models were tested (method: maximum likelihood, free estimation). First, a one-factor model was examined to determine whether the construct could be simplified to a unidimensional measure, with all items loading onto a single latent factor. Second, a two-factor correlated model was tested, based on the theoretical background and findings from previous studies. In this model, the items were grouped into two correlated factors: Individual Opposition (IO) and Normative Opposition (NO). Third, a bi-factor model was assessed. In this model, a general factor (General Opposition to same-sex parenting) accounted for the shared variance across all items, while the two specific factors (IO and NO) explained additional unique variance. This model examined whether the two specific factors contributed meaningful variance beyond the general factor. Finally, a hierarchical model was applied to explore the relationship between the two first-order factors (IO and NO) and to test whether these two subfactors are nested under a broader construct (General Opposition). The results for all models are presented in Table 2.

The comparison of the models showed that the two-factor correlated model, and the bi-factor model yielded the best fit indices. The bi-factor model demonstrated slightly better fit on incremental indices (CFI, TLI, SRMR), whereas the two-factor correlated model exhibited superior parsimony (lower AIC) and better absolute fit (lower χ² and RMSEA). Overall, although the bi-factor model is statistically plausible, the two-factor correlated model aligns more closely with the literature supporting the distinctiveness of IO and NO as related but separate constructs (see Figure 1), as proposed by the scale’s authors (Frias-Navarro & Monterde-i-Bort, 2012). In addition, it should be noted that the fit indices are identical to those reported in the study by Barrientos et al. (2013) (TLI = 0.94, CFI = 0.96, and RMSEA = 0.07). Cronbach α coefficients for BCASSFS total were found 0.92 and 0.90 for IO and 0.88 for NO. Item-total correlations were found to reflect acceptable internal consistency (α > .88), while all inter-item correlations were more than .25 (p < .001). Therefore, the two-factor correlated model is preferred while each factor had high internal consistency, confirming H1.

To demonstrate the construct validity of the BCASSFS, both convergent validity and discriminant validity were examined (H2 to H4). Previous research suggests that there are strong correlations between high scores in political identity (political conservatism) and opposition to LGBTQIA+ people, support of gay rights and same-sex parenting. In the present study we controlled political identity and examined these latter correlations. With respect to convergent validity, the BCASSFS was expected to show significant positive correlations with religiosity, because religiosity often aligns with traditional values and is associated with greater opposition to same-sex parenting. Additionally, it was hypothesized that negative attitudes toward same-sex families would show significant negative correlations with the belief that sexual orientation is learned. Strong negative correlations with low scores in political identity (liberal political opinions), general positive attitudes toward lesbians and gay men, and support of LGBTQIA+ rights were also expected, as these constructs are theoretically opposed to prejudice against same-sex parenting.

The hypotheses were partially confirmed, as presented in Table 3. Religiosity showed a significant positive correlation with both Individual Opposition (r = .25, p < .001) and Normative Opposition (r = .12, p = .007), supporting the hypothesis that higher religiosity aligns with greater opposition to same-sex parenting. Political Identity (where higher scores indicate more right-wing political views) showed a significant positive correlation with Individual Opposition (r = .27, p < .001), but its correlation with Normative Opposition was weaker and nonsignificant (r = .10, p = .070). Further, as hypothesized, Learned Etiology of sexual orientation demonstrated strong positive correlations with both Individual Opposition (r = .77, p < .001) and Normative Opposition (r = .51, p < .001). This finding indicates that participants endorsing the belief that sexual orientation is learned expressed significantly higher levels of opposition to same-sex parenting. Similarly, Genetic Etiology of sexual orientation showed a weak positive correlation with Individual Opposition (r = .20, p < .001), but no significant relationship with Normative Opposition (r = .06, p = .233). General attitudes toward lesbians and general attitudes toward gay men were significantly and negatively correlated with both Individual Opposition (r = -.76 and r = -.82, p < .001, respectively) and Normative Opposition (r = -.52 and r = -.52, p < .001, respectively). Support of gay rights was also negatively correlated with both Individual Opposition (r = -.78, p < .001) and Normative Opposition (r = -.51, p < .001), further supporting the hypothesized relationships.

For discriminant validity, it was expected that variables such as Genetic Etiology of sexual orientation and political identity would show weaker relationships with BCASSFS scores compared to attitudinal variables like support of gay rights or general attitudes toward lesbians and gay men. The results supported this distinction: the correlation between Genetic Etiology and Normative Opposition was nonsignificant (r = .06, p = .233). The correlation between political identity and Normative Opposition was also nonsignificant (r = .10, p = .070). These findings indicate that the BCASSFS is not conflating unrelated constructs and remains distinct from weaker predictors, providing evidence for discriminant validity. It should also be noted that political identity, the control variable, has a minor impact on these correlations, slightly reducing their strength but not reversing their direction or significance.

A multiple linear regression analysis was conducted to examine the predictors of Individual and Normative Opposition scores. The predictors included gender, religiosity, having contact with LGBTQIA+ couples, political identity, genetic and learned etiology for sexual orientation, support of gay rights and general attitudes towards lesbians and gay men.

For the Individual Opposition, the regression model was statistically significant, F(9, 351) = 141.83, p < .001, and accounted for 78.4% of the variance in Individual Opposition scores (R² = .784, Adjusted R² = .779). The results of the model fit indicate that the regression analysis explains a substantial amount of the variance in the dependent variable. Learned Etiology (β = .34, p < .001) was the strongest positive predictor of Individual Opposition, suggesting that participants who endorsed the belief that sexual orientation is learned exhibited significantly higher levels of Individual Opposition. Support of gay rights (β = -.19, p < .001), General Attitudes Towards Lesbians (β = -.22, p < .001), and General Attitudes Towards Gay Men (β = -.25, p < .001) were also significant predictors, indicating that more favorable attitudes toward LGBTQIA+ rights and individuals predicted lower Individual Opposition. The other predictors, gender, religiosity, having contact with LGBTQIA+ couples, political identity, and Genetic Beliefs, were not statistically significant (p > .05). The results suggest favorable attitudes toward LGBTQIA+ rights, and general attitudes toward lesbians and gay men significantly predict lower Individual Opposition whereas etiology about sexual orientation being learned, predict higher. These findings highlight the role of attitudinal and belief-based predictors in shaping opposition toward same-sex parenting.

For the Normative Opposition, the regression model was statistically significant, F(9, 351) = 22.79, p < .001, and accounted for 36.9% of the variance in Normative Opposition scores (R² = .369, Adjusted R² = .353). The model explained a moderate amount of the variance in Normative Opposition. Once again, learned etiology (β = .32, p < .001) emerged as the strongest positive predictor, suggesting that participants who endorsed the learned etiology about sexual orientation exhibited significantly higher normative opposition. General Attitudes Toward Lesbians (β = -.27, p = .001) and Genetic Etiology (β = -.11, p = .023) were also significant predictors, indicating that more favorable attitudes toward lesbians and endorsement of genetic explanations for sexual orientation predicted lower normative opposition. The other predictors, gender, religiosity, having contact with LGBTQIA+ couples, and political identity, were not statistically significant (p > .05). Similarly, General Attitudes Toward Gay Men showed no meaningful contribution to the model (p = .994). The results indicate that favorable attitudes toward lesbians, and genetic explanations for sexual orientation significantly predict lower Normative Opposition scores while beliefs about sexual orientation being learned predict higher. Overall, the stronger explanatory power of predictors in the Individual Opposition model suggests that explicit prejudice is more directly tied to attitudinal and belief-based variables (e.g., learned etiology, general attitudes toward LGBTQIA+ individuals). Normative opposition, on the other hand, appears to be a more complex construct, potentially influenced by broader cultural or contextual factors that warrant further exploration. Results are shown in Table 4.

Two independent samples t-tests were conducted to compare scores on Individual Opposition and Normative Opposition between male and female participants (H5). Male participants scored higher than female participants on both Individual Opposition (Males: M = 3.21, SD = 0.89, Females: M = 2.82, SD = 0.83), t(359) = -4.30, p < .001, Cohen’s d = 0.45, and Normative Opposition (Males: M = 3.54, SD = 0.67, Females: M = 3.24, SD = 0.70), t(359) = -4.04, p < .001, Cohen’s d = 0.43. These results suggest that gender plays a significant role in shaping opposition, with males exhibiting greater opposition than females.

For testing H6, we employed a paired samples t-test. The highest scores were obtained for Normative Opposition. The mean score on the Individual Opposition scale was 3.01 (SD = 0.88), whereas on the Normative Opposition scale, was 3.38 (SD = 0.70). The paired samples t-test revealed that the observed difference was statistically significant, t(360) = -10.06, p < .001, Cohen’s d = .70. Therefore, it is confirmed that the two constructs are empirically distinct and that societal-level concerns (Normative Opposition) are significantly more prevalent among participants.

The confirmatory factor analysis supported the hypothesized two-factor structure of the BCASSFS, distinguishing between Individual Opposition (IO) and Normative Opposition (NO). These findings are consistent with prior validations of the scale (Frias-Navarro, 2009; Frias-Navarro et al., 2017) in other sociocultural contexts, such as Spain (Frias-Navarro & Monterde-i-Bort, 2012), Chile (Barrientos et al., 2013) and Italy (Badenes-Ribera et al., 2020). Reliability analyses further underscored the internal consistency of the scale, with high Cronbach’s alpha values for both subscales, replicating previous studies (Frias-Navarro et al., 2017).

Additionally, confirmatory factor analyses were conducted to compare alternative models to determine the best representation of the BCASSFS’s structure in the Greek context. Among the models tested, a bi-factor model and a correlated two-factor model were of particular interest. The bi-factor model assumes that all items contribute to a general factor (e.g., overall opposition to same-sex parenting) while simultaneously accounting for variance unique to each subscale, namely Individual Opposition (IO) and Normative Opposition (NO). This model demonstrated marginally superior fit indices, suggesting that a single overarching construct explains much of the shared variance in the scale items. However, the correlated two-factor model, which conceptualizes IO and NO as distinct yet related constructs, better aligns with the theoretical distinction between personal biases (IO) and societal-level concerns (NO). This distinction is critical because IO reflects explicit, direct opposition often rooted in personal beliefs, such as moral or religious objections, whereas NO captures more implicit, socially acceptable forms of prejudice, such as concerns for children’s well-being in non-traditional families. While both factors are interconnected, treating them as separate constructs provides a deeper understanding of the multifaceted nature of prejudice. The decision to retain the correlated two-factor model underscores the importance of acknowledging the dual dimensions of opposition toward same-sex parenting. By doing so, researchers can better differentiate between individual-level attitudes and broader societal norms that influence prejudice. These findings highlight the interdependence of personal and societal opposition, illustrating that while they share common roots, they are manifested with different attitudes and behaviors.

The validation of the BCASSFS structure within the Greek sociocultural context is particularly significant. Greece represents a unique case where traditional cultural and religious values coexist with relatively recent advancements in LGBTQIA+ rights (Daoultzis & Kordoutis, 2024), such as the legal recognition of same-sex civil unions and same-sex marriage. These opposing forces create a dynamic environment where prejudice is more likely to be influenced by both conventional conservative norms and emerging societal changes. The ability of the BCASSFS to distinguish between IO and NO within this context confirms its utility in capturing the complex interplay of individual beliefs and societal expectations.

Convergent validity was demonstrated through significant correlations with theoretically relevant variables, in line with previous studies (Cárdenas et al., 2012; Corrêa-Ribeiro et al., 2019; Frias-Navarro & Monterde-i-Bort, 2012; Frias-Navarro et al., 2017). As hypothesized, religiosity positively predicted both IO and NO scores. This finding aligns with prior research linking religiosity to traditional family values and opposition to same-sex parenting (Baiocco et al., 2018). Interestingly, the correlations were stronger for IO than for NO, suggesting that religiosity may more strongly influence personal biases than societal-level concerns.

Support for LGBTQIA+ rights and positive attitudes toward lesbians and gay men were negatively correlated with both IO and NO. These results underscore the importance of advocacy and inclusive attitudes in reducing prejudice, consistent with prior findings (Frias-Navarro et al., 2014). From a philosophical perspective of particular interest is the finding that Learned Etiology of sexual orientation positively correlated with opposition to same-sex parenting. As Frias-Navarro et al. (2017) stated, the belief that homosexuality is a learned behavior is associated with higher levels of disapproval of gay and lesbian parenting. This finding is also supported by the attributional theory of stigma (e.g., Hegarty & Golden, 2008), which suggests that attitudes toward same-sex parenting are influenced by the perception of the controllability of homosexuality (as learned behavior).

Discriminant validity was supported by weaker and nonsignificant correlations between certain predictors and BCASSFS scores. For example, Genetic Etiology of sexual orientation showed a weak positive correlation with IO while there was no significant relation with NO, indicating its limited explanatory power. Despite the recent popularity of genetic explanations in general, the Genetic Etiology of homosexuality may still appear too controversial or complex to be accepted as plausible by a significant number of laypeople of diverse ideopolitical origins (Sheldon et al., 2007), hence its low approval in the present data, M = 2.65, SD = 1.18 compared to M = 2.91, SD = 1.01 for Learned Etiology on a scale 1–5, t(360) = -2.61, p = .009, Cohen’s d = 0.14, and the absence or weak associations with opposition to same-sex parenting. Similarly, political identity (where higher scores indicate more conservative views) had a significant but weaker relationship with IO and a nonsignificant association with NO. These findings suggest that while political conservatism may influence personal biases, societal-level concerns about same-sex parenting transcend political ideology (Grigoropoulos, 2019, 2022; Grigoropoulos & Kordoutis, 2015).

The analysis revealed significant gender differences, with male participants scoring higher on both IO and NO compared to female participants. These findings align with prior research indicating that men are more likely to express prejudice toward LGBTQIA+ individuals and same-sex parenting (Grigoropoulos & Kordoutis, 2015). This gender gap may stem from traditional gender norms and the reinforcement of heteronormativity, which are more strongly associated with male identity. Furthermore, NO scores were significantly higher than IO scores across genders, reflecting the socially acceptable nature of societal-level concerns about same-sex parenting. This supports the theoretical framework of modern prejudice, which posits that biases often persist under socially acceptable justifications, such as concerns for children’s well-being (Morrison & Morrison, 2002).

Religiosity was correlated with opposition to same-sex parenting and more strongly so with IO. The finding highlights the pervasive influence of religion in shaping personal biases against same-sex parenting in Greece, a country where Orthodox Christianity plays a central role in cultural and societal norms (Voultsos et al., 2019). On the other hand, the weaker correlation between religiosity and NO suggests that societal concerns about same-sex parenting may not be solely rooted in religious beliefs but could also reflect broader cultural values.

The relationship between political identity and opposition was weaker than expected. While conservatism was associated with higher IO scores, its association with NO was nonsignificant. These results in part deviate from prior studies linking conservative political views to strong opposition to same-sex parenting (Baiocco et al., 2018). One possible explanation is the influence of recent sociopolitical changes in Greece, including on the one hand the bipartisan support of the legal recognition of same-sex civil union and marriage and on the other an apprehensive stance of a significant number of parliamentarians across parties specifically towards same-sex parenthood (Tryfonidou, 2024). Societal concerns about same-sex parenting seem to be to a certain extent disconnected from the conservative-liberal continuum.

Several limitations should be acknowledged. First, the sample’s reliance on snowball sampling and its overrepresentation of younger and urban populations limits the generalizability of the findings and future studies should aim to recruit more diverse and representative samples. Second, the exclusion of non-heterosexual participants, while aligning with the study’s focus on majority-group attitudes, may have limited the exploration of diverse perspectives. Third, the use of a single-item measure for political identity may have constrained the depth of analysis. Future research should incorporate multi-item scales to capture this construct more comprehensively. Additionally, longitudinal studies could examine how attitudes toward same-sex parenting evolve over time in response to sociopolitical changes. Fourth, the order in which the survey items were presented may have created serial effects (priming negative attitudes towards same-sex parenting, since all participants first responded to the BCASSFS). Lastly, even though we kept the term “same-sex” as in the original scale, the term “same-gender” is more accurate and should be used in the future. In this line, we also kept the sole distinction between gay and lesbian parents, excluding many other identities (e.g., queer, intersex, trans). Future research should take into consideration all the above and develop standardized scales relating to attitudes toward same-gender family rights and investigate the link between attitudes and behaviors. Finally, although most participants did not personally know same-sex couples wanting children (85.7% for men and 66.7% for women), nearly 1 in 10 participants who did know such couples reported this desire. This finding underscores the importance of studying same-sex parenting in Greece and provides a compelling rationale for further research.

The study attempted to validate the Greek version of the BCASSFS, confirming its psychometric properties and identify the main factors predicting its subscales, IO and NO. Nonetheless, for investigating such complex associations, alternative research designs might be more appropriate, such as semi-structured interviews and the use of vignettes. As suggested also by Webb and colleagues (2020), the use of vignettes that describe several sexually diverse family scenarios (e.g., “traditional family”, “single-parent family”, “same-gender family”) may be more effective in probing participants’ attitudes and perform meaningful comparisons. Having as reference point the term “gender” and not “sex”, can result in more precise identification and capture the diversity of the sample, along with societal changes in this regard. As Lefkowitz and colleagues (2014) found, taking into consideration merely sex (and not multiple gender identities) is restrictive and insufficient when examining attitudes.

The validated Greek BCASSFS provides a valuable tool for assessing attitudes toward same-gender parenting in future research and intervention programs. Educational campaigns targeting traditional beliefs and promoting positive contact with LGBTQIA+ individuals may help reduce both IO and NO. Fostering positive attitudes towards LGBTQIA+ individuals is crucial, as recent research has shown that homophobic prejudice can have significant negative impacts on the mental and physical health of LGBTQIA+ individuals, as well as their social and economic well-being (Meyer, 2013). Additionally, addressing misconceptions about the developmental outcomes of children raised by same-sex parents could mitigate societal concerns. For example, highlighting findings from meta-analyses that show no significant differences in child development between same-sex and different-sex parent families (e.g., Golombok et al., 2018) could help counter Normative Opposition.

Greece undergoes a transitional phase with respect to marriage, as adherence to traditional family and parenthood models is declining and families are socially allowed to be more diverse (e.g., single-parent families). At the same time, social tolerance or even acceptance towards same gender parents is expanding among the population although there are still people with strong oppositional views. In recent years, Greece has shifted away from a past characterized by authoritarian politics, and the dominant role of the Orthodox Church (Avdela et al., 2020). Traditional (religious) marriage coexists with alternative forms of partnerships such as civil unions and civil marriages, including those open to same-sex couples. In this regard, studies and interventions in the Greek context, aimed at attitudes towards LGBTQIA+ people and alternative forms of partnerships including parenthood, are essential to further curtail outright and modern prejudice, social stigma, minority pressure and normative oppression.