Applying Multilevel Modeling to Understand Individual and Cross-Cultural Variations in Attitudes Toward Homosexual People Across 28 European Countries

Abstract

Individual and cross-cultural factors associated with attitudes toward homosexual people were examined in this study. Using cross-sectional data from the sixth biennial European Social Survey, which represents 36,959 individuals nested within 28 European countries, successive nested models were tested using multilevel modeling (MLM). Results found that attitudes varied cross-culturally as a function of people’s country of residence—this clustering effect was controlled for in all subsequent models. Individual-level predictors (Level 1) of male gender, older age, less education, being an immigrant to one’s residing country, conservative political affiliation, high religiosity, perceptions that politics in one’s country were unfair, low openness to change values, low self-transcendence values, high conservation values, and high self-enhancement values were significantly linked with anti-homosexuality attitudes. At the country level (Level 2), a high emphasis on social conservatism and fewer civil rights for homosexuals was connected with more unfavorable attitudes. Findings indicate main effects of predictors at both levels; however, country-level variables tended to yield stronger coefficients than individual-level factors, highlighting the contributions of macro- and microfactors in simultaneously shaping attitudes toward homosexuality. Beyond these effects, interactions of country- and individual-level variables show political affiliation, religiosity, self-enhancement values as stronger predictors in liberal countries, but openness to change values, younger age, and higher education as stronger predictors in conservative countries. Implications are discussed for understanding the wide continuum of views toward homosexuality across people and countries.

Keywords

multilevel modeling attitudes homosexuality social policies homophobia discrimination political cultures

The topics of homosexuality and the legalization of gay marriage are subject to heated social and political debate in many European countries (Kuntz, Davidov, Schwartz, & Schmidt, 2015; van den Akker, van der Ploeg, & Scheepers, 2012). Eastern and Central Europe have continued resistance against gay and lesbian rights, whereas Western European countries adopt the ideology that homosexuals should be granted the same civil rights as heterosexuals (Stulhofer & Rimac, 2009). In line with these findings, research supports the existence of cross-national variations in people’s attitudes toward homosexuality (ATH; Kuntz et al., 2015; Morrison, Morrison, & Franklin, 2009; van den Akker et al., 2012). Considering that individual attitudes regarding an object or topic are shaped concurrently by their personal beliefs and social environment (Lewin, 1951), a comprehensive micro- and macro-approach to examining factors that contribute to ATH should generate insights and advance research in this area. The current study, applying multilevel modeling (MLM), is designed to investigate the collective and unique associations of individual and cross-cultural factors with ATH in a representative sample of people residing across 28 European countries.

Individual-Level Variables

At the individual level, socialization theories argue that person-to-person differences are responsible for the spectrum of attitudes about morality (Kohlberg & Hersh, 1977). Demographic characteristics such as younger age (Kuntz et al., 2015; Wojcieszak & Price, 2010), female gender (Poteat & Mereish, 2012; van den Akker et al., 2012), and higher educational attainment (van den Akker et al., 2012; Wojcieszak & Price, 2010) are related to the acceptance of homosexual people. In addition, immigration status is linked with ATH, with immigrants less accepting of homosexuality (Röder, 2015). Furthermore, personal beliefs, such as less religiosity (Poteat & Mereish, 2012) and a more liberal political affiliation (Steffens & Wagner, 2004), have been documented to be connected with less prejudiced attitudes about homosexuality.

Similarly, political fairness perceptions have been postulated to guide attitude development and formation. These perceptions may influence people’s thoughts and feelings about their country’s governing body and reactions to critical political issues. Believing that the political process is fair compels support of, and compliance with, governmental laws and decisions (Lind, Kulik, Ambrose, & de Vera Park, 1993), whereas believing that the system is unfair induces protest attitudes and behaviors (Vermunt, Wit, van den Bos, & Lind, 1996). Governments deemed to be equitable might conjure beliefs that other people have been treated equally, whereas perceiving one’s governing body as unfair might prompt less tolerant attitudes toward minority and marginalized groups and objections to equality.

Basic Human Values

The considerable pivotal role of personal values on attitudes has also been demonstrated in previous research (e.g., Davidov & Meuleman, 2012; Kuntz et al., 2015; Sagiv & Schwartz, 1995; Schwartz, Caprara, & Vecchione, 2010). Values are core beliefs that transcend situations and contexts, and help guide people’s decisions and worldviews (Hitlin & Piliavin, 2004). Values are reasoned to develop during early socialization processes and remain relatively stable across the life span (Bardi, Buchanan, Goodwin, Slabu, & Robinson, 2014). In contrast, attitudes represent beliefs that are specific to issues and situations. Compared with values, attitudes tend to be more malleable, less stable across the lifetime, and are less central to one’s self-concept (Hitlin & Piliavin, 2004). Theoretically, values serve as motivational underpinnings for the formation of attitudes about a topic (Schwartz et al., 2012). Thus, the study of esteemed values held by individuals can provide insights on ATH.

Schwartz (1992) put forth that 10 basic values could be useful for understanding how people around the world think and behave: universalism, benevolence, tradition, conformity, security, power, achievement, hedonism, stimulation, and self-direction. Together, these 10 subordinate values comprise four higher-order human value constructs (Davidov, Schmidt, & Schwartz, 2008): openness to change, conservation, self-enhancement, and self-transcendence.

The openness to change value dimension encompasses self-direction, stimulation, and hedonism,¹ and is defined as having autonomous thoughts and actions, and receptivity to novel experiences. The value of conservation is characterized by conformity, tradition, and security, and is conceptualized as compliance with traditional values and customs. Self-enhancement values are comprised of achievement and power, and are defined as placing importance and concern on self-interests and personal enrichment of status. Furthermore, the value of self-transcendence captures universalism and benevolence, and is operationalized as the concern for the welfare of others including those who have been marginalized (Davidov et al., 2008; Schwartz, 1992; Schwartz et al., 2010).

Research by Kuntz and colleagues (2015) supports that people’s values serve as motivational guides to ATH, showing that openness to alternative lifestyles and the acceptance of goals pursued by others (openness to change values), and support of justice for others (self-transcendence values) are associated with more approving ATH. In contrast, people who embrace authority, conformity, and traditional conceptualizations of family and society (conservation values), and pursue status and prestige (self-enhancement values) are less tolerant of homosexuality.

Group-Level Variables

A nation’s collective norms represent socially acceptable implicit and explicit standards used to judge the appropriateness of personal attitudes and behaviors in a group context (Cialdini & Trost, 1998). Social climates, legal institutions, and cultural values shape prevailing norms (Inglehart & Baker, 2000; van den Akker et al., 2012), and societal norms influence personal attitudes as early as childhood (Monteiro, França, & Rodrigues, 2009). Furthermore, intergroup biases emerge when a group’s beliefs and actions are challenged by the well-being or goal attainment of out-group members (Riek, Mania, & Gaertner, 2006). Accordingly, the predominant population of heterosexuals in any nation may consider homosexuals to be a threat to their esteemed beliefs and values. As such, the examination of cross-cultural variations, over and beyond individual differences, is expected to provide insights into the contextual role of social environment in ATH.

Societal-level social conservatism and the extent that civil rights are granted to homosexuals have been identified as relevant contributors to ATH. In more socially conservative cultures (that embrace more traditional ideologies), the connection has been evidenced between conventional views about family structure and the perceived threat of external forces such as homosexuality (Brambilla & Butz, 2013; Poteat & Mereish, 2012). In addition, research shows that social conservatism promotes feelings of disgust and increased negative attitudes toward homosexuals and other out-groups (Terrizzi, Shook, & Ventis, 2010). In terms of civil rights, research indicates links between societal levels of anti-gay prejudice and not legally recognizing same-sex marriage in a country (van den Akker et al., 2012). Countries imposing more restrictive laws against homosexual behavior tend to foster less tolerant norms (Andersen & Fetner, 2008). Thus, the number of legal rights granted to homosexuals should be related to and how they are perceived (van den Akker et al., 2012).

The Current Study

The current study aimed to apply a pertinent set of individual- and country-level factors to concurrently explain the outcome of ATH. Kuntz et al. (2015) and van den Akker et al. (2012) showed the importance of individual- (e.g., gender, age, education, religiosity) and country-level variables (e.g., laws regarding homosexuality, cultural values) in understanding ATH. Expanding upon these previous investigations, the current study tests additional variables, such as the unique association of political ideology and perceptions of political fairness at the individual level, and socially conservative norms at the country level with the outcome of ATH. Furthermore, the current study extends past research by testing country-level social conservatism as a statistical moderator of the connection from individual-level predictors to ATH.

Hypotheses and Predictions

Analyses were conducted with MLM (hierarchical linear model), a technique that takes into account and controls for hierarchically nested data, and permits person- and country-level predictors to explicate variance in ATH. Due to social and cross-cultural variations (see Kuntz et al., 2015; van den Akker et al., 2012), ATH is hypothesized to vary based on country of residence (Hypothesis 1 [H1]), as evaluated by the test of random intercepts. As informed by the cited literature, the 11 person-level variables of female gender (Hypothesis 2 [H2]), younger age (Hypothesis 3 [H3]), higher education (Hypothesis 4 [H4]), being born in one’s current residing country (Hypothesis 5 [H5]), liberal political ideology (Hypothesis 6 [H6]), higher perceptions of political fairness (Hypothesis 7 [H7]), lower religiosity (Hypothesis 8 [H8]), higher openness to change values (Hypothesis 9 [H9]), lower conservation values (Hypothesis 10 [H10]), lower self-enhancement values (Hypothesis 11 [H11]), and greater self-transcendence values (Hypothesis 12 [H12]) are hypothesized to be associated with favorable ATH. Furthermore, after controlling for individual-level effects, countries scoring lower in social conservatism (Hypothesis 13 [H13]) and countries granting more civil rights to gays and lesbians (Hypothesis 14 [H14]) were hypothesized to foster stronger correlations with more favorable ATH.

Exploratory analyses tested potential interaction effects involving national climate (country-level social conservatism and gay civil rights) and individual-level predictors on ATH. Although previous research has not examined these particular combinations of interaction effects (while controlling for the aforementioned set of person- and control-level predictors), the testing of these moderating effects may be guided and informed by a theoretical framework. Specifically, Inglehart’s (1987, 1990, 1997) post-materialist thesis posits that cultural-level liberal principles result from democracy, innovation, and economic development, allowing individuals in these societies to shift their attention from material concerns to post-materialist concerns, such as social issues and self-expression. In support of this theory, research shows that the level of industrialization within a nation may shape tolerance for non-conventional groups and opinions (Adamczyk & Pitt, 2009; Inglehart, Norris, & Welzel, 2002). Inglehart and Baker (2000) explain that as nations become more modernized, attitudes and standards shift from concerns about physical and economic security (e.g., food, shelter, and safety) to views that are progressively reasonable, accepting, and trusting, with an increasing interest on issues such as quality of life and self-expression. Along with a shift from survival to self-expressive orientation, people are more easily able to accept new ideas and counter-normative behaviors such as homosexuality. In contrast, when nations are faced with economic or political uncertainty, people are more likely to support ideas that are familiar, resulting in less tolerance of non-traditional ideologies.

Based on this framework, it was hypothesized that stronger associations of individual-level predictors with ATH would be exhibited in more progressive societies (conceptualized as countries that are lower on social conservatism and grant more legal rights for homosexuals): People residing in these nations have the ability to focus on self-expression and challenging traditional views.

Method

Dataset

Data were drawn from the sixth round of the biennial European Social Survey (ESS; 2012). Responses were collected from 28 European countries (see Table 2),² using a cluster sampling design with an equal probability method of selecting participant units (Crano, Brewer, & Lac, 2015). The ESS is conducted through in-person interviews with residents aged 15 years and older. Design weights and population weights, provided by the dataset, were applied to ensure the representativeness and generalizability of the results. The final weight value was computed by multiplying the design and population weights. Additional information about the sampling design and procedures can be found on the ESS website (ESS, 2012).

The individual-level variables, including the outcome of ATH, were drawn from the dataset. The group-level measures were culled from various external data sources to supplement the dataset. For all group-level measures, individuals residing within the same country were assigned the same value for the purpose of analysis (Heck, Thomas, & Tabata, 2013).

Participants

The sample consisted of 36,959 individuals drawn from 28 European countries (see Table 1 for the list of countries). Respondents ranged from 15 to 103 years old (M = 46.81, SD = 18.30). Females comprised 54.8% of the sample. The vast majority (91.5%) of participants were born in their residing country.

Table 1.

Mean Scores for all Variables by Country of Residence.

	Country-Level Variables		Individual-Level Variables
	Social Conservatism	Civil Rights	Attitudes Toward Homosexual People		Age		Education		Political Affiliation		Fairness		Religiosity		Openness to Change Values		Conservation Values		Self-Enhancement Values		Self-Transcendence Values
Country			M	SD	M	SD	M	SD	M	SD	M	SD	M	SD	M	SD	M	SD	M	SD	M	SD
Albania	5	6	2.54	1.35	41.96	18.25	2.93	1.57	5.02	3.54	4.44	2.23	7.18	2.98	4.33	.79	4.87	.69	4.27	.91	5.09	.61
Belgium	4	17	4.26	1.00	47.33	19.09	3.65	1.97	4.96	1.98	5.89	1.64	4.49	3.16	4.26	.76	4.41	.70	3.66	.89	4.97	.55
Bulgaria	4	6	3.50	1.18	52.48	16.74	3.49	1.86	4.81	2.57	3.68	1.85	4.62	2.54	3.84	1.14	4.83	.76	4.20	.96	4.95	.75
Cyprus	4	1	3.51	1.20	45.94	17.99	3.44	1.92	4.84	3.04	5.53	1.98	7.08	2.36	4.33	.95	4.87	.64	3.95	1.10	5.22	.61
Czech Republic	3	9	3.60	1.15	45.49	17.06	3.48	1.42	4.71	2.53	4.52	1.77	2.01	2.66	4.00	.93	4.41	.82	3.86	1.02	4.50	.78
Denmark	1	9	4.46	.83	48.70	19.03	3.93	1.91	4.66	2.37	7.43	1.35	4.20	2.69	4.33	.82	4.14	.80	3.54	.95	4.99	.62
Estonia	3	2	3.17	1.21	49.41	19.59	3.91	1.83	4.55	2.07	4.72	1.95	3.48	2.96	3.85	.89	4.30	.72	3.35	1.00	4.75	.65
Finland	1	11	4.01	1.07	49.75	18.90	3.68	1.96	4.34	2.04	6.96	1.47	4.97	2.73	4.10	.81	4.23	.80	3.11	.97	4.99	.63
France	1	6	4.15	1.16	49.21	18.17	3.35	1.81	5.02	2.53	5.53	1.61	4.44	3.37	4.03	.84	4.12	.84	3.05	.96	4.87	.73
Germany	2	20	4.15	.91	47.87	18.55	3.90	1.71	5.37	1.88	6.19	1.58	4.52	3.01	4.20	.78	4.25	.79	3.56	.92	5.06	.54
Hungary	3	14	3.16	1.29	47.20	18.18	3.36	1.50	4.59	2.11	4.86	2.30	4.05	3.00	4.34	.85	4.61	.73	4.18	.94	4.94	.70
Iceland	3	12	4.56	.72	44.14	19.10	3.77	1.86	4.66	2.10	5.94	1.82	5.27	3.02	4.19	.83	3.89	.82	3.17	.99	4.97	.66
Ireland	1	9	4.22	.84	44.82	17.66	3.54	1.82	4.85	1.75	6.28	2.01	5.01	2.64	4.11	.86	4.44	.83	3.85	.92	4.92	.71
Italy	3	2.5	3.91	1.11	47.08	18.27	3.25	1.91	5.33	2.68	4.13	1.84	5.91	2.92	4.11	.88	4.79	.77	4.00	.96	5.16	.64
Kosvo	5	2	2.29	1.49	41.90	17.34	2.55	1.38	2.96	2.94	3.95	2.22	7.34	2.32	4.44	.93	5.23	.63	4.80	.89	5.35	.58
Lithuania	4	3	2.36	1.13	45.03	18.26	3.79	1.79	5.08	2.18	4.10	1.89	5.41	2.57	3.74	1.05	4.12	.76	4.02	1.00	4.13	.80
Netherlands	1	16	4.46	.76	48.83	17.85	3.54	2.00	4.64	2.10	6.59	1.35	4.54	3.18	4.30	.75	4.17	.76	3.57	.89	4.84	.61
Norway	1	15	4.18	.94	46.00	18.18	4.10	1.85	4.36	2.11	7.40	1.37	3.81	2.85	4.03	.81	4.15	.79	3.38	.87	4.79	.64
Poland	4	2	3.30	1.20	46.10	18.87	3.52	1.72	4.33	2.39	4.80	1.76	6.25	2.59	3.87	.92	4.79	.70	3.95	.91	4.98	.60
Portugal	4	15	3.84	1.01	49.96	18.96	2.10	1.70	5.09	2.10	4.29	1.63	5.22	2.75	3.86	.79	4.30	.72	3.79	.80	4.57	.69
Russian Federation	5	-4.5	2.48	1.24	43.64	17.34	4.75	1.79	4.65	1.99	3.75	2.10	4.48	2.71	3.98	.95	4.56	.81	4.23	.99	4.69	.80
Slovakia	4	9	3.08	1.34	46.73	16.07	3.84	1.75	4.74	2.52	4.57	1.80	5.97	3.08	4.08	.99	4.82	.72	4.17	.98	4.80	.65
Slovenia	3	9	3.35	1.25	48.31	18.89	3.39	1.46	5.29	2.35	4.66	1.80	4.50	3.16	4.39	.83	4.86	.63	4.12	.83	5.14	.53
Spain	3	20	4.17	.94	47.64	17.99	2.79	2.23	5.48	2.40	4.50	1.88	4.48	3.05	4.08	.92	4.65	.77	3.35	1.00	5.21	.58
Sweden	1	18	4.35	.80	47.83	19.02	3.76	1.85	4.69	2.23	7.35	1.77	3.17	2.90	4.21	.83	4.04	.89	3.33	.99	4.98	.67
Switzerland	3	6.5	3.99	1.00	47.41	18.79	3.82	1.74	4.85	1.96	6.87	1.46	4.99	2.98	4.45	.74	4.37	.77	3.81	.94	5.13	.54
UK	2	21	4.12	.85	49.07	19.02	3.39	1.99	4.94	1.83	6.26	1.90	4.13	2.98	4.09	.86	4.44	.81	3.66	.98	4.96	.64
Ukraine	5	-4	2.40	1.40	45.84	18.20	4.50	1.80	4.55	2.35	3.29	1.91	5.03	2.66	3.67	1.02	4.34	.87	3.90	1.07	4.50	.92
Total	3.15	7.68	3.59	1.32	46.84	18.29	3.79	1.96	4.94	2.25	4.93	2.17	4.72	3.00	4.05	.90	4.46	.83	3.77	1.04	4.89	.73

Individual-Level Measures³

Gender and age

Gender was dichotomously coded (0 = male; 1 = female). Age was measured as an open-ended quantitative variable.

Education level

Educational attainment was measured with the question, “What is the highest level of education you have successfully completed?” Response options ranged from 0 (no education or very low education) to 8 (highly educated: college or graduate degree).

Born in residing country

Respondents were asked if they were born in their residing country. Responses were coded 0 (not born in residing country) and 1 (born in residing country).

Political affiliation

Political affiliation was operationalized with the item, “In politics people sometimes talk of left and right. Using this card, where would you place yourself on this scale, where 0 means the left and 10 means the right?” Responses were reverse coded to yield a range from 0 (right) to 10 (left). A similar measure of political affiliation was used in a previous study conducted with people from European countries (Anderson & Singer, 2008).

Perceptions of political fairness

A six-item measure of political fairness perceptions (α = .85) was informed by Kluegel and Mason’s (2004) scale of political and fairness evaluations. Respondents were asked, “To what extent do you think each of the following statements applies in [participant’s country of residence]?” They were presented with the stem “In [country] . . .” followed by each item: (a) “Opposition parties are free to criticize the government,” (b) “The media are free to criticize the government,” (c) “Courts treat everyone the same,” (d) “The government protects all citizens against poverty,” (e) “The government explains its decisions to voters,” and (f) “The government takes measures to reduce differences in income levels.” Responses ranged from 0 (does not apply at all) to 10 (applies completely). A mean score was computed to represent the construct, with higher values representing higher fairness perceptions.

Religiosity

Religiosity was measured with an item used in previous research (Zmerli & Newton, 2008). Respondents were asked, “Regardless of whether you belong to a particular religion, how religious would you say you are?” Responses were assessed on a Likert-type scale ranging from 0 (not religious) to 10 (very religious).

Human values

The Human Values Scale (Schwartz, 2003) was used to assess personal values. This 21-item scale involves a two-sentence vignette describing a fictitious person who is gender matched to the respondent. After reading the scenario, respondents indicated how similar or dissimilar this person was to himself or herself using response options ranging from 1 (very much like me) to 6 (not like me at all). Respondents’ values were deduced based on their similarity ratings with the person described in the vignette. Items were scored such that higher numbers denoted higher similarity of values. Four mean subscales were computed (Davidov et al., 2008; Kuntz et al., 2015): openness to change (OPEN; six items; α = .77), conservation (CONS; six items; α = .74), self-enhancement (SE; four items; α = .75), and self-transcendence (ST; five items; α = .75). Higher values indicated greater perceptions of openness to change, conservation, self-enhancement, and self-transcendence (for a full list of items, see Davidov et al., 2008).

Attitudes toward homosexual people

Attitudes toward homosexual people were captured with a measure applied in previous research (Takács & Szalma, 2011; van den Akker et al., 2012). Participants indicated the extent of agreement with the statement: “Gays and Lesbians should be free to live their own life as they wish.” Likert-type response options were 1 (disagree strongly), 2 (disagree), 3 (uncertain), 4 (agree), and 5 (agree strongly).

Group-Level Measures

Social conservatism

Guided by the conceptual framework of the Traditional Family Values Scale (Newman & Muzzonigro, 1993), external data about religion, marriage, children, and conventional family roles represented the extent to which countries were marked by socially conservative norms. Religious unimportance in a country was obtained from the Global Index of Religiosity and Atheism (see Gilani, Shahid, & Zuettel, 2012), a measure of the percentage of people in a country that did not view themselves as religious. The average age of first marriage (United Nations Economic Commissions for Europe, 2012), average age of first child (Central Intelligence Agency, 2012), and percentage of the workforce that was female for each country (The World Bank, 2012) were also recorded. All four items (α = .81) were standardized (Z scored), and then averaged. Next, the mean of the standardized scores was recoded into five groupings of equal proportions to create a 5-point scale. This composite was then reverse coded to yield a scale from 1 (low social conservatism) to 5 (high social conservatism).

Gay civil rights

Gay civil rights were measured using the Rainbow Europe Country Index (International Lesbian, Gay, Bisexual, Trans, and Intersex Association Europe [ILGA Europe], 2012). This index provides an overall reflection of gay and lesbian legal rights in each European country. The organization responsible for this index assigned different points to each facet. Positive points were assigned for countries with marriage equality, adoption rights, and non-discrimination laws. Negative points were assigned to countries for violating freedoms of assembly for homosexuals, limiting freedoms of association or expression for homosexuals, making homosexual activity illegal, and having discrimination laws against homosexual contact, to name some examples. The final index of gay civil rights ranges from −4 (least civil rights) to 21 (most civil rights).

Analytic Plan

MLM, using the SPSS mixed procedure (Heck et al., 2013; Peugh & Enders, 2005), was performed with 36,959 individuals (Level 1) across 28 countries (Level 2) to control for the nested data structure. All group- and individual-level predictors, and the outcome variable, were standardized (Z scored) prior to conducting analyses (Heck et al., 2013). This transformation is appropriate for fixed-effect models and enables the interpretation of standardized regression coefficients, which represent effect sizes that have statistically removed the contribution of rival predictors in the same model. The statistical assumptions for MLM were satisfied—the quantitative variables did not depart drastically from normality (Hair, Anderson, Tatham, & Black, 1995). Moreover, multicollinearity was not an issue as variance inflation factor (VIF) values were below 4.00 (Hair et al., 1995); the highest VIF in Model 2 was 1.75, in Model 3 it was 2.21, in Model 4a it was 2.34, and in Model 4b it was 2.47.

The multilevel analyses were sequentially specified, by incorporating additional predictors into each successive model, to produce nested models that could be statistically compared. Models were fitted using maximum-likelihood (ML) estimation. Three indices were interpreted to scrutinize the fit of the nested models. The −2 log likelihood (−2LL) indicates the extent to which the model coincides with the underlying data (Heck et al., 2013; Peugh & Enders, 2005)—a lower −2LL signifies a better fitting model. Whether a subsequent model rendered a significant improvement over the prior model was tested using the change in chi-square statistic (χ²), computed from the difference in the −2LL of two nested models. Akaike’s information criteria (AIC) were also inspected to judge model fit, with a lower value indicating better fit (Heck et al., 2013; Wright & London, 2009).

Participant country of residence served as the grouping variable in all analyses. Model 1 was specified as a null (baseline) model, by permitting random intercepts only, to determine whether mean scores in ATH were significantly discrepant across the 28 countries. This model was used to compute the intraclass correlation (ICC), an indication of the extent that people within the same countries were similar on their ATH scores relative to the total variation in individual attitudes across all countries. A high ICC value beyond the null hypothesis of .00 signifies that participant units are not statistically independent within countries, and therefore the nested design should be taken into account by conducting a multilevel model.

Models 2 and 3 involved random intercepts with fixed-effect predictors. Model 2 builds upon the previous model by including 11 fixed-effect predictors at the individual level. Model 3 subsequently incorporated two fixed-effect predictors at the country level. Thus, Models 2 and 3 controlled for the nested structure by permitting intercepts to vary, while estimating fixed effects. Fit statistics from Models 1, 2, and 3 were compared to determine whether adding each subsequent block of predictors enhanced model fit.

Building upon Model 3, Models 4a and 4b incorporated cross-level interactions of individual- and group-level variables. Nation-level contextual characteristics (social conservatism and gay civil rights) were specified as moderators of the relation between all individual variables and ATH. Two separate models were conducted for each country-level moderator to reduce the likelihood of multicollinearity. Significant interactions were decomposed with tests of simple slopes, and were specified and interpreted following recommended guidelines (Bauer & Curran, 2005; Curran, Bauer, & Willoughby, 2006; Preacher, Curran, & Bauer, 2006). Each predictor variable was assessed at 1 standard deviation below the mean, at the mean, and 1 standard deviation above the mean for the moderator variables.

Results

Raw (unstandardized) descriptive information for all continuous variables is presented in Table 1. The correlation matrix of all standardized predictor variables is presented in Table 2.

Table 2.

Correlation Matrix for All Standardized Predictor Variables.

	1	2	3	4	5	6	7	8	9	10	11	12	13
1. Gender (L1)	—
2. Age (L1)	.03	—
3. Education (L1)	.01	−.13	—
4. Born in residing country (L1)	−.01	.01	−.01	—
5. Political affiliation (L1)	−.01	−.02	.03	−.03	—
6. Fairness (L1)	−.06	.06	−.04	−.09	−.07	—
7. Religiosity (L1)	.15	.15	−.08	−.09	−.16	.03	—
8. Openness to change values (L1)	−.09	−.30	.08	.00	−.01	.09	−.06	—
9. Conservation values (L1)	.07	.19	−.11	−.04	−.11	−.05	.26	.08	—
10. Self-enhancement values (L1)	−.08	−.27	.14	−.01	−.09	−.09	.02	.48	.28	—
11. Self-transcendence values (L1)	.06	.07	−.03	−.05	.05	.09	.12	.32	.52	.16	—
12. Social conservatism (L2)	.06	−.09	.20	.08	−.07	−.45	.06	−.10	.16	.30	−.16	—
13. Civil rights (L2)	−.06	.08	−.21	−.09	.09	.42	−.08	.11	−.09	−.23	.18	−.71	—

Note. The reference group for gender is male. L1 = individual level; L2 = country level. Correlations of at least .01 are significant at p < .05. Correlations of at least .03 are significant at p < .001.

MLM

Null model

H1 proposed that characteristics attributed to country of residence would explain variation in ATH, and was tested using Model 1. Analysis revealed a significant ICC of .27, Wald Z = 2.66, p < .01, signifying that 27% of the variance in one’s ATH was explained exclusively by variations across countries. This provided sufficient evidence that a multilevel regression model was warranted (Heck et al., 2013; Peugh & Enders, 2005). Given significant mean (intercept) differences across countries on the outcome measure, Table 1 details that people from Iceland, Denmark, and The Netherlands were most accepting, whereas individuals residing in Kosovo, Lithuania, Ukraine, and Russia were most disapproving of homosexuals.

Random intercepts with person-level predictors

Model 2 tested the hypothesized 11 fixed-effect predictors at the individual level (Level 1). As detailed in Table 3, gender, education level, birth in resident country, political affiliation, fairness perceptions, openness to change values, and self-transcendence values were positively associated with favorable ATH, whereas age, religiosity, conservation values, and self-enhancement values were negatively linked with ATH. In particular, female gender, younger age, higher education level, being born in one’s residing country, liberal political affiliation, higher perceptions of political fairness, lower religiosity, higher openness to change values, lower conservation values, lower self-enhancement values, and higher self-transcendence values were uniquely and significantly associated with positive ATH. These results supported H2 through H12. As reported in the chi-square difference test in Table 4, Model 2 was significantly (p < .001) better than Model 1 (null model).

Table 3.

Multilevel Model Predicting Attitudes Toward Homosexual People.

	Model 1	Model 2	Model 3	Model 4a	Model 4b
Individual-level variables (L1)
Gender		.08***	.08***	.08***	.08***
Age		−.09***	−.09***	−.10***	−.10***
Education		.08***	.08***	.08***	.08***
Born in residing country		.04***	.04***	.02***	.03***
Political affiliation		.08***	.08***	.08***	.08***
Fairness		.06***	.06***	.06***	.06***
Religiosity		−.12***	−.12***	−.11***	−.12***
OPEN		.05***	.05***	.04***	.05***
CONS		−.10***	−.10***	−.09***	−.09***
SE		−.05***	−.05***	−.03***	−.04***
ST		.12***	.12***	.12***	.12***
Country-level variables (L2)
SC (L2)			−.30**	−.30**	−.31**
CR (L2)			.22*	.22*	.21*
Cross-level interactions
Gender × SC				.00
Age × SC				−.02***
Education Level × SC				−.01
Born in Residing Country × SC				−.04***
Political Affiliation × SC				−.01***
Fairness × SC				.02***
Religiosity × SC				.02***
OPEN × SC				.02***
CONS × SC				−.01*
SE × SC				.03***
ST × SC				−.04***
Gender × CR					.00
Age × CR					.00
Education Level × CR					−.01*
Born in Residing Country × CR					.02***
Political Affiliation × CR					.01
Fairness × CR					−.03***
Religiosity × CR					−.01*
OPEN × CR					−.02**
CONS × CR					.00
SE × CR					.00
ST × CR					.04***
Covariance parameters
Residual variance	.58***	.58***	.58***	.58***	.58***
Intercept variance	.05**	.05**	.05*	.05*	.05*

Note. OPEN = openness to change values; CONS = conservation values; SE = self-enhancement values; ST = self-transcendence values; SC = social conservatism; CR = civil rights.

p < .05. **p < .01. ***p < .001.

Table 4.

Model Fit Indices.

Model	Type	Description	df difference	−2 log likelihood	AIC	χ² difference
1	Null model	Individuals nested within their country with no other predictors	—	111,943.06	111,949.06	—
2	Random intercepts and fixed predictors	Individual (L1) level	11	84,918.67	84,946.67	27,024.39***
3	Random intercepts and fixed predictors	Individual (L1) and country (L2) level	2	84,898.77	84,930.77	19.90***
4a	Random intercepts and fixed interactions	Cross-level interactions with social conservatism (L2) and individual-level predictors (L1)	11	84,532.51	84,586.51	366.26***
4b	Random intercepts and fixed interactions	Cross-level interactions with civil rights (L2) and individual-level predictors (L1)	11	84,707.85	84,761.85	190.92***

Note. Models are compared with the previous models for comparison of model fit. AIC = Akaike’s information criteria.

p < .05. **p < .01. ***p < .001.

Random intercepts with country- and person-level predictors

All individual- and group-level (i.e., social conservatism and gay civil rights of countries) predictors were simultaneously tested in Model 3. As reported in Table 4, Model 3 produced a significantly superior model fit compared with Model 2 (p < .001). Both group-level predictors emerged as significant in the model. Lower social conservatism and granting more gay civil rights were linked with more favorable ATH scores, supporting H13 and H14.

Random intercepts with cross-level interactions of country-level social conservatism and person-level predictors

In Model 4a, cross-level interactions of social conservatism with individual-level variables were entered into the model. Compared with Model 3, Model 4a produced a superior fit (p < .001; Table 4). Results revealed statistical interactions of social conservatism with age, birth country, political affiliation, fairness, religiosity, openness to change values, conservation values, self-enhancement values, and self-transcendence values (Table 4).

The slopes of significant interaction terms are decomposed in Table 5. Conservative culture as the moderator was examined first. In more conservative cultures, the negative connection between age and ATH, and the positive association between fairness perceptions and ATH, was stronger than in less traditional societies. Similarly, the connection from strong openness to change values to supportive ATH, and from high conservation values to unsupportive ATH, exhibited stronger magnitudes in more socially conservative societies. Thus, younger age, greater fairness perceptions, placing a high emphasis on independence, freedom, and openness (high openness to change values), and placing a low importance on conformity to tradition (low conservation values) were associated with a stronger acceptance of homosexuality in more socially conservative countries, holding more traditional ideologies at the group level.

Table 5.

Simple Slopes of the Moderating Effect of National-Level Social Conservatism and Gay Civil Rights With Individual-Level Predictors on Attitudes Toward Homosexual People.

	Low social conservatism (−1 SD)	Moderate social conservatism (0 SD)	High social conservatism (+1 SD)
Age	−.08***	−.10***	−.12***
Born in residing country	.06***	.02***	−.02**
Political affiliation	.09***	.08***	.07***
Fairness	.04***	.06***	.08***
Religiosity	−.13***	−.11***	−.09***
Openness to change values	.02**	.04***	.06***
Conservation values	−.08***	−.09***	−.10***
Self-enhancement values	−.06***	−.03***	.00
Self-transcendence values	.16***	.12***	.08***
	Low civil rights (−1 SD)	Moderate civil rights (0 SD)	High civil rights (+1 SD)
Education level	.09***	.08***	.07***
Born in residing country	.01	.03***	.05***
Fairness	.09***	.06***	.03**
Religiosity	−.11***	−.12***	−.13***
Openness to change values	.07***	.05***	.03**
Self-transcendence values	.08***	.12***	.16***

p < .05. **p < .01. ***p < .001.

The positive association between birth country and ATH was greater in less conservative cultures, but this relation was instead negative in highly conservative societies (Table 5). That is, immigrants were less accepting of homosexuality in more progressive countries but were more accepting in countries scoring the highest in terms of social conservatism.

Furthermore, in more progressive countries, which scored low in social conservatism, the relation from both liberal political affiliation and low religiosity, to favorable ATH, was stronger than in countries that were more socially conservative (Table 5). In addition, the positive relation between self-transcendence values and ATH, and the negative link between self-enhancement values and ATH were stronger in cultures scoring low in social conservatism. Thus, liberal political affiliation, placing a low emphasis on religiosity, assigning a low importance to power and self-achievement (i.e., low self-enhancement values), and placing a strong emphasis on social justice concerns for others (i.e., high self-transcendence values) were most strongly associated with more accepting ATH in countries scoring lowest in social conservatism.

Random intercepts with cross-level interactions between country-level gay rights and individual-level predictors

Gay civil rights as a moderator were examined next in tests of simple slopes (Table 5). Model 4b, cross-level interactions of granted gay rights with all individual-level predictors were tested. Compared with Model 3, Model 4b produced a superior fit (p < .001; see Table 4). Analyses demonstrated significant interactions between civil rights and education level, immigration status, fairness perceptions, religiosity, openness to change values, and self-transcendence values on ATH.

In cultures granting homosexual people fewer civil rights, the positive relationships between both education level and fairness perceptions with accepting ATH were stronger than in those permitting greater civil liberties to the homosexual community (Table 5). Furthermore, greater openness to change values were more strongly associated with favorable ATH in countries granting fewer civil liberties to sexual minorities. As such, being more educated, perceiving one’s government as fair, and valuing openness to change were more strongly connected with accepting ATH in cultures granting homosexuals the fewest civil rights.

Compared with immigrants, non-immigrants born in their residing country were more accepting of homosexuals in countries granting more gay rights. In addition, high religiosity and lower self-transcendence values were associated with less accepting ATH in countries where homosexuals were granted more civil liberties. Thus, being born in one’s residing country, being less religious, and placing a high emphasis on the social welfare of others (i.e., high self-transcendence values) were more strongly linked with favorable ATH in countries granting more civil liberties to sexual minorities.

Discussion

The aim of this research was to elucidate factors associated with the extent of approval (or disapproval) of homosexuality, at both the individual- and national levels, using data collected from 28 European countries. The current study extends past research examining ATH by utilizing innovative multilevel analyses, testing novel variables not previously examined, and applying high-quality international data for the purpose of estimating the contribution of individual- and country-level factors across European countries.

Main Effects

Consistent with previous research (e.g., Gerhards, 2010; Hodson & Busseri, 2012; Poteat & Mereish, 2012; Röder, 2015), results supported that the individual-level measures of female gender, younger age, higher education, being born in one’s current residing country, liberal political affiliation, low religiosity, and perceptions that politics in one’s country are fair contributed to more positive ATH. In addition, our findings corroborated previous studies, showing that personal values are linked with ATH (Heaven & Oxman, 1999; Kuntz et al., 2015)—openness to change and self-transcendence values were associated with positive ATH, whereas conservation values and self-enhancement values were linked with negative ATH.

It is not surprising that gender played a key role in the likelihood of adopting pro-homosexuality attitudes. This association can be explained using feminist theorizing, contending that, as marriage is historically a patriarchal institution, women are more receptive to supporting homosexual liberties and marriage rights afforded to straight people (Ferguson, 2007). In terms of age, older cohorts have been socialized to perceive homosexuality as a disease and a sin; thus, they possess more unfavorable attitudes (Van de Meerendonck & Scheepers, 2004).

This study ascertained unique associations of education, religiosity, and political ideology on attitudes as supported in the previous literature (e.g., Poteat & Mereish, 2012; Wojcieszak & Price, 2010). Greater educational attainment has been linked with more liberal attitudes and greater tolerance for counter-traditional lifestyles (van den Akker et al., 2012). Religions often impose strict norms and codes of conduct regarding appropriate behaviors for their followers (Andersen & Fetner, 2008; Van de Meerendonck & Scheepers, 2004): Negative views toward homosexuality are embedded in many religions. As such, religious people tend to be less tolerant of homosexuality. Supporting research reveals a strong correlation between conservative political ideologies and right-wing authoritarianism (Van Hiel, Duriez, & Kossowska, 2006), with both constructs connected too less favorable ATH (Steffens & Wagner, 2004).

Research also contends that culturally instilled behaviors and beliefs regarding sexuality are indispensible aspects of the collective consciousness of countries (Bourdieu, 2001). Thus, whether an individual is native to their country of residence, or immigrated and socialized elsewhere, has an impact on his or her values and attitudes. Results of the current investigation showed that immigrants were less approving of homosexuality. In line with these results, research (Röder, 2015) shows that first-generation immigrants are less accepting of homosexuality because they typically migrate from countries with less supportive ATH.

Beyond the effects, a review of the literature reveals that this is likely the first examination of political fairness perceptions in the context of ATH. Findings shed light on psychological processes that underlie fairness judgments and the extent that homosexuality is accepted. Consistent with the study hypothesis, perceiving one’s own government, laws, and politics as fair was linked with more pro-homosexual attitudes. This finding is also supported by research (Van den Bos & Lind, 2002) showing that perceptions of being treated fairly by police and other authorities boost personal acceptance of public laws and opinion (Tyler & Degoey, 1996; Tyler & Wakslak, 2004). Thus, fairness judgments might motivate and guide attitude formation and individual reactions to important issues, such as homosexual rights and acceptance.

The relations among measures of individual human values and approval of homosexuality provide additional support for prior research, indicating that individual-level priorities can have powerful influences on moral attitudinal development (e.g., Schwartz et al., 2010; Kuntz et al., 2015). People view homosexuality as a threat to the traditional family (Haddock & Zanna, 1998), so the acceptance of homosexuality as a moral behavior involves abandoning traditional views of sexual and gender roles in support of changing customs.

Openness to change values (emphasizing personal freedom and autonomy) and self-transcendence values (valuing social justice and tolerance for others) both fostered approving ATH. Thus, accepting alternative lifestyles and valuing tolerance and caring for others contribute to individual approval of counter-traditional behaviors, such as homosexuality. In contrast, conservation values (a concern for upholding tradition, and conforming to rules and regulations) and self-enhancement values (emphasizing power, social status, and achieving personal goals) were associated with less approving ATH. As such, individuals who prioritize adherence to traditional customs and ideologies, and those who are self-interested and concerned with their own personal status are less likely to accept homosexuality.

As hypothesized, low national-level social conservatism was associated with more approving ATH (Terrizzi et al., 2010). This finding might be explained by observing trends in the redefinition of marriage, as research shows that people who perceive gay marriage as a threat to the sanctity of heterosexual marriage are usually less approving of homosexuality (Brumbaugh, Sanchez, Nock, & Wright, 2008). In line with this proposition, people’s perception that homosexuals are responsible for breakdown of the traditional family structure, the view that marriage should remain between a man and a woman, and the disapproval of divorce, are prevailing forces impinging on negative views about homosexuals. Thus, it is possible that cultures bound by more socially conservative customs might be more willing to fault homosexuals for shifts in traditional views of marriage, promoting policies that strengthen the exclusivity of heterosexual marriage (Brumbaugh et al., 2008).

A significant link between progressive governemental policies supporting homosexuality (i.e., granting homosexuals more civil liberties) and the acceptance of homosexuality also emerged (van den Akker et al., 2012). This finding might potentially show the bidirectional association between legal institutions and individual attitudes, as supported by Coleman’s (1990) boat hypothesis. Progressive policies may promote tolerant ATH, which in turn encourage action that influences the formation of progressive laws. Despite this idea, many changes in European legal regulations have occurred due to governmental pressures from the European Union (EU) rather than individual-level attitude change (Schlüter, Meuleman, & Davidov, 2013). Thus, future research should examine the feedback loop between individual attitudes and legal regulations.

Interaction Effects

Results of exploratory estimation of interaction terms showed that cross-cultural variations in social conservatism and gay civil rights statistically moderated the strength of the connections between individual political, religious, and human values on ATH. In line with the post-materialist thesis (Inglehart, 1987, 1990, 1997), and research showing stronger variance in individual attitudes for people residing in self-expressive rather than survival-oriented cultures (Adamczyk & Pitt, 2009; Inglehart, Norris, & Welzel, 2002), it was predicted that individual-level factors would have the strongest associations in more progressive societies (i.e., low in social conservative, granting homosexual people more civil liberties).

In support of the study hypotheses, having a non-immigrant status, liberal political affiliation, low religious perceptions, placing a low emphasis on power and social status (i.e., low self-enhancement values), and valuing social equality (high self-transcendence values) were linked with more favorable ATH in countries with lower levels of social conservatism. Similarly, being born in one’s residing country, possessing lower levels of religiosity, and valuing the social justice and the welfare of others (low self-transcendence values) were most strongly linked with positive ATH societies granting sexual minorities more civil rights. Contrary to study predictions, younger age, perceiving a country’s government as fair, valuing freedom, autonomy, and change (high openness to change values), and placing a low emphasis on concern for conformity and tradition (low conservation values) were most strongly linked with positive ATH in the most socially conservative cultures. Furthermore, higher education level, stronger fairness perceptions, and stronger openness to change values were more strongly connected with accepting attitudes in countries granting homosexuals fewer civil liberties (less progressive nations).

The hypothesized association between immigrant status and attitudes in more tolerant societies can be explained using the persistence model (Davies, 1965), which argues that cultural customs are instilled as early as childhood and are reinforced over time, leading to less malleable, persistent values. Furthermore, previous research has demonstrated that immigrants typically hold less understanding perceptions of homosexuals, as they usually migrate from less industrialized countries that are typically less accepting of homosexuals (e.g., Goldenberg & Saxe, 1996; Röder, 2015; Wong & Poon, 2013). Findings from cross-level interactions support this proposition—respondent ATH were consistent with their country’s ideologies regarding homosexuality. In progressive countries that ranked low in social conservatism, natives to the country had significantly more positive ATH compared with immigrants. Likewise, in countries granting homosexuals more civil liberties, non-immigrants had significantly greater positive ATH compared with immigrants.

In support of the post-materialist thesis, high religiosity and conservative political beliefs were most strongly connected with unaccepting views of homosexuality in the most progressive cultures. This relation is reinforced by research showing that individual ideologies are more strongly linked with ATH in countries that emphasize self-expression rather than survival (Adamczyk & Pitt, 2009). Thus, it is likely that countries with more country-level progressive beliefs are likely to encourage thinking divergent from traditional ideologies.

In contrast to the original hypotheses, younger age, higher education level, and perceiving one’s government as fair had stronger associations with accepting ATH in less progressive cultures. These contradictory results might show that people holding conflicting views from the societal-level status quo feel a sense of social instability or uncertainty due to the differing values of one’s country and oneself, resulting in the adoption of more extreme values in the direction away from the rival view (Hogg, 2015). As a result, it is likely that group polarization occurs (Wojcieszak & Price, 2010), a process in which individuals react by forming more extreme views, contrasting societal opinions, on political issues such as the legalization of gay marriage and equality for the homosexual community.

Extending research presented by Kuntz and associates (2015), the current study examined country-level social conservatism as a group-level variable moderator, in addition to legally sanctioned civil rights. Contrary to Inglehart’s (1987, 1990, 1997) theorizing, Kuntz and colleagues (2015) found that individual-level values of openness to change and conservation were most strongly associated with attitudes in cultures granting fewer civil rights. In corroboration with these results, our findings also revealed that the link between one’s personal openness to change values and conservation values with ATH was strongest in the least progressive societies. In contrast, self-transcendence values and self-enhancement values were most strongly associated with ATH in more liberal cultures, supporting Inglehart’s (1987, 1990, 1997) post-materialist thesis. These opposing results may provide further support to a theory of group polarization, as self-focused personal values (i.e., self-enhancement, openness to change) had stronger respective connections to ATH in countries holding opposing social and legal standards, whereas other focused personal values (i.e., self-transcendence, conservatism) had the strongest relations in countries with national views aligned with their own. Future research should further examine the mechanisms underlying a potential group polarization in this context.

Limitations and Future Directions

Results should be interpreted in light of several limitations. Data were cross-sectional and correlational, so causal claims about the findings should be avoided. Future research should assess predictors of ATH using longitudinal designs to unravel the temporal direction of these processes. The self-report measures of individual-level variables may also present some limitations. However, validity concerns of self-report bias tend to be minor and unlikely to alter study results (Cornelius, Leech, & Goldschmidt, 2004). The one-item measure of ATH may also be a limitation, but this measure has been used extensively in published research (Takács & Szalma, 2011; van den Akker et al., 2012). In further support of including a one-item outcome measure, research shows that people’s ATH reflect their global approval/disapproval of homosexual behavior and the lesbian, gay, bisexual, and transgender (LGBT) community at large (Herek, 2000). Thus, the one-item measure employed in this study is likely to serve as a valid proxy to individual acceptance of the homosexual community. Regardless, future research should investigate the relation between multi-item measures of ATH and support of the legalization of gay marriage, which is relevant in current political debate. Furthermore, large datasets often present overabundant statistical power for inferential analyses; however, the current research also reported standardized coefficient estimates to determine the magnitude of each predictor.

Conclusion

Based on an internationally representative sample of Europeans, the current study offers unique cross-cultural insights on ATH. Taken together, this study reveals a multitude of demographic, social, political, and legal factors that are simultaneously associated with attitudes regarding the acceptance of homosexual people, as well as offering a glimpse into the political climates of countries that garner progressive movement toward equality. These results contribute to the literature for unraveling cross-cultural and individual factors effecting one’s position on homosexuality across countries.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Supplemental Material

The supplemental materials are available at

Notes

References

Adamczyk

Pitt

(2009). Shaping attitudes about homosexuality: The role of religion and cultural context. Social Science Research, 38, 338-352.

Andersen

Fetner

(2008). Economic inequality and intolerance: Attitudes toward homosexuality in 35 democracies. American Journal of Political Sciences, 52, 942-958.

Anderson

C. J.

Singer

M. M.

(2008). The sensitive left and the impervious right: Multilevel models and the politics of inequality, ideology, and legitimacy in Europe. Comparative Political Studies, 41, 564-599.

Bardi

Buchanan

K. E.

Goodwin

Slabu

Robinson

(2014). Value stability and change during self-chosen life transitions: Self-selection versus socialization effects. Journal of Personality and Social Psychology, 106, 131-147. doi:10.1037/a0034818

Bauer

D. J.

Curran

P. J.

(2005). Probing interactions in fixed and multilevel regression: Inferential and graphical techniques. Multivariate Behavioral Research, 40, 373-400.

Bilsky

Janik

Schwartz

S. H.

(2011). The structural organization of human values—Evidence from three rounds of the European Social Survey (ESS). Journal of Cross-Cultural Psychology, 42, 759-776. doi:10.1177/0022022110362757

Bourdieu

(2001). The male domination. Cambridge, UK: Polity Press.

Brambilla

Butz

D. A.

(2013). Intergroup threat and outgroup attitudes: Macro-level symbolic threat increases prejudice against gay men. Social Psychology, 44, 311-319.

Brumbaugh

S. M.

Sanchez

L. A.

Nock

S. L.

Wright

J. D.

(2008). Attitudes toward gay marriage in states undergoing marriage law transformation. Journal of Marriage and Family, 70, 345-359.

10.

Central Intelligence Agency. (2012). Mother’s mean age at first birth (The World Factbook). Retrieved from https://www.cia.gov/library/publications/the-world-factbook/fields/2256.html

11.

Cialdini

R. B.

Trost

M. R.

(1998). Social influence: Social norms, conformity, and compliance. In Gilbert

D. T.

Fiske

S. T.

Lindzey

(Eds.), The handbook of social psychology (4th ed., Vol. 2, pp. 151-192). Boston, MA: McGraw-Hill.

12.

Coleman

J. S.

(1990). Foundations of social theory. Cambridge, MA: Harvard University Press.

13.

Cornelius

Leech

Goldschmidt

(2004). Characteristics of persistent smoking among pregnant teenagers followed to young adulthood. Nicotine & Tobacco Research, 6, 159-169.

14.

Crano

W. D.

Brewer

M. B.

Lac

(2015). Principles and methods of social research (3rd ed.). New York, NY: Routledge.

15.

Curran

P. J.

Bauer

D. J

Willoughby

M. T.

(2006). Testing and probing interactions in hierarchical linear growth models. In Bergeman

C. S.

Boker

S. M.

(Eds.), The Notre Dame Series on Quantitative Methodology: Vol. 1. Methodological issues in aging research (pp. 99-129). Mahwah, NJ: Lawrence Erlbaum.

16.

Davidov

Meuleman

(2012). Explaining attitudes towards immigration policies in European countries: The role of human values. Journal of Ethnic and Migration Studies, 38, 757-775.

17.

Davidov

Schmidt

Schwartz

S. H.

(2008). Bringing values back in: The adequacy of the European Social Survey to measure values in 20 countries. Public Opinion Quarterly, 72, 420-445.

18.

Davies

J. D.

(1965). The family’s role in political socialization. Annals of the American Academy of Political and Social Science, 361, 10-19.

19.

European Social Survey. (2012). ESS methodological overview. Retrieved from http://www.europeansocialsurvey.org/methodology/

20.

Ferguson

(2007). Gay marriage: An American and feminist dilemma. Hypatia, 22, 39-57.

21.

Gerhards

(2010). Non-discrimination towards homosexuality: The European Union’s policy and citizens’ attitudes towards homosexuality in 27 European countries. International Sociology, 25, 5-28.

22.

Gilani

I. S.

Shahid

Zuettel

(2012). Global index of religiosity and atheism. Gallup International, 1-25.

23.

Goldenberg

Saxe

(1996). Social attitudes of Russian immigrants to the United States. The Journal of Social Psychology, 136, 421-434.

24.

Haddock

Zanna

M. P.

(1998). Authoritarianism, values, and the favorability and structure of antigay attitudes. In Herek

G. M.

(Ed.), Psychological Perspectives on Lesbian and Gay Issues: Stigma and sexual orientation: Understanding prejudice against lesbians, gay men, and bisexuals (pp. 82-107). Thousand Oaks, CA: Sage.

25.

Hair

J. F.

Jr. Anderson

R. E.

Tatham

R. L.

Black

W. C.

(1995). Multivariate data analysis (3rd ed.). New York, NY: Macmillan.

26.

Heaven

P. C.

Oxman

L. N.

(1999). Human values, conservatism and stereotypes of homosexuals. Personality and Individual Differences, 27, 109-118.

27.

Heck

R. H.

Thomas

S. L.

Tabata

L. N.

(2013). Multilevel and longitudinal modeling with IBM SPSS (2nd ed.). New York, NY: Routledge

28.

Herek

G. M.

(2000). The psychology of sexual prejudice. Current Directions in Psychological Science, 9, 19-22.

29.

Hitlin

Piliavin

J. A.

(2004). Values: Reviving a dormant concept. Annual Review of Sociology, 30, 359-393.

30.

Hodson

Busseri

M. A.

(2012). Bright minds and dark attitudes lower cognitive ability predicts greater prejudice through right-wing ideology and low intergroup contact. Psychological Science, 23, 187-195.

31.

Hogg

M. A.

(2015). Social instability and identity-uncertainty: Fertile ground for political extremism. In Forgas

J. P.

Fiedler

Crano

W. D.

Forgas

J. P.

Fiedler

Crano

W. D.

(Eds.), Social psychology and politics (pp. 307-320). New York, NY: Psychology Press.

32.

Inglehart

(1987). Value change in industrial societies. American Political Science Review, 81, 1289-1303.

33.

Inglehart

(1990). Culture shift in advanced industrial society. Princeton, NJ: Princeton University Press.

34.

Inglehart

(1997). Modernization and postmodernization: Cultural, economic, and political change in 43 societies. Princeton, NJ: Princeton University Press.

35.

Inglehart

Baker

W. E.

(2000). Modernization, cultural change, and the persistence of traditional values. American Sociological Review, 65, 19-51.

36.

Inglehart

Norris

Welzel

(2002). Gender equality and democracy. Comparative Sociology, 1, 321-345.

37.

International Lesbian, Gay, Bisexual, Trans, and Intersex Association Europe. (2012). Rainbow Europe map and index 2012. Retrieved from http://www.ilga-europe.org/rainboweurope/2012

38.

Kluegel

J. R.

Mason

D. S.

(2004). Fairness matters: Social justice and political legitimacy in post-communist Europe. Europe-Asia Studies, 56, 813-834.

39.

Kohlberg

Hersh

R. H.

(1977). Moral development: A review of the theory. Theory Into Practice, 16, 53-59.

40.

Kuntz

Davidov

Schwartz

S. H.

Schmidt

(2015). Human values, legal regulation, and approval of homosexuality in Europe: A cross-country comparison. European Journal of Social Psychology, 45, 120-134.

41.

Lewin

(1951). Field theory in social science: Selected theoretical papers ( Cartwright

, Ed.). Oxford, UK: Harpers.

42.

Lind

E. A.

Kulik

C. T.

Ambrose

de Vera Park

M. V.

(1993). Individual and corporate dispute resolution: Using procedural fairness as a decision heuristic. Administrative Science Quarterly, 224-251.

43.

Monteiro

M. B.

França

D. X.

Rodrigues

(2009). The development of intergroup bias in childhood: How social norms can shape children’s racial behaviours. International Journal of Psychology, 44, 29-39.

44.

Morrison

M. A.

Morrison

T. G.

Franklin

(2009). Modern and old-fashioned homonegativity among samples of Canadian and American university students. Journal of Cross-Cultural Psychology, 40, 523-542. doi:10.1177/0022022109335053

45.

Newman

B. S.

Muzzonigro

P. G.

(1993). The effects of traditional family values on the coming out process of gay male adolescents. Adolescence, 28, 213-226.

46.

Peugh

J. L.

Enders

C. K.

(2005). Using the SPSS mixed procedure to fit cross-sectional and longitudinal multilevel models. Educational and Psychological Measurement, 65, 717-741.

47.

Poteat

V. P.

Mereish

E. H.

(2012). Ideology, prejudice, and attitudes toward sexual minority social policies and organizations. Political Psychology, 33, 211-224.

48.

Preacher

K. J.

Curran

P. J.

Bauer

D. J.

(2006). Computational tools for probing interaction effects in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics, 31, 437-448.

49.

Riek

B. M.

Mania

E. W.

Gaertner

S. L.

(2006). Intergroup threat and outgroup attitudes: A meta-analytic review. Personality and Social Psychology Review, 10, 336-353.

50.

Röder

(2015). Immigrants’ attitudes toward homosexuality: Socialization, religion, and acculturation in European host societies. International Migration Review, 49, 1042-1070.

51.

Sagiv

Schwartz

S. H.

(1995). Value priorities and readiness for out-group social contact. Journal of Personality and Social Psychology, 69, 437-448. doi:10.1037/0022-3514.69.3.437

52.

Schlüter

Meuleman

Davidov

(2013). Immigrant integration policies and perceived group threat: A multilevel study of 27 Western and Eastern European countries. Social Science Research, 42, 670-682.

53.

Schwartz

S. H.

(1992). Universals in the content and structure of values: Theoretical advances and empirical tests in 20 countries. In Zanna

(Ed.), Advances in experimental social psychology (Vol. 25, pp. 1-65). San Diego, CA: Academic Press.

54.

Schwartz

S. H.

(2003). A proposal for measuring value orientations across nations. Questionnaire Package of the European Social Survey, 259-290.

55.

Schwartz

S. H.

Caprara

G. V.

Vecchione

(2010). Basic personal values, core political values, and voting: A longitudinal analysis. Political Psychology, 31, 421-452.

56.

Schwartz

S. H.

Cieciuch

Vecchione

Davidov

Beierlein

Ramos

Konty

(2012). Refining the theory of basic individual values. Journal of Personality and Social Psychology, 103, 663-688. doi:10.1037/a0029393

57.

Steffens

Wagner

(2004). Attitudes toward lesbians, gay men, bisexual women, and bisexual men in Germany. The Journal of Sex Research, 41, 137-149.

58.

Stulhofer

Rimac

(2009). Determinants of homonegativity in Europe. Journal of Sex Research, 46, 24-32.

59.

Takács

Szalma

(2011). Homophobia and same-sex partnership legislation in Europe. Equality, Diversity and Inclusion: An International Journal, 30, 356-378.

60.

Terrizzi

J. J.

Shook

N. J.

Ventis

W. L.

(2010). Disgust: A predictor of social conservatism and prejudicial attitudes toward homosexuals. Personality and Individual Differences, 49, 587-592. doi:10.1016/j.paid.2010.05.024

61.

Tyler

T. R.

Degoey

(1996). Trust in organizational authorities: The influence of motive attributions on willingness to accept decisions. Trust in Organizations: Frontiers of Theory and Research, 331-356.

62.

Tyler

T. R.

Wakslak

C. J.

(2004). Profiling and police legitimacy: Procedural justice, attributions of motive, and acceptance of police authority. Criminology, 42, 253-282.

63.

United Nations Economic Commissions for Europe. (2012). Mean age at first marriage by sex (Statistical database). Retrieved from http://w3.unece.org/PXWeb2015/pxweb/en/STAT/STAT__30-GE__02-Families_households/052_en_GEFHAge1stMarige_r.px/?rxid=280b4a3f-7a19-4f99-baf8-f740afa242ec

64.

Van de Meerendonck

Scheepers

(2004). Denial of equal civil rights for lesbians and gay men in The Netherlands, 1980-1993. Journal of Homosexuality, 47, 63-80.

65.

van den Akker

van der Ploeg

Scheepers

(2013). Disapproval of homosexuality: Comparative research on individual and national determinants of disapproval of homosexuality in 20 European countries. International Journal of Public Opinion Research, 25, 64-86.

66.

Van den Bos

Lind

E. A.

(2002). Uncertainty management by means of fairness judgments. Advances in Experimental Social Psychology, 34, 1-60.

67.

Van Hiel

Duriez

Kossowska

(2006). The presence of left-wing authoritarianism in western Europe and its relationship with conservative ideology. Political Psychology, 27, 769-793. doi:10.1111/j.1467-9221.2006.00532.x

68.

Vermunt

Wit

van den Bos

Lind

E. A.

(1996). The effects of unfair procedure on negative affect and protest. Social Justice Research, 9, 109-119.

69.

Wojcieszak

Price

(2010). Bridging the divide or intensifying the conflict? How disagreement affects strong predilections about sexual minorities. Political Psychology, 31, 315-339. doi:10.1111/j.1467-9221.2009.00753.x

70.

Wong

J. P.

Poon

M. K.

(2013). Challenging homophobia and heterosexism through storytelling and critical dialogue among Hong Kong Chinese immigrant parents in Toronto. Culture, Health & Sexuality, 15, 15-28. doi:10.1080/13691058.2012.738310

71.

The World Bank. (2012). Employment in services, female (% of female employment). Retrieved from http://data.worldbank.org/indicator/SL.SRV.EMPL.FE.ZS

72.

Wright

D. B.

London

(2009). Multilevel modeling: Beyond the basic applications. British Journal of Mathematical and Statistical Psychology, 62, 439-456.

73.

Zmerli

Newton

(2008). Social trust and attitudes toward democracy. Public Opinion Quarterly, 72, 706-724.

Applying Multilevel Modeling to Understand Individual and Cross-Cultural Variations in Attitudes Toward Homosexual People Across 28 European Countries

Abstract

Keywords

Individual-Level Variables

Basic Human Values

Group-Level Variables

The Current Study

Hypotheses and Predictions

Method

Dataset

Participants

Individual-Level Measures 3

Gender and age

Education level

Born in residing country

Political affiliation

Perceptions of political fairness

Religiosity

Human values

Attitudes toward homosexual people

Group-Level Measures

Social conservatism

Gay civil rights

Analytic Plan

Results

MLM

Null model

Random intercepts with person-level predictors

Random intercepts with country- and person-level predictors

Random intercepts with cross-level interactions of country-level social conservatism and person-level predictors

Random intercepts with cross-level interactions between country-level gay rights and individual-level predictors

Discussion

Main Effects

Interaction Effects

Limitations and Future Directions

Conclusion

Footnotes

Declaration of Conflicting Interests

Funding

Supplemental Material

Notes

References

Individual-Level Measures³