Europeanization of attitudes towards homosexuality: exploring the role of education in the transnational diffusion of values

Abstract

How does exposure to EU integration affect social attitudes in third countries, and what is the role of education in this process? These questions matter because the EU aspires to be not only a regulatory regime but also a community of values. Addressing both the common elision of attitudes and the underappreciation of education in Europeanization research, this article analyses the EU’s impact on attitudes towards homosexuality in third countries via national education systems. This article offers not only an exposition of how we can expand our theoretical framework to include the transmission of values; its empirical findings demonstrate the need for this exercise, as some of the literature’s underlying assumptions are shown to be unsupported.

Keywords:

• Europeanization
• social learning
• education
• attitudes toward homosexuality
• socialization
• LGBT

ORCID

Koen Slootmaeckers http://orcid.org/0000-0002-1189-5095

Notes

1. Although we are aware of and sympathetic to the critical debates regarding the EU fundamental rights identity, it is beyond the scope of this article to unpack the so-called European values in much detail. In fact, we are here interested in how the EU’s investment in the promotion of the values it has identified as its foundation, such as anti-discrimination, can influence third countries’ societies.

2. The notion of Europeanization of attitudes is here not used to denote attitudes becoming more European, as this would erase the negative attitudes that exist within the older EU member states. Rather we use the notion to refer to how attitudes change as part of the European integration process. In other words, although the term can be read from an orientalist perspective, the reader should refrain from doing so.

3. Although this article deals primarily with lesbian, gay, and bisexual issues, we use the acronym LGBT to be consistent with EU documents and the literature’s broader use of the term. To avoid an overload of acronyms, we also use the terms gay and LGBT interchangeably.

4. For example, van den Akker, van der Ploeg, and Scheepers (2013, 1) hypothesize that the “changing European political structure towards increasingly close cooperation [has] led to the demand for more universal policies”. Others have studied the impact of same-sex marriage legislation on individuals’ attitudes towards homosexuality (Hooghe and Meeusen 2013; Takács and Szalma 2011). However, with the exception of a study by Zapryanova and Surzhko-Harned (2016), neither the EU nor the processes of Europeanization that undergird it have been explored as a causal factor on attitudes vis-à-vis gay people.

5. Börzel and Risse (2007) later clarified the relationship between misfit and adaptational pressure. They refined their argument by stating that misfit only results in adaptational pressure with an active intervention of actors (European or domestic).

6. The external incentives approach is by many authors equated with the conditionality principle used by the EU to govern its relations with other countries (see e.g. Glüpker 2013; Koinova 2011; O’Dwyer 2012). While the two concepts are indeed closely related, using them interchangeably conflates the distinct nature of both concepts. While the external incentives model should be regarded as a theoretical framework explaining the impact of the EU on domestic policies, conditionality is a EU tool to influence third countries and thus an example of the causal mechanisms proposed by the external incentives model.

7. At the level of activism – both supporting gay rights and opposing them – we may expect, of course, that international pressure will shape the issue’s politics (see O’Dwyer 2012), but that is an arena of attitudinal change beyond the one of concern in this article.

8. Börzel and Risse (2007) in a later account distinguished processes of persuasion from processes of socialization, arguing that the former follows the logic of communications rather than the logic of appropriateness. However, as persuasion can be translated into the logic of appropriateness, we consider both processes as part of the latter logic.

9. We want to stress that we are aware of the problematic nature of the notion “teacher of norms” as used in the literature and do not wish to subscribe to the underlying idea of the EU’s exceptionalism of human rights protection and value-based superiority. Although we agree that such critical analysis of EU values and their political nature is needed, it is beyond the scope of this article to do so.

10. Despite this, most existing studies of the EU’s impact on LGBT rights focus on the public policy aspect (Ames 2004; Kochenov 2006; Langenkamp 2003; Slootmaeckers and Touquet 2016).

11. The countries included in data are as follows: Albania, Armenia, Azerbaijan, Bosnia and Herzegovina, Bulgaria, Belarus, Switzerland, Cyprus, Czech Republic, Estonia, Georgia, Croatia, Hungary, Iceland, Lithuania, Latvia, Moldova, Montenegro, Macedonia, Malta, Poland, Romania, Serbia, Russia, Slovenia, Slovakia, Turkey, and Ukraine. The 28 selected countries are those countries that were not part of the EU at the time when LGBT issues became officially part of the EU’s legal framework (cf. the Amsterdam treaty that introduced anti-discrimination based on sexual orientation into the Treaties). Another reason for excluding the old EU member states is that doing so would distort the analysis due to the continuously observed attitudinal gap between the old and new EU member states. Including the old member states in the analysis would make it impossible to attribute whether the effects observed of EU membership are due to becoming an EU member or the result of pre-existing attitudes.

12. See also Finlay and Walther (2003), Herek (2007) and Sears (1997). Although the standard model of attitudes towards homosexuality includes a measure of contact with homosexual, we did not include this measure due to the unavailability of the data.

13. We have modeled our analysis with a categorical version of the education variables which delivered similar results to those reported. We have decided to report the continuous version of the variable as this version eases the interpretation of the interaction term in model 5.

14. We are aware of the measurement issues associated with the ISCED in relation to comparative research (for more details see Schneider and Kogan 2008), but as the best data available and the with our focus on a relative comparison between those with higher levels of education and those with low levels of education, this measure suffices for our analyses. We are not interested in comparing the different categories against each other, but rather look at the additional impact of being on a higher level of education people’s attitudes. We are also aware that by treating the variable as a continuous variable we presume a linearity of the educational system, which might not necessarily reflect the reality, but for our explorative purposes this suffice. Moreover, we have run the analyses using a categorical version of the data, with education measured across three categories, which has yielded the same results. We have opted for the continuous version of our model as the results are easier to interpret.

15. For proof that the incentive of future integration is not trivial for ENP states, one need only consider the example of Ukraine, whose Maidan protests were initially sparked by the threat of losing its path to future integration with the EU in November 2013.

16. In East Central Europe such association treaties were referred to as “Europe Agreements” whereas in the former Yugoslavia they are generally called Stability and Association Agreements.

17. While this operationalization of conditionality might be criticized as too broad-brush, it has the advantage of avoiding the other peril in measuring this variable cross-nationally, i.e. introducing ad hoc assessments of each country case based on a shifting set of context-specific factors. Recognizing the potentially greater bias of that approach, a country’s legal status in the accession process is the standard that other scholars have used in research such as ours. Ayoub (2014), for example, uses application/accession to the EU as a proxy for conditionality in his study of the European integration’s influence over policies regarding LGBT people. Probably the most fine-grained operationalization of conditionality in large-n research on EU accession and domestic politics is that of Schimmelfennig and Scholtz (2008, 195–196). They create a typology of accession and potential accession states that incorporates two dimensions: the size and credibility of incentives on offer over the course of the accession process. We have decided against using their more fine-grained operationalization of conditionality, as in our dataset, it would not allow sufficient variation in the variable. Given the no-effect hypothesis for conditionality, we believe a more parsimonious operationalization enables us to test a general effect of the external incentives model. Moreover, practically speaking, they only mention one instance where a significant lag obtained between the size and credibility of incentives, and it regarded Euro-Mediterranean states, not post-communist ones (2008, 196).

18. As a robustness check we have also included a measure of GDP per capita: however, as this control variable was not significant in our analysis and did not substantively change our results, we have opted to not report this model.