123 4 as a concrete policy measure to match the policy measures I use for the expectations conjoint experiment. The question on same-sex adoption is an excellent proxy for views on homosexuality because it correlates strongly with many other measures of views on homosexuality (Takács et al., 2016: 1795). Furthermore, while the largescale European Social Survey asks respondents whether they agree that “gays and lesbians are free to live their lives as they wish,” this question is becoming increasingly vulnerable to ceiling effects. More than 80 percent of people in France and Germany and more than 90 percent in the Netherlands “agree” and “strongly agree” with this statement (Fischer, 2019: 3). Tolerance towards homosexual marriages is higher than tolerance towards adoption by homosexual couples, arguably because child-rearing is not considered private, but part of the public sphere (Takács et al., 2016: 1787, 1789). I measured perceived similarity through asking: “How different or the same do you consider yourself to be from:” 1) “native French/German/Dutch people,” 2) “French/ German/Dutch Turkish people,” 3) “people who practice Islam,” 4) “people who practice Christianity” and 5) “people who do not practice any religion” on a scale of 0 to 1023. The initial scale ranged from 0 being “completely the same” to 10 being “completely different.” In order to enhance readability I recoded the scale to range from 0 being “completely different” to 10 being “completely the same.” Figure 1 shows that attitudes towards same-sex adoption have somewhat of a tri-polar structure. I therefore categorize the views of voters as anti (answering 0), middle (answering 1-9) and pro (answering 10). I call people who answered 0 antiflankers, people who answered 10 pro-flankers, and those who gave some position in the middle moderates. This categorization of respondees is the most symmetrical and results in groups that are relatively similar in size. In the appendix, I show the results for different definitions of flankers and moderates and demonstrate that the results remain similar. I stacked the data for France, Germany and the Netherlands. I chose these three countries as they offer three sampling frames in large West European democracies. All three countries have witnessed mass immigration since the Second World War and report similar levels of acceptance of Muslim immigrants (Alba and Foner, 2015; Foner and Alba, 2008). All three countries have a history of politicians venting xenophobic and particularly Islamophobic rhetoric in their national parliaments (Brubaker, 2013). The transnational LGBTQAI movement has made similar inroads in terms of social attitudes towards homosexuality partially due to the transnational character of the movement (Ayoub, 2016, 2019). Although The Netherlands was the first in partnership, marriage and adoption rights, all three countries are considered “early” compared to other countries across Europe (Dotti Sani and Quaranta, 2022: 132). The attitudes towards homosexuality have developed very similarly in all three countries (Abou23 LISS (2018) LISS Panel Data Archive. Available at: https://www.dataarchive.lissdata.nl/.
RkJQdWJsaXNoZXIy MjY0ODMw