Abstract
We investigate the relationship between education and antisemitism using unique individual-level survey data on antisemitism from more than 100 countries. Our findings show that education is associated with greater favorability toward Jews, but the relationship between education and endorsement of antisemitic stereotypes and conspiracy theories varies between countries. In countries that actively supported recent statements condemning Holocaust denial and antisemitism at the United Nations—which we use as a proxy for country-level opposition to antisemitism in education and politics—greater education is associated with reduced endorsement of antisemitic stereotypes. By contrast, more educated people are more likely to endorse antisemitic stereotypes than less educated people in countries that declined to endorse those statements. These descriptive findings provide new evidence about the association between education and intolerance.
Introduction
It is commonly argued that education promotes greater openness and political diversity and decreases prejudice (Golebiowska, 1995). For example, previous research finds that higher levels of education are associated with reduced prejudice and outgroup hostility (e.g., Borgonovi, 2012; Easterbrook et al., 2016). These findings suggest that education can challenge prejudice and promote critical thinking and tolerance.
However, education can also promote intolerance in illiberal states, especially if teaching or curricula promote or reinforce negative views about outgroups (Zhang and Brym, 2019). For example, Saudi textbooks were found to contain antisemitic and anti-Western stereotypes after the 9/11 terrorist attacks (O’Hara, 2006).
Similarly, the association between education and anti-American attitudes and conspiracy beliefs in predominantly Muslim countries was stronger in countries whose education systems had less Western influence (Gentzkow and Shapiro, 2004). Similarly, education can promote greater political knowledge and engagement, which can have positive or negative effects on outgroup hostility depending on the content of political debate. If messages from elites reinforce norms of tolerance and inclusion, greater knowledge should lead to higher levels of tolerance. Alternatively, exposure to political messages can encourage outgroup prejudice and hostility.
In practice, elite attacks on outgroups are unfortunately common in political debate, especially in places with historical animosity toward outgroups (Blaydes and Linzer, 2012; Fearon and Laitin, 2000; Glaeser, 2005; Voigtländer and Voth, 2012; Zaller, 1992). Recent examples include Hungarian Prime Minister Viktor Orbán deploying antisemitic stereotypes and tropes against Jewish financier George Soros to rally his supporters and Pakistani Foreign Minister Shah Mahmood Qureshi accusing Israelis of “controlling the media” (Forman, 2018; Gul, 2021). Exposure to messages like these may explain, for instance, why the association between holding negative views of Jews and Israel and belief in conspiracy theories about the West, Jews, and Israel was found to be greater among people in Egypt and Saudi Arabia who are more knowledgeable about politics (Nyhan and Zeitzoff, 2018).
We test these competing expectations by examining the association between education and antisemitism, one of oldest and most virulent forms of prejudices. Using unique survey data collected in more than 100 countries, we measure the prevalence of belief in antisemitic stereotypes, many of which relate to conspiratorial beliefs that Jews have inordinate power and that Jewish-led conspiracies are responsible for many of the problems in the world (Hersh and Royden, 2021; Lipstadt, 2019).
Our results demonstrate that education is associated with greater favorability toward Jews, but the relationship between education and endorsement of antisemitic stereotypes and conspiracy theories varies between countries. Education is associated with reduced endorsement of antisemitic stereotypes in countries that actively supported recent statements condemning Holocaust denial and antisemitism at the United Nations, which we use as a proxy for country-level opposition to antisemitism in education and politics. By contrast, however, education is associated with greater antisemitic stereotype endorsement in countries that declined to endorse those statements at the U.N. These novel descriptive findings provide an important basis for future research on the relationship between education and prejudice.
Research design
Our study examines unique cross-national survey data on antisemitism collected by Anzalone Liszt Grove Research on behalf of the Anti-Defamation League (ADL), a U.S.-based NGO that monitors and campaigns against antisemitism. In 2013, 2015, and 2017, Anzalone Liszt Grove conducted nationally representative surveys in over 100 countries with more than 60,000 respondents—the largest survey to date measuring individual-level attitudes toward Jews and endorsement of antisemitic stereotypes and conspiracy theories. 1 The surveys asked respondents to indicate whether they have a favorable or unfavorable opinion of Jews and other religious groups. Respondents were then asked if a set of 11 common antisemitic stereotypes and conspiracy theories about Jews are probably true or probably false (e.g., “Jews have too much control over the global media”). These items exhibit high reliability as a scale (Cronbach’s alpha = 0.88). Country-level sample sizes, survey timing, and exact question wording are provided in the Appendix.
As a proxy for country-level differences in the content of educational curricula and political messages, we examine whether each country in the data sponsored a U.N. resolution in 2007 that “[c]ondemns without any reservation any denial of the Holocaust” and/or supported a 2015 U.N. statement that condemned “an alarming increase in Antisemitism worldwide.” 2 Our measure, which combines countries’ stances on both issues, corresponds closely with membership in the International Holocaust Remembrance Alliance (IHRA; see https://www.holocaustremembrance.com/about-us/countries-membership)—in total, 30 of 32 countries that appear in our data supported both statements are IHRA members (93.8%) compared to 7 of 33 that supported only the Holocaust denial statement (21.2%). (No countries supported the antisemitism statement but not the Holocaust denial statement.) None of the 37 countries that did not sign either statement are IHRA members (0%).
In our empirical analysis below, we estimate how country-level differences in opposition to antisemitism moderates the relationship between education and (a) intolerance of Jews and other minorities and (b) belief in antisemitic stereotypes and conspiracy theories. In our regressions, we interact a four-point scale measuring respondent education with whether their country actively supported neither statement, only the Holocaust denial statement, or both of the statements against antisemitism listed above, which we use as a proxy measure of country-level opposition to antisemitism in education and politics. 3 These regressions include fixed effects by country and survey wave to account for time-invariant differences between countries and differences over time across countries, respectively. Because we include country fixed effects, the indicators for number of statements supported are omitted. The interaction terms thus allow the slope of the respondent education term to vary by the number of statements supported at the country level, testing for the heterogeneous relationship of interest.
Results
We first examine the association between education and attitudes toward Jews and other minorities, and how it varies at the country level by our proxy measure of country-level opposition to antisemitism in education and politics.
We begin with the raw data in Figure 1, which reports the average proportion of “probably true” and “probably false” responses for antisemitic stereotypes by education level and whether their country supported neither statement, only the Holocaust denial statement, or both statements.
4
As respondent education increases in countries that supported neither statement (left plot), the proportion of antisemitic stereotypes endorsed increases, the proportion rejected stays relatively consistent, and the number of “don’t know” responses decreases. These results suggest that the positive association between education and stereotype endorsement is driven in part by reduced uncertainty. As a result, net endorsement increases from 23% to 36% as education goes from its minimum to its maximum. We observe the opposite in countries that supported both statements. For respondents in those countries, more education is associated with an increase in the rejection of antisemitic stereotypes and a decrease in endorsement (right plot). Stereotype endorsement by respondent education and country support for statements against antisemitism. Survey data from 2013 to 2014 measuring percentage of stereotypes endorsed by respondent education level (see Appendix for countries and question wording). Graph separates countries by whether they supported only the U.N. General Assembly Resolution 60/7 condemning Holocaust denial, both that statement and a Joint Statement against Antisemitism issued at the U.N. General Assembly in 2015, or neither.
Outgroup unfavorability and antisemitic stereotype endorsement by respondent education and country support for statements against antisemitism.
OLS regressions with standard errors clustered at country level in parentheses; *p
Higher levels of education are associated with less unfavorable (more favorable) feelings toward Jews and other religious outgroups, including in countries that did not support either statement against antisemitism. The magnitudes of these effects vary, however. In countries that supported both statements (which are majority or plurality Christian), this relationship was attenuated for Jews and Christians (p < .05 and p < .001, respectively). The opposite was observed in countries that supported neither statement, where the relationship between education and outgroup hostility was stronger for Jews and Christians than for other groups. As education moves from its minimum (1) to its maximum (4) on our four-point scale, unfavorable attitudes toward Jews decreased by 0.23 points on our three-point outcome measure (3 × 0.077, 0.27 sd) in countries that supported neither statement compared to 0.12 points in countries that supported only the Holocaust denial statement (3 × 0.040; 0.14 sd) and 0.17 points in countries that support both statements (3 × 0.058; 0.20 sd).
By contrast, as Table 1 indicates, higher levels of education have different directional associations with antisemitic stereotype endorsement. In countries that supported neither statement against antisemitism, education is positively associated with antisemitic stereotype endorsement (proportion rated “probably true” out of 11) and net stereotype endorsement (proportion rated “probably true” minus “probably false” out of 11), a pattern of results that is consistent with more educated people in those countries offering answers to both questions at higher rates (vs. saying they don’t know). The expected proportion of antisemitic stereotypes that respondents in those countries endorse increases by 6 percentage points (3 × 0.020, 0.19 standard deviations) as education moves from its minimum (1) to its maximum (4) on our 4-point scale. Because the coefficient for proportion true is larger, the net proportion of stereotypes endorsed increases correspondingly by 3.9 percentage points (3 × 0.013, 0.07 standard deviations) over the range of the education variable.
In countries that supported only the Holocaust statement or both statements, the interaction coefficients are instead negative for the net proportion endorsed (p < .001 for both). The magnitude of the latter is larger (−0.067 vs −0.016). As a result, the combined marginal association reported at the bottom of Table 1 is null for countries that supported only the Holocaust statement and negative for countries that supported both statements. The expected change in the proportion of antisemitic stereotypes as a respondent goes from the minimum to maximum level of education levels is −16.2 percentage points (−0.054 × 3, 0.29 standard deviations) in these countries.
We explore region-specific variation in these results in the Appendix and find results that largely match those in the main text, particularly once we account for baseline differences between regions. Education is generally negatively associated with less animosity toward Jews across regions (see Table A3). As we discuss there, the positive association between education and stereotype endorsement among countries that endorsed neither statement reported in Table 1 is observed most clearly in Eastern Europe (Belarus) and MENA. The association between supporting both statements and a more negative relationship between education and stereotype endorsement is observed most clearly in the Americas and Eastern Europe.
Conclusion
Using novel survey data from more than 100 countries, we show that education is consistently positively associated with greater favorability toward Jews and other minority outgroups. In contrast, the association between education and antisemitic stereotype endorsement varies by country depending on their support for statements condemning antisemitism and Holocaust denial at the U.N (our proxy measure for country-level opposition to antisemitism in education and politics). People with higher levels of education were less likely to endorse antisemitic stereotypes and conspiracy theories in countries that supported both U.N. statements. Conversely, people with higher levels of education were actually more likely to endorse antisemitic stereotypes and conspiracy theories than those who were less educated in countries that opposed both statements, illustrating how greater education is not always associated with reduced prejudice and that the association may even be reversed in countries where education and politics promote intolerance.
These findings represent an important first step toward understanding the relationship between education and prejudices like antisemitism, but face several limitations that should be addressed in future research. Most importantly, our findings are correlational; we document an association that deserves further study using designs that allow for causal inference. Second, our measure of country-level opposition to antisemitism is a proxy. Future research should seek to create comparable country-level measures of antisemitism in political discourse and educational materials such as official curricula or textbooks to provide greater insight into potential mechanisms for our findings. Finally, future studies should seek to capture behavioral outcomes along with survey measures.
Despite these limitations, our results suggest we should not assume education is a panacea for reducing prejudice. Future research should compare the direct effects of educating students directly about antisemitism and the Holocaust (e.g., UNESCO, 2018) with education alone.
Supplemental Material
Supplemental Material - How the relationship between education and antisemitism varies between countries
Supplemental Material for How the relationship between education and antisemitism varies between countries by Brendan Nyhan, Shun Yamaya, and Thomas Zeitzoff in Research & Politics
Footnotes
Authors’ note
We are grateful to the Anti-Defamation League for sharing the data used in this study and to Anna Getmansky, Joshua Kertzer, and Ethan Porter for helpful comments. The authors alone are responsible for the conclusions and any errors, however.
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
Supplemental material for this article is available online
Notes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
