You don't need a pollster to get the message here. Photographer: Uriel Sinai/Getty Images
You don't need a pollster to get the message here. Photographer: Uriel Sinai/Getty Images

Wouldn't it be interesting to poll people in 100 countries and find out what prejudices they have? Well, yes, if you have the resources. Unfortunately, the Anti-Defamation League, which has released a poll claiming to study anti-Semitism globally, didn’t quite do that. Instead of designing a neutral questionnaire that would compare anti-Jewish attitudes to negative attitudes about other religions or ethnicities, it asked a series of 11 questions that stacked the deck in favor of anti-Semitic answers. It then defined you as an anti-Semite if you answered yes to six of the 11 questions.

The results might tell you something about relative degrees of anti-Semitism in different places -- surprise: Saudis have a more negative attitude toward Jews than Danes do. But other than the rhetorical effect of announcing that a quarter of the world’s people are anti-Jewish, the poll offers precious little in the way of actual knowledge.

The poll’s methodology is what's most disappointing. Derived from a survey model first developed more than 50 years ago, when social science knowledge about polling was in its infancy, the questionnaire offers a series of statements made essentially without context. The person being questioned has only three options to answer: probably true, probably false, or “don’t recognize,” meaning the person basically has no idea what the pollster is talking about.

To give you a flavor, the first four read like this: Jews are more loyal to Israel than to this country; Jews have too much power in the business world; Jews have too much power in international financial markets; Jews still talk too much about what happened to them in the Holocaust”

Taken individually, these questions are problematic. Start with the first one, about Jewish loyalty, which the pollsters altered in countries with tiny numbers of Jews to read, “Jews are more loyal to Israel than to the countries they live in.” It's not at all clear to me that such a statement necessarily reflects an anti-Semitic worldview. I myself have no idea what Jews around the world would say in an anonymous poll about where their loyalties lie. But would it really be so terrible if, for example, Jews living under Vladimir Putin were more loyal to Israel than to Mother Russia? If I believed that some or even most Jews would take the side of Israel in the case of a serious conflict with their own countries, would that belief help qualify me as an anti-Semite? Don't get me wrong, it would be bad for Jews in many countries if such polls were performed. But thinking you know what they might say isn’t bias.

Questions two and three are worrisome because they’re essentially the same question -- ordinary people don't differentiate “the business world” from financial markets. Repetitiousness wouldn't be a problem in itself -- except that the anti-Semitism index asks whether you answered six questions affirmatively. Two questions based on the same prejudice distorts the results.

Then there's the Holocaust. Considering that about half of respondents had never heard of the Holocaust, it’s asking a lot for them to judge whether that tragedy is being discussed “too much.” But more fundamentally, suppose a person thought that discussion of the Holocaust does take up too much air-time in political discourse compared to other genocides or massive global crimes on the order of African slavery. Such a view might be wrong or right -- but is it by definition anti-Semitic?

Taken collectively, the questions are in some ways even more troubling. As the pollster makes a continuing series of anti-Semitic statements, the addressee may well have the experience of feeling increasingly reminded of whatever latent anti-Semitic beliefs he may have. Each statement makes the ones after it more salient in the listener's mind, and each may prime the listener for a particular response. Today we have a sophisticated understanding of how certain questions in a poll can shape the answers to others. A poll consisting of nothing but a series of anti-Semitic propositions runs the risk of becoming a cascade that the respondent will feel drawn to join.

Admittedly, not everyone is sucked in. Countries where people know little or nothing of the Jews -- say, Laos or Vietnam -- have low index scores. And in Nordic countries, as well as the U.K. and the U.S., where people know that the attitudes in question count as anti-Semitic, low scores are also produced. Countries with large concentrations of Muslims generally produce higher indices, explaining for example the 53 percent score for Senegal, much higher than other central African states with fewer Muslims. In a gross way, the poll reveals something.

But given its methodology, it's fair to ask what the purpose of the poll really is. And it's tempting speculate that it is aimed to convince us that the world needs the ADL, and that we should donate money to it.

Then there is a final, telling omission: Why didn’t the ADL poll in Israel? They did poll in the West Bank and Gaza, where they got the highest anti-Jewish result they found anywhere. Is it because no one really wants to hear how the world's only majority Jewish population would have done on the quiz?

To contact the writer of this article: Noah Feldman at noah_feldman@harvard.edu.

To contact the editor responsible for this article: Tobin Harshaw at tharshaw@bloomberg.net.