1.2 million more US Jews: Hanukka gift or flawed study?

AJC population study's methodology challenged by HU expert.

hanukkah mall 298.88 (photo credit: Ariel Jerozolimski [file])
hanukkah mall 298.88
(photo credit: Ariel Jerozolimski [file])
For the American Jewish Committee, it's a Hanukka miracle. A new study, conducted for the AJC's American Jewish Yearbook, may add over a million American Jews to the currently accepted count of just over five million. However, some, such as renowned Hebrew University demographer Prof. Sergio DellaPergola, have expressed serious doubts regarding the alleged miracle, claiming that the new figures are marred by intrinsic problems that call the findings into question. The study is a synthesis of some 50 local community studies that counts a US Jewish population of some 6.4 million, fully 1.2 million Jews more than were counted in the National Jewish Population Survey from 2000. The gap seems inexplicable. How can American Jewry number 5.2 million, while the communities count 6.4 million Jewish members? According to DellaPergola, this is due to the innate duplication built into local studies. The findings of the new study's coauthors, Dr. Ira Sheskin of the University of Miami and Dr. Arnold Dashefsky of the University of Connecticut ("with whom I've corresponded and argued as good friends," DellaPergola emphasizes), are marred by this double-counting and other inconsistencies, which he believes stem from several sources. The surveys are conducted "in different years, with some from 15 and even 20 years ago. They employ different research methodologies by different researchers, and are funded by different organizations, each with a different agenda," he told The Jerusalem Post on Thursday. Their localism also exposes them to more nonscientific pressures, since "local communities, which pay for the studies, want to show more members," DellaPergola added. "I'm not saying this happened," he emphasizes, but insists "there's a danger of [wanting to] 'improve' the results through methodology, phrasing of the questions and altering definitions." The third problem with the study, he said, is that "America has incredible internal mobility, with more than a quarter of Americans moving house every five years." This, coupled with the fact that the studies are from different years, adds inevitably to the potential duplication in the counting. "Being written into the Phoenix, Arizona, phone book doesn't erase you from the New York City phone book," he noted. Furthermore, many Jews - hundreds of thousands, according to DellaPergola - own two homes, one in the northern states and another in a warm southern state such as Florida. They can thus belong to two Jewish communities at the same time, further increasing the duplication in local studies. "In the national survey," he concludes, "everyone exists once." Study coauthor Dr. Ira Sheskin, director of the Jewish Demography Project at the University of Miami, partly agrees with DellaPergola's critique. But he told the Post on Thursday that the new findings have partly compensated for it. "Some Jews spend part of the year in one community and part in another community," he concedes, noting that "in our [Yearbook] article we say about 12% of Jewish households do that." However, he argued that the new study tried to compensate for it. "We haven't counted the snow-bird calculation in Florida's numbers. There's a separate estimate in our article of how many Jews we estimate are living in Florida for less than 8 months a year," he says. "In Palm Beach, Florida, for example, 20% of the Jewish population isn't counted." In addition, Sheskin admits, "5% of American Jews are students, and a student studying in Atlanta with a family in Miami is probably being counted twice. However, we've eliminated the students living in dormitories. Only students living off-campus are counted in the community studies." Thus, "We've done what we could to eliminate the overlap." So what's the real number? "I'm willing to bet it's around 6 million, or at least 5.9 million," Sheskin says. And while he concedes that the figure of 6.4 million is almost certainly too high due to the difficulties associated with the counting, "There's little doubt in anybody's mind that the NJPS in 2000 underestimated the number of Jews." Sheskin's explanation for this axiom follows a test he conducted after the national survey in 2000. The accepted method for discovering the number of American Jews, he says, is called "random digit dialing." If 30,000 random phone numbers are called in a certain area code, and 300 of them say they're Jewish, then the area is 1% Jewish. In this way, for example, researchers could determine that some 20% of Palm Beach, Florida, households are Jewish. In North Dakota, meanwhile, the figure is closer to one household in a thousand. Once researchers have the percentage in hand, they simply multiply that figure by the number of households in the area. Thus the percent figure discovered by the national study was multiplied by 105 million American households. But, says Sheskin, in the 2000 NJPS, "tens of thousands didn't stay on the phone long enough to answer the questionnaire." So he developed a small test to determine the significance of the lost information. "We took a list of distinctive Jewish names, such as Goldberg or Schwartz. Without compromising the identity of those questioned, we found that 0.37% of those who failed to respond had distinctive Jewish names. For those who responded, only 0.16% had a distinctive Jewish name." In other words, Jewish households probably made up a higher percentage of those who refused to respond than of those who responded, pushing the percentage determined by NJPS 2000 down significantly. And, multiplied by 105 million American households, even a 0.1% error excludes a significant number of people. Asked why Jews likely were less responsive than non-Jews to the survey, Sheskin replied, "I think we have a situation where response rates on surveys have gone way down. It was a lot easier in the 1970's and 1980's. People were intrigued by surveys. Very often today it's difficult to get people to stay on the phone." This means that "all the estimates, both the national study and the local studies" are "a little more suspect," he says, since "Jewish households are less likely to stay on the phone." Another issue with which DellaPergola criticizes the findings is the lack of "historical continuity." "American Jewry exists at least 350 years," he told the Post on Thursday. "The question is how many [Jews] remained year after year, and whether there is continuity in the numbers." There are only six factors that affect the number of American Jews, according to DellaPergola: birth, death, immigration, emigration, assimilation and conversion. "Today's number must be the result of a previous figure plus the changes brought by these factors," and a new study must therefore explain "how these [factors] affected today's number," he said. For example, "between 1945 and 2005 there's been a new survey approximately every 10 years, and they always fit the previous ones, considering the factors of change," he noted. The factors DellaPergola cites are well-known: "Childbirth dropped since the 1970's, and the population has aged, so that slowly more Jews are dying than are being born. There has been enormous growth in intermarriage since the 1960's, reaching more than 50% today, and most children of intermarried couples are defined as non-Jewish by their own parents. "According to the surveys - as opposed to the declarations of the religious movements - there are few converts, slightly less than the number of Jews converting to Christianity, and this is another negative balance." Thus, he believes, "the broad perspective of historical continuity says that [the 2000 NJPS figure of] five to 5.5 million Jews makes sense." For Sheskin, however, what DellaPergola demands is impossible. "In the American environment, in which we depend on survey results, we can't deal with birthrate or death rate over 10 years," he says, since they are unavailable to researchers. Thus, lacking an organized census, researchers must depend more on statistical reasoning to infer population data. Finally, says DellaPergola, "we have to think seriously about who is a Jew." The definition of "Jewish" was also a problem in the national survey, he notes. "The definitions have loosened to the point of impossibility [in terms of demographic analysis]," he insists. "People with two religions - say Catholic and Jewish - are they Jewish? The national survey decided that they are part of the count. So the [figure of] five to 5.5 million includes a few hundreds of thousands - I estimate less than half a million - who say they're Jews but also something else." Since "the survey focuses on the Jewish side," he explains, "we don't ask about the non-Jewish identity, which also exists." The problem of defining "Jewish" makes the national study's figures relatively large. "Even the lower count [of the NJPS] includes a distant periphery," which DellaPergola estimates at "about 800,000 people who were thought, by their answers, to be non-Jewish at first." These are people who answered "nothing" or "atheist" as their religious affiliation, and only "with effort" did researchers conclude that they were Jews who don't identify religiously. These difficulties are magnified significantly, he adds, in the local studies, where "there is no uniformity to the definition of the Jew." While Sheskin insists "we didn't count those who converted out," he agrees with DellaPergola's general critique, even though he believes they work in both directions. "If in Israel you only counted Jews who call themselves religious, you'd have a lot less there too," he notes. "Sergio wants to say that there are more Jews in Israel now than there are in the United States. In order to do that, he has to define "Jewish" by the definition in the United States." Yet, while "there's no doubt that the national study underestimated" the American Jewish population, the amalgamation of the local studies overestimated it, Sheskin admits. "The truth is somewhere in between," he concludes.