The Pew Research Center’s recent independent survey of Jewish America is the most comprehensive national poll of the U.S. Jewish population in more than a decade, and we are pleased that it has captured the interest of the American public, and beyond.
Our survey builds on the Jewish community’s long and proud tradition of conducting pioneering research on the size and characteristics of U.S. Jewry. However, there are many differences between previous surveys of Jews and the new survey, which means that direct, exact comparisons between previous surveys and the new Pew Research study are not possible.
In his op-ed, Forward columnist J.J. Goldberg suggests, wrongly, that the Pew Research survey of American Jews should be compared to the 1990 National Jewish Population Survey rather than to the 2000-2001 NJPS and that using this earlier comparison point would result in a very different picture of demographic trends. Goldberg’s analysis is mistaken in several ways.
First, all of the National Jewish Population Surveys (conducted in 1970-71, 1990 and 2000-2001) differ from the Pew Research survey – as well as from each other – in their sampling methods, question wording and categorizations of Jews. As a result, it is highly misleading to take findings from any of the NJPS studies and, without further analysis, simply stack them against figures from the Pew Research survey. The importance of paying attention to different methods, questions and definitions is highlighted in some detail in our report.
In particular, Goldberg appears to take the category labeled “Born Jews With No Religion” in the report on the highlights of the 1990 study and compares it with the “Jews of No Religion” category in the Pew Research survey.
The 1990 study found that of the “core” Jewish population of 5.515 million people, 1.120 million, or 20%, were “Born Jews With No Religion.” Since the numbers look similar (20% in the 1990 study vs. 22% in the new survey), he mistakenly concludes that there has been no increase in the share of Jews of No Religion. His error is to assume that the two categories are the same. They are not.
The category called Born Jews With No Religion in the 1990 NJPS is considerably broader than the Jews of No Religion category in the new Pew Research survey. The new survey defines Jews of No Religion as U.S. adults who identify themselves, religiously, as atheist, agnostic or nothing in particular but who have at least one Jewish parent or were raised Jewish and who say they consider themselves Jewish or partially Jewish aside from religion.
This question – do you consider yourself Jewish? – was included in both the 2000-2001 NJPS and the Pew Research survey. But it was not asked in the same way, at least not of all respondents and about all Jewish adults in surveyed households, in the 1990 NJPS.
Because of the lack of comparability on this key question, there is no way to construct, retroactively, a narrower category in the 1990 dataset that can be compared with the Jews of No Religion category in the new survey. But there are at least two ways to try to see what has happened, over time, to the NJPS 1990’s Born Jews With No Religion category.
One is to take the Jews of No Religion category in the Pew Research survey and expand it to include people who describe themselves as atheist, agnostic or having no particular religion and who have a Jewish parent or were raised Jewish and who do not currently consider themselves Jewish aside from religion. This is, admittedly, a very rough comparison that ignores methodological differences between the NJPS 1990 and the new survey, but at least it creates categories that are similarly defined.
Using this definition, people Born Jews With No Religion made up 20% of all Jews in the 1990 survey (and about 19% of Jewish adults) and about 26% of all Jewish adults in the Pew Research survey. If we also include those who have a Jewish parent or were raised Jewish and who decline to specify their religion or who report practicing a religion other than Christianity (as the Born Jews With No Religion category in 1990 seems to have done), then the Pew Research figure rises still further, to 31%.
The second (and we think much better) way to try to discern how this category has changed over time is to compare the 1990 NJPS with two subsequent surveys that were designed to replicate its research design – the 2001 and 2008 American Jewish Identification Surveys conducted by Barry A. Kosmin of Trinity College in Hartford, Conn. Using his methods and definitions, the share of Jewish adults in the U.S. with a Jewish parent and “no religion” rose from 19% in 1990 to 28% in 2001 and 32% in 2008, as calculated from the adult counts shown in Table 2 on page five of the 2008 AJIS report. Both of these comparisons indicate – contrary to Goldberg’s assertion – that there has been substantial growth in the share of U.S. Jews who do not identify with Judaism as a religion.
In addition, there is another serious error in Goldberg’s column that we would like to correct. He suggests that the main demographic findings of the Pew Research survey, including trends in intermarriage rates, are based on comparisons with the 2000-2001 NJPS. This is not the case. On the contrary, the Pew Research Center’s report on the new survey clearly and repeatedly cautions that such comparisons are complicated by differences in methodology, question wording and categorizations, as well as by the well-known flaws in the 2000-2001 NJPS. In consultation with experts on the 2000-2001 NJPS, we reanalyzed the 2000-2001 data to create comparable categories, and we make comparisons only sparingly, always with appropriately cautious language.
The main demographic findings of the Pew Research survey are, in fact, based on data from the new survey itself. We can see from the survey, for example, that 17% of respondents who got married before 1970 say their spouse is not Jewish, while among those who got married in the 1980s, about 40% have a non-Jewish spouse, and among those who have gotten married since 1995, nearly six-in-ten have a non-Jewish spouse.
We can also see evidence of the increase of Jews of no religion in our survey itself. Among Jews in the aging “greatest generation” of Americans (born 1914-1927), 93% identify their religion as Jewish and just 7% say they have no particular religion, yet consider themselves Jewish aside from religion. That share rises in subsequent generations, and among Jews in the youngest cohort of current adults, the Millennial generation, fully 32% say they have no religion.
There are a number of additional problems with Goldberg’s analysis. For example, he appears to compare findings from previous studies as they pertain to the total Jewish population (adults and children) with findings from the new survey that pertain only to adults.
On some other points, such as whether intermarriage rates finally have leveled off over the last 15 years, Goldberg is entitled to his interpretation. But we would strongly caution readers against taking population totals or other figures from either the NJPS 1990 or the NJPS 2000-2001 and stacking them against findings from the new survey without the kind of careful re-analysis contained in our report. The first chapter of the report goes into some detail on how the population estimates differ depending on the definitions used, and it puts the new findings into historical context as well.
We hope our response helps to clear up any confusion that may have resulted from Goldberg’s op-ed. We at the Pew Research Center take great care in designing and implementing our surveys, and we strive to present our findings accurately, impartially and transparently. It is our wish that our survey of American Jews will continue to serve as a useful source of information for those interested in the characteristics, attitudes, and experiences of the American Jewish population.
Alan Cooperman is deputy director, Religion & Public Life Project, Pew Research Center. Greg Smith is director of U.S. Religion Surveys, Religion & Public Life Project, Pew Research Center