Notes on the Pew survey

(Crossposted to Mah Rabu.)

The Pew Research survey, “A Portrait of Jewish Americans,” released yesterday, has received a lot of attention in both the Jewish and the mainstream media. I don’t have anything more to add about the results themselves; many pages have already been written in the last 48 hours. But after reading both the data and some of the spin, I have several comments about what we can and can’t conclude from the data.

1) Orthodox Retention

There has been discussion of the retention rates among various age cohorts of Orthodox Jews, i.e. what percentage of Jews raised as Orthodox currently identify as Orthodox. This percentage is significantly higher among the younger age cohorts than among the older cohorts, leading some to conclude that the Orthodox world is more effective at retention at the present time than in the past.

This conclusion is not supported by the data. Let us consider an alternate hypothesis: The attrition rate of Orthodox Jews has remained constant over time. What results would we expect from this hypothesis? The percentage of raised-Orthodox Jews who currently identify as Orthodox should decrease with increasing age (since older people have had more time to leave Orthodoxy), and this is in fact what we see in the data. But we can be more precise in our predictions from this model: The percentage should decay exponentially.

To test this, I fit the numbers to an exponential curve. I made the following assumptions and simplifications (which were quick-and-dirty, but you can try it yourself with different assumptions): I assumed that 100% of Orthodox-raised Jews identified as Orthodox at age 18 (and all attrition occurred after this). I collapsed each age range (e.g. 18-29) to a single data point at the center of the age range. For the highest age group (65+), I assumed it went up to 90.

The result was that the data fit the exponential very closely (R2 = 0.9932), with an attrition rate of about 2.4% per year:

Of course we can’t conclude that there has in fact been a steady rate of attrition either! My point is just that this would be consistent with the data. There are many possibilities – it would also be consistent with the data that everyone who leaves Orthodoxy leaves during their 20s (which would mean that the attrition rate is in fact much lower for the current 20somethings). There’s just no way to determine from these data (which only provide a snapshot of the present time) which model is correct, without data from past generations.

2) Denominational Identification

First of all, the exact question asked on the phone survey was “Thinking about Jewish religious denominations, do you consider yourself to be [RANDOMIZE: Conservative, Orthodox, Reform] something else, or no particular denomination?” (Kudos to the survey writers for randomizing the order of denominations! Shame on the report writers, who did not randomize or alphabetize the order when reporting the results.) So the survey participants were asked about denominational self-identification. Any discussion of the results (whether by Pew itself or by the media) that references denominational affiliation is not accurately reporting the results; there was no question that asked about denominational affiliation.

This distinction is particularly important when it comes to the “Reform” label. Reform Jewish identity is complex and multifaceted, but there are many people for whom the “Reform” label doesn’t mean affiliation with the Reform movement or affinity to Reform Judaism as a system of religious belief and practice; it means “I don’t do anything; I’m Reform[ed].”

This phenomenon is supported by the survey data themselves: 20% of the “Jews with no religion” category (i.e. people who first described their religion as “none”, then answered in response to a followup question that they consider themselves Jewish aside from religion) identified as “Reform”. Of all the Jews who identified as “Reform”, only 34% are synagogue members (so this category does not represent URJ members), and 16% said that they never attend Jewish services (including high holidays).

This means that any results about the “Reform” population have to be taken with a grain of salt, and can’t be translated into generalizations about Reform Judaism or the Reform movement. This goes both for the results that make Reform look bad (e.g. low rates of Hebrew literacy or of seeing being Jewish as very important) and for the results that make Reform look good (being the largest denomination and having the highest “retention” rate).

What we can conclude is not that Reform Jews are likely to be Jewishly inactive in various ways, but that people who are Jewishly inactive in various ways are much more likely to identify as “Reform” than any of the other denominational labels.

3) Intermarriage

The intermarriage rates are based on the percentage of the “net Jewish” population, i.e. those who consider themselves Jewish. Now that there has been substantial intermarriage for more than one generation, there are many people who were raised by one Jewish and one non-Jewish parent and who have gotten married themselves. Such people may identify as Jewish or as non-Jewish, or may be on the fence. It’s probably fair to assume that people who are on the fence are more likely to marry non-Jews than people who identify unequivocally as Jewish.

Therefore, in the same way that unscrupulous school administrators can improve their average test scores and graduation rates by getting rid of students who are likely to lower those stats, Jewish leaders who want to lower the intermarriage rate may be motivated (consciously or otherwise) to alienate intermarried families so that their children don’t identify as Jewish (so that if those children go on to marry non-Jews, it won’t count as an intermarriage). This may explain the behavior of some parts of the Jewish community.

Note also that the intermarriage rates by denomination are based on current self-identification. If someone was raised Orthodox and then marries a non-Jew, it is unlikely that s/he currently identifies as Orthodox. Given that, the most surprising part of the 2% Orthodox intermarriage rate is that it’s so high – these are Jews who are married to non-Jews and continue to identify as Orthodox.

Finally, the intermarriage statistics are only for current, intact marriages. Therefore, the apparent rise in intermarriage over time may be somewhat misleading. The report notes that “some research indicates that “in-marriages” (marriages between people of the same religion) tend to be more durable than intermarriages; if this is the case, then the percentage of intermarriages in the 1970s and 1980s may have been higher than it appears from looking only at intact marriages today.”

4) Money

One result that raised some eyebrows was that only 76% of Ultra-Orthodox respondents refrain from using money on Shabbat (suggesting that 24% do use money on Shabbat). The explanation for this may be simple: The question wording was “Do you personally refrain from handling or spending money on the Jewish Sabbath, or not? [INTERVIEWER NOTE: IF ASKED, “REFRAIN” MEANS TO NOT DO SOMETHING]” This double negative may have been confusing (especially for respondents whose first language wasn’t English), and the clarification was only for people who requested it. Some significant number of Ultra-Orthodox respondents may have answered “no”, thinking they were saying “No, I don’t handle or spend money on Shabbat.”

10 Responses to “Notes on the Pew survey”

  1. Great post. One caveat: “Therefore, in the same way that unscrupulous school administrators can improve their average test scores and graduation rates by getting rid of students who are likely to lower those stats, Jewish leaders who want to lower the intermarriage rate may be motivated (consciously or otherwise) to alienate intermarried families so that their children don’t identify as Jewish (so that if those children go on to marry non-Jews, it won’t count as an intermarriage). This may explain the behavior of some parts of the Jewish community.”

    frankly, I doubt this.

    KRG · October 3rd, 2013 at 1:27 am
  2. Completely agree regarding points 2,3, and 4. My own observation is that the older one gets, especially if if she is married and is raising a family, the likely she is to leave Orthodoxy. Somewhere between 20 and 30, most people seem to settle on a religious identity. Of course that identity continues (or should continue)to develop and evolve but the changes become much more nuanced.

    Avraham · October 3rd, 2013 at 2:05 am
  3. Looking at just the data of this report, you’re right about point 1. Perhaps 17% of Orthodox Jews defect in their 20s, 26% in their 30s and 40s, and so forth.
    But there are other data from other surveys, so there’s no need to conjecture.
    Also, I’d like to see you plot out the rate of attrition to the Reform and Conservative movements. Between 18 and 30, only 1% of the Orthodox who leave join R or C. Among the older generation, it reaches 52%.

    Regarding your point about Reform identification, Orthodoxy may have the opposite problem. More people affiliate with Orthodox institutions than consider themselves Orthodox (or any other denomination). Internally, the terms of choice are “frum” and “shomer Shabbos,” not Orthodox. And Orthodox Jews tend not to see themselves as part of a movement or denomination, but as “just Jews” doing what Jews are supposed to do.

    Elli · October 3rd, 2013 at 10:35 am
  4. KRG-
    To make the same point in a way that sounds less like a conspiracy theory: Some of the explanation of the increase in the intermarriage rate may be not that Jews who in an earlier generation would have married Jews (or not married at all) are marrying non-Jews, but that people marrying non-Jews who in an earlier generation would have identified as non-Jewish are identifying as Jewish.

    BZ · October 3rd, 2013 at 10:47 am
  5. Re point 1: given how much more popular Conservative Judaism was in the 1950s, it seems to me that the Orthodox attrition among older Americans was essentially a cohort effect: that is, for some reason Orthodox Judaism just wasn’t very popular among Americans of that generation, and Conservative Judaism was.

    If I am right, the idea that younger generations will experience the same level of “Orthodox attrition” as older generations is questionable: those generations haven’t had the same life experiences as the 80 somethings who left Orthodoxy for other streams of Judaism. They weren’t first- or second-generation immigrations, and they don’t feel the same pressure to assimilate in order to feel like Americans.

    Can we verify which view is correct? We actually could, if the Pew poll had asked one more question: asking people who switched denominations “when did you switch”? If my view is right, the overwhelming majority of defectors would have defected in early adulthood, which means that the low attrition rates of post-Baby Boom Orthodox Jews will stay low as those Jews age.

    Woodrow · October 3rd, 2013 at 2:05 pm
  6. One of the most egregious errors seem to be the assumption that no longer identifying with mainstream orthodoxy means a lesser religious observance. I would suggest that 90% of those who left institutional orthodoxy do not identify at all with Conservative or Reform Judaism, and are likely as halachic ally observant as before. The fastest growing trend in orthodoxy is alienated bnei Torah, now forming something more similar to European or British Orthodoxy.

    GARY PICKHOLZ · October 5th, 2013 at 1:01 pm
  7. I finally got around to looking into the survey details and figured this is a good place to share a few things. First page 119 of has the survey margins of error. Calculations on the full population are a respectable +/- 3%. Measures in subpopulations are between +/-4% to +/-12% and this doesn’t even include the sub-sub populations. For example, the MoE for “Orthodox” is +/-9.1% (There were 517 Orthodox people interviewed) The MoE for ages 18-29 is +/-8.1%. The MoE for Orthodox within an age range isn’t listed, but it’s obviously much higher. Given this level of noise, its surprising the results fit within any type of reasonable looking curve.

    Another observation is that the Orthodox group included 326 people who were ultra-Orthodox and 154 who were Modern Orthodox. With the caveat that the margins of error are 12%, the questions where they divide these two streams of Orthodoxy show quite a lot of differences. When comparing these results to past surveys, it’s very possible that earlier surveys were skewed towards modern Orthodox and this survey has much more ultra-Orthodox.

    Dan Ab · October 9th, 2013 at 12:04 am
  8. Thanks, BZ. Can you help me parse out the effect of the ecological fallacy in the interpretation of this data? The ecological fallacy, IIRC from my stats academic friends, is that you can’t apply general data to individuals. The wikipedia article will provide a better summary than I can:

    I think that the implication for us here, is that–and I say this as a rabbi–relating to people is always as individuals, and never as aggregates. And it would be mistaken to take the broad demographic information as predictive of individual behavior.

    But it feels like it is the profession of many Jewish population demographers to get us to do just that–not in a corrupt way, but in the sense that demographers are probably somewhat blind to the limitations of their field.

    Does this make sense? I think it does. And I think it helps to explain why so often our individual experience, sometimes derided as “merely anecdotal,” doesn’t square with the larger survey data. Your thoughts?

    EW · October 9th, 2013 at 9:36 am
  9. The issue is that in most Orthodox synagogues Shabbat and Yom Tov attendance pack the synagogues with multiple minyanim. Home practice and observance as a family in addition to regular service attendance sets the tone. Most Jews do not attend services because they have other priorities such as sports events, dance recitals etc. They feel it is more important for their children to be active sports participants on Shabbat than to attend regular services. How many post Bar and Bat Mitzvah students attend services on a regular basis? Many synagogues are bar and bat mitzvah mills where parents drop membership after the last bar or bat mitzvah or just do not attend except for the High Holy Days.RABBI DR. BERNHARD ROSENBERG

    ROSENBERG · October 17th, 2013 at 7:02 pm
  10. I think the larger issue here is that the Jewish communities of North America are drowning in a sea of multiple expensive, top-heavy infrastructures which don’t have tons of overlap with each other – and upon which way too many people depend for employment.
    Don’t believe me? Check out the listings at — and don’t count the number of organizations posting, count the actual positions available. That right there should be raising red flags as to how unsustainable the old model of movement-based Judaism has become.
    The Reform movement alone, in its efforts to protect its brand, pends untold hundreds of thousands of dollars to prop up outdated structures and a way of doing non-profit work that no longer matches either demographic OR financial trends in North America.
    Add to this the ongoing struggle to engage younger adults (too often at the expense of all other age demographics) and how miserably it is failing in some quarters, and it makes complete sense that we’re seeing an uptick in the number of smaller, more agile, and UNaffiliated shuls and havurot. Traditional North America Jewish institutions are going to end up being their own version of the LaBrea Tarpits if leadership doesn’t get its collective heads out of the and very soon.

    batiya · November 6th, 2013 at 7:40 pm

Leave a Reply

If your comment does not immediately appear, do not freak out and repost your message a dozen times. Please note that all new visitors must have their first comment approved by the editor, and you must provide a legitimate e-mail address and use the same username for the system to "remember" you. The editor maintains the right to refuse comments deemed inappropriate or unhelpful. Users who repeatedly delve into ad hominem attacks or other troll-like behavior will be banned.

Trackback (Right-click & 'Copy Link...') | Comments RSS

"I may attack a certain point of view which I consider false, but I will never attack a person who preaches it. I have always a high regard for the individual who is honest and moral, even when I am not in agreement with him. Such a relation is in accord with the concept of kavod habriyot, for beloved is man for he is created in the image of God." —Rav Joseph Soloveitchik