Data Deniers’ Delight: The Mormon Gender Gap Is Unremarkable

Pew’s 2007 Religious Landscape Survey (of the US) found that 56% of respondents who identified as LDS were women. Is that a lot?

Women in the US are more religiously observant in general, but even against that background, Dave of T&S has argued that 56% is a lot. Here’s where his conclusion comes from: If you look at this chart from the Pew survey, you’ll see Mormons are third highest in percent women. (Sorry the link points to a PDF. Although I can’t link to it directly, you can get the same chart by going to the Comparisons page, clicking “Demographics” and then “Gender Compositions of Religious Traditions.”)

But this chart is misleading if we’re interested in where the LDS Church stands relative to other churches in terms of what percentage of members are women. The chart aggregates people into broad religious traditions rather than considering individual churches. For example, Protestants are grouped into Evangelical, Mainline, and Historically Black, and Jews are all lumped together. The same thing happens with Mormons, which include other followers of Joseph Smith’s restoration in addition to Latter-day Saints, but here it makes little difference, as LDS make up 96% of the Mormons surveyed.

In order to get a better picture of the gender gap in the LDS Church compared to other churches, I looked at two more detailed charts in the Pew survey report. First, I looked at this one, which breaks out religious traditions into churches for everyone but Protestants, and then I looked at this one, which breaks out Protestant denominations. Using these charts allows us to compare the gender breakdown of churches to other churches rather than comparing the LDS Church to broad religious traditions, as the first chart effectively does.

In order to get a better look at the data, I put the percentage female for each church/denomination in the following chart. It is sorted by percent female.

LDS are certainly on the high side, but not dramatically so. We’re lower than 10 other churches, tied with 4, and higher than 25 (actually 22 if we ignore those who aren’t religious–unaffiliated secular, agnostic, and atheist).

But even beyond our unremarkable position, there’s still the question of which differences between churches are even large enough that we can be confident they’re not just a result of the particular sample of people interviewed. In order to answer this question, I used a series of statistical tests (Fisher’s exact test) to compare each other church against the LDS Church in terms of their percentage female.

In the chart above, churches whose percentages female are reliably different from the LDS percentage have gray bars. Churches not reliably different from the LDS church have blue bars. Not surprisingly, it’s the churches at the low end that the LDS Church can be statistically distinguished from. (Note that the difference with Southern Baptists is reliable while the differences with the next two aren’t because lots of Southern Baptists were interviewed and the ability of the test to distinguish differences depends on sample size.) But more interesting is the fact that the LDS percentage is statistically indistinguishable from percentages from 27 other churches.

So, the Church is in the upper half (maybe on the border of the upper third) of churches in terms of gender gap. But, and this is a big but, the sample size by church in the Pew study isn’t big enough to draw firm conclusions about most of the differences between churches. Therefore, I think it’s clearly an overstatement to say that the LDS Church has an unusually large gender gap.


  1. Thanks for contributing to the discussion, Ziff. I appreciate your concession that “LDS are certainly on the high side” of the gender gap.

    Your use of the term “statistically indistinguishable” seems likely to mislead some readers — it suggests the data support the claim that the mean LDS female participation rate is equal to the mean female participation rates of those denominations for which the data was not sufficient to reject the null hypothesis at your chosen level of significance. But failing to reject the null hypothesis is not the same as statistical support for the null — and should not be described as such.

    Even when the data don’t permit rejection at the chosen statistical level, they are still strongly suggestive, aren’t they? Even the disaggregated data you presented suggest the LDS gender gap is wider than average, as you noted.

  2. In spite of Dave’s critiques, even if there really is a bigger gender gap among LDS as opposed to these other Christian congregations, a difference of one or two percent may not be substantively significant. The main finding here is that this gender gap is common among many Christian denominations; most of these congregations are somewhere between 51% and 60% female, but it is unusual to be more than 60% female. LDS are pretty similar to most other Christians in this respect.

  3. But failing to reject the null hypothesis is not the same as statistical support for the null — and should not be described as such.

    That’s a fair point. Perhaps you could suggest a better wording? At some value, though, the differences really are indistinguishable. For example, 19 of the p-values were greater than .30. 16 were greater than .40. At some point, the null just isn’t going to be rejected for any reasonable alpha level.

    Even when the data don’t permit rejection at the chosen statistical level, they are still strongly suggestive, aren’t they?

    No, I don’t think they are. The gender ratios are all clustered so close to 1:1, and the sample sizes so small (relative to the differences) that I think it’s a stretch to conclude that they’re strongly suggestive.

    Clearly we both have our biases, but as I read it, your argument is that the Church has a dramatic, outlying gender gap, that is a big finding feminists must account for if we’re going to say the Church doesn’t treat women well. I think it looks like the Church’s gender gap is pretty middle of the road, and if above average, certainly nothing like an outlier.

  4. Maybe the LDS church is like Wal-Mart: Lots and lots of female employees, but all the managers are men.

    Sounds like a gender gap.

  5. As a single woman, I’d say the gender gap cannot be understated. If we’re all supposed to marry good Mormon men, a large population of Mormon women are going to remain single.

  6. Whitney: “a difference of one or two percent may not be substantively significant.”

    Depends on how precise your measurement is. Let’s say for sake of argument that the measured value (56/44%) is exactly the actual population value. There’s what, 7 million American Mormons? If the gender gap is indeed because the Church is doing something to chase men away, then we’re talking ~840,000 missing men. Is that substantial? Even just a 1% gap—70,000 “missing” persons—is worth considering.

    Ziff: nice post, especially the part where you break down each church/denomination. In my work, I commonly hear colleagues present data that is not significantly different but, in hopes of still defending their hypotheses, they note that “it is trending in [a certain direction].” I always think, If it isn’t different, then it isn’t trending.

    I’m not sure I would conclude as you that “LDS are certainly on the high side.” The high side of the overall study, yes, but when compared to other Christian denominations—a fair restriction, I think—LDS is right in the middle; i.e., it may not be so much our LDS-ness as our Christianity-ness that skews our gender gap. (In contrast, compare Jewish-reform to Jewish-conservative.)

  7. I also find the Hindu numbers interesting. One thing it leads me to wonder is what percentage of Indian Americans are male/female. Any idea where to get that number, Ziff?

  8. Howard – did the poll include all the (disproportionately) black men currently incarcerated? Kind of hard to go to church with the wife and kids when you’re in jail.

  9. That’s a great question, Macha. I really don’t know, but it seems like a really consistent pattern, at least in the US.


Comments are closed.