Recently FAIR (via Ask the Apologist feature) was queried about whether the Church’s website was knowingly misusing the popular 6% divorce from temple marriage statistic. Its main promoter has been Daniel K. Judd. He gave a BYU devotional in 2006 and defended the 6% figure. A partial transcript of Judd’s comments about his prior (2000) LA Times newspaper interview can be found here. Judd explains that divorce statistics are very dependent on how one collects and calculates the data. My co-blogger, Steven Danderson, pointed out that the high divorce rates that people are most familiar with are calculated (for example by the government) on a yearly basis by dividing the number of recorded divorces by the number of recorded marriages. As will be shown, the research that Judd refers to uses a different counting scheme, which is nevertheless well within the norms of academic journals. I think Judd and the LDS Church can continue to use the figure in good faith.

It is kind of silly all the criticism it has gotten on ex-mormon forums and sites without anyone tracking the original research to see how the numbers were calculated. Even a neutral site like this one gets a lot of things wrong. The study used by Judd was reprinted in a book he editted for the Religious Studies Center. The original article used 1981 data and was published as “Religion and Family Formation” in Review of Religious Research June 1985, Vol. 26:4 by authors Tim Heaton and Kristen Goodman. Heaton’s co-author was from the Church Correlation Department.

The Mormon data in the study was gathered by taking a random sample (n = 7446) of adults (18+) in the US and Canada based on Church records. They mailed or phoned surveys with a 81% response rate, but 15% (included in the 81%) came from the person’s Bishop based on what he knew of the person aided by church membership records. Of the missing 19%, 4% refused to answer, 1% had died or had officially left the Church, and about 14% didn’t respond (the authors likened that category of people who are Mormons, but who don’t self identify as such. The self-identification problem is fairly typical in religious research, but scholars usually forge ahead after acknowledging the limitations. The divorce rates are calculated as number ever-divorced divided by the number ever-married. This percentage was compared to that found for Catholics and Protestants, using prior survey results over several years (this data not originally collected by Heaton and his co-author).

Let me quickly summarize the numbers for male and female divorce rates calculated this way:

Catholic: 19.8/23.1
Liberal Protestant: 24.4/30.8
Conservative Protestant: 27.7/30.9
Mormon (total): 14.3/18.8

Compare those numbers to the 1999 Barna Survey (used here) which calculated its percentage by ever-divorced divided by total number of adults. In this calculation the denominator is relatively larger than ever-married population denominator used in the study above.

Non-denominational 34%
Baptists: 29%
Mainline Protestants: 25%
Mormons: 24%
Catholics: 21%
Lutherans: 21%

As another baseline of comparison, Michael Quinn in Mormon Hierarchy: Extensions of Power page 828 has an entry for 26 JAN 1942 which reads:

first counselor J. Reuben Clark tells reporter from Look magazine: “Our divorces are piling up.” Church Historian’s office in 1968 compiles divorce statistics since 1910 for temple marriages, “church civil” marriages, and “other civil” marriages. Although temple marriages have the lowest divorce rate, in 1910 there was one “temple divorce” (cancellation of sealing) for every 66 temple marriages performed that year, 1:41 in 1915, 1:34 in 1920, 1:27 in 1925, 1:30 in 1930, 1:23 in 1935, 1:27 in 1939, 1:17 in 1945, 1:31 in 1950. 1:30 in 1955, 1:19 in 1960 and 1965. Last rate for temple divorce is almost ten times higher than Utah’s civil divorce rate [a] century earlier.

Finally readers will interested in Tim Heaton’s remarks at the 2002 FAIR conference where he discussed studies that where then more up-to-date and estimated the active Mormon divorce rate using the governmental metric.