Federal Sex Data Revised

The US Centers for Disease Control and Prevention have changed data in a report issued, originally, in September, 2005. This seems like a highly unusual thing to do, but it is hard to see any political motive for it, as the changes are mostly small and don't tend to make any group look significantly better or worse than before.

I have cited the survey, titled Sexual Behavior and Selected Health Measures: Men and Women 15–44 Years of Age, United States, 2002, numerous times. It is a fascinating look into what sexual behaviors adult Americans actually practice -- it is incredibly interesting to find out what your neighbors are really doing -- and I quote it pretty often to make the important point that anal sex is predominantly a heterosexual activity.

The survey is kind of hard to find (some web-censor software blocks the CDC site), so I keep a copy of it on our web server: HERE, but the official version is HERE, at the Centers for Disease Prevention and Control.

Today I looked at the CDC version. It now has a yellow box at the top of the first page that says:

Data in Tables 8, 10, 12, and Appendix table 4 have been revised. Numbers cited in text on pages 3, 4, and 13 have been revised.

This report has been sitting here on the Internet for a year and a half, and suddenly, apparently within the last month or so since I last looked at it, they have decided to revise their estimates.

Don't you want to know what has changed?

Here we go.

In Table 8, they increased the number of adult Americans who have had same-sex sex but never had sex with anyone of the opposite sex. Yes, the word "sex" is working too hard here. Sorry.

Noticeably, the raw numbers of men and women who have had sex with the same sex, but never with the opposite sex, has just about doubled in the new data. These are hard data to understand; it apparently reports the number of people who have had a category of same-sex experience, of those who have no opposite-sex experience. The number of strictly homosexual men (never had sex with a female) who have had either oral or anal sex (with another man) has jumped by about ten percent, with 25.7 percent now saying they have had anal sex, compared to 15.5 percent in the original, and 26.7 percent now reporting oral sex, compared to the original 16.2. Women who have had sex only with other women jumped from 6.3 to 12.2 percent.

So I guess that's saying that gay people are more exclusively homosexual than previously reported, though the "None" row of this table is hard to conceptualize.

I don't see anything else on that page changing.

In Table 10, the number of lifetime men's female partners has increased slightly: the median number went from 5.4 to 5.6 partners. Because these are reported as percents, it's again a little tricky to figure out where the change was -- if you change one percent, you have to change them all. It appears that the number of men reporting a large number of female partners has been decreased slightly, and the lower reports, percentagewise, have increased. That is, more guys report having had zero, one, two, or three-through-six partners in the revised data, while proportionally fewer report having had seven-through-fourteen or fifteen-or-more. Changes are reflected in every age group, it looks like.

There were also changes in the percentages of "Never married, not cohabitating" males who have had more than two partners in their lifetime, and percentages are shuffled around among the educational categories, but I can't see what the trend is. While medians didn't change, the percentages reported by those with and without military service changes, as well. (FWIW, it looks like military guys have about twice as many female partners as non-military.)

Hispanic percentages went up in almost every category, though their median didn't change. Fewer white guys have had zero partners, and more have had one. The survey changes represent a 20 percent jump in the number of black males who have had a prudent 3-6 partners, with a nearly 25 percent decrease in the 15-or-more category.

Table 12, which is about men only, has a change in the last section, "Number of opposite-sex partners in lifetime," in particular the percentage of guys who say they've never had sex with a female.

           Het   Hom   Bi  Other   NA
Original  74.5   9.7  1.5   7.3   7.1
Revised   73.7  10.1  1.5   7.3   7.4

These numbers are the percent of men reporting having had no opposite-sex partners who say they are heterosexual, homosexual, bisexual, other, or didn't answer. The change shows fewer straight men, and more gay men, reporting they have never had sex with a female.

Interesting: most guys who never have had sex with a woman are straight, by more than a 7-to-1 margin. Yes, this survey is fascinating.

In the Appendix, Table IV has changes, it looks like, in every section. It's hard to say what the trend is. It appears to show more male teenagers having some sexual experience with a female partner, and fewer guys in their early twenties. Fewer men now have had no sexual experience with a female, and more Hispanic and black (but not white) men have had fifteen or more female partners than previously reported.

On page 3, the percentage of Hispanic women 15-44 years of age who have had fifteen or more partners increased from 4 to 4.6 percent.

On page 4, the percent of men 15-44 who had used a condom in their most recent sex "event" was 39, now it's leaped to 40 percent (the category is a little more complex than that, you can find it if you look).

On page 13, the original report states that "In table 10, the median number of partners reported by men 15–44 years of age" is 5.6, but the revised report gives that number as 5.4. This is just a summary of Table 10 results cited above.

OK, those are the differences.

In my experience, if a government survey changes its published results, they usually have to send an SES out to face the TV cameras. They don't like to do it, let's say. It means there was something wrong before, and it makes the agency look bad.

In this case, as I said, I can't see any important political statement that is made by the revisions. It looks like they might have realized they had some weights screwed up or something. I imagine they'll blame it on a "computer glitch."