maradydd: (Default)
maradydd ([personal profile] maradydd) wrote2008-11-03 04:01 pm

Because I hate bad statistics

The following is a response to this post about California's Proposition 8. I left it as a comment there, but comments are moderated, and somehow I don't think it will get posted. Thus, y'all get to read it here.

Amy writes:
"After legalizing same-sex marriage 5 short years ago HIV/AIDS has increased in Massachusetts with more than 40,000 being infected each year."
I don't know what Amy's source on this figure is, but I did some research, and this claim is not only wrong, it's wrong on several orders. The first same-sex marriage in Massachusetts was performed on May 17, 2004. Since then, the prevalence of HIV/AIDS -- "prevalence" meaning "how many people have it" -- has increased, but the rate of increase has fallen off sharply.

First of all, according to the Massachusetts Office of Health and Human Services' epidemiology department, whose 2007 report you can read for free, as of November 1, 2007 there were only 16,866 people known to be living with HIV/AIDS in Massachusetts.

How could there be 40,000 new cases a year if the total number of cases in the state is less than half that?

The report also examines the trends in HIV infection. As you can see in the chart (the right-hand one on the first page), the number of newly diagnosed HIV infections dropped sharply between 2004 and 2005 and again in 2006. If you look at the first page of the data tables, you'll see that in 2003 the total number of cases was 14,992 and in 2004 it was 15,633. That's an increase of 641 cases. In 2005 the number was 16,217 -- an increase of 584 cases. In 2006 the number was 16,621 -- an increase of 404 cases. For the first ten months of 2007 it was 16,866 -- an increase of 245 cases.

What we can see from this is that the rate of new infections in Massachusetts has not only fallen since the introduction of gay marriage, it has fallen more quickly. 57 fewer people got infected in 2004 than in 2003. 180 fewer people got infected in 2005 than in 2004. And 159 fewer people got infected in the first ten months of 2007 than in all of 2005. If we project that trend out to the end of 2007, that would be 190 fewer new infections.

I'm sorry, Amy, but your argument doesn't hold up.

[identity profile] kragen.livejournal.com 2008-11-04 09:25 pm (UTC)(link)
Bravo for bringing evidence into this discussion.

She didn't claim the rate of new infections had risen; she said "HIV/AIDS", which presumably means "HIV/AIDS prevalence", had risen. She is correct about that, as you confirmed. That rise is hardly surprising, though; as far as I know, HIV prevalence has risen every year in every state since HIV was discovered, because there's no cure, it takes a long time to kill you, it's a contagious disease that currently affects a small fraction of the population, and humans have no natural immunity. This means that the increase is not due to same-sex marriage in Massachusetts, and [livejournal.com profile] maradydd's analysis, looking at the rate of change in the rate of change, is the right approach.

Perhaps Amy is parroting someone who actually knew what they were talking about, and left out qualifiers like "40 000 new cases per year in the US"?

For what it's worth, the number of new infections is not the same as the change in the number of cases. The number of cases can increase because people with HIV immigrate from other states (and actually you would expect the same-sex marriage law to encourage this), and it can decrease, or increase by a smaller amount, when people die or emigrate to other states.

A plausible case for legal same-sex marriage increasing the incidence of domestic violence can be constructed as follows:


  • Legal same-sex marriages will increase the number of same-sex partnerships by making them more attractive to people who might otherwise end up single or with an opposite-sex partner. (Seems likely; in fact, you might even say reducing the number of same-sex partnerships busted up by societal pressures is kind of the point of same-sex marriage!)
  • Male-male partnerships have higher rates of domestic violence than opposite-sex partnerships. (I think this is empirically supported but I forget where to find the evidence. It is hardly surprising, though; men are, on average, far more aggressive than women.)
  • The people who commit that domestic violence in male-male partnerships would commit less of it if they were single or with an opposite-sex partner. (Plausible (for example, maybe fighting back increases the amount of domestic violence, and women are less likely to fight back), but I would want to see evidence to support it.)
  • Either female-female partnerships don't have lower rates of domestic violence than opposite-sex partnerships, or the rate is lower by less than the contribution from male-male partnerships, or the increase in the number of female-female partnerships will be much smaller than the increase in the number of male-male partnerships. (All three of these are plausible.)


Rather than speculating among like-minded people about how those who disagree with us are "noxious harridans", it would probably be more productive to sit down with them and try to empathize with them. Even when they're wrong, and even when they're not 100% genuine. In the long run, that's the only way to resolve contentious issues for real. Getting angry at strawman caricatures we construct in our own minds will just deepen the divisions in our society.

Orson Scott Card wrote a pretty good (clear, logical, somewhat misinformed, based on premises I disagree with) explanation of why he thought gay marriage was a threat to marriage in the Mormon Times: State job is not to redefine marriage, 2008-07-24.

See also my related comment on another friend's blog.