pageok
pageok
pageok
Careful With Those Surveys:

A commenter on the attitudes-towards-homosexuality thread writes:

The Pew Research Center did an interesting poll on the subject recently, actually. The number of people who "strongly oppose" gay marriage has sharply declined in the last two years, from 42 percent in February of 2004 to 28 percent in March 2006. The greatest decline is among seniors, Republicans and moderate religious groups. Fully 58 percent of Americans age 65 and older strongly opposed gay marriage in 2004; only 33 percent are strongly opposed now. Seems to support the idea that old dogs can learn new tricks, if people make the case well enough.

Another picks this up:

The Pew study doesn't surprise me. It seems to me that people are more likely to have favorable viewpoints regarding issues associated homosexuality as we as a society talk about it more. Gay marriage or gay parents are no longer little discussed issues shoved in the very far back of the closet. The more those sorts of issues are discussed, the less bizare they sound. It's easy to whip people up into a fervor about something they haven't thought very much about and is never done. It's much harder to freak people out about something when they are informed (regardless of their conclusion) and when they see that other states and countries have done it and the world didn't come screetching to a halt.

To his credit, the first commentator provided a link to the study, and the study paints a rather different picture than the comment suggests. Strong opposition did fall from 42% to 28% between February 2004 and March 2006 -- but only after it had risen from 30% to 42% between July 2003 and February 2004. The change between July 2003 and March 2006 is thus a quite small and statistically insignificant -2%, not -14%. Similar results are present for the subgroups within the sample.

So be careful when reading accounts of surveys. A seemingly striking change may be less "old dogs learn[ing] new tricks" and more the old dogs returning to their old tricks. Perhaps there has been a long-term change in attitudes about same-sex marriage over the years (though then we'd have to ask whether the change stems from individuals' changed views, or people with one set of attitudes dying and being replaced by people with a different set); but the survey that the commenter pointed to didn't really show such a long-term change, and certainly not in the magnitude that the commenter suggested.

Medis:
What will be interesting to see is whether there is any similar bump in this even-numbered year.
6.19.2006 4:08pm
Splunge (mail):
[Alt]hough then we'd have to ask whether the change stems from individuals' changed views, or people with one set of attitudes dying and being replaced by people with a different set

We do? I'd say the latter hypothesis not only smells unpleasantly of social Darwinism but is plain silly to boot.

Why should we entertain the hypothesis that people are just born with unchangeable attitudes towards complex adult issues? Or that they are fixed in these attitudes at a young age? In the first place, this reminds me unpleasantly of the notion that black people are "just born" (or permanently shaped by their environment at a young age) to be musically talented or shiftless, or (John McCain's hilarious implication a while ago) that Mexicans are "just born" to be superior lettuce-pickers, or that girls are "just born" (or trained at a young age to be) uninterested in science or medicine and to want to take care of babies instead.

This quite denies the utility of education, experience and introspection. It's been historically an attractive point of view when one favors a heriditary aristocracy, the idea that people are more or less just born different, and hence some are born better than others and the latter can't change. But I think it strikes the modern American ear unpleasantly like a vestige of a European-style birth-class system we've consciously rejected over and over again.

Secondly, I suggest the hypothesis is just plain silly. The idea that people's attitudes change with age and stage in the life cycle is very well accepted amongst the psychologists, and even (for what it's worth) in folk wisdom. Who doesn't think "young marrieds" have different attitudes about marriage than folks celebrating their golden anniversary? Who is surprised that parents change certain attitudes towards child-rearing when they become grandparents, and in pretty predictable ways? Why would anyone be surprised to learn that people change their attitudes about Social Security taxes the closer they get to retirement themselves?

That we should toss this common-sense wisdom overboard and consider the alternate possibility that people have attitudes determined primarily by the accident of what the social mores were like in the year they were born (or turned 12, or whatever) seems goofy.
6.19.2006 5:43pm
Medis:
Splunge,

If you clicked through, you should know that Professor Volokh was referring to an empirical result. Whether we want to entertain it or not, it does seem like we need to entertain the hypothesis that with respect to this particular issue, and at least in the aggregate, there is limited change within birth cohorts over time.
6.19.2006 6:15pm
Splunge (mail):
Medis, I did read the original post, but I disagree with you. I think Professor Volt is making an interpretation of the data, not noting a fact unequivocally present in the data itself.

It is, furthermore, based on the flimsiest of statistical bases, namely that two numbers in widely-separated polls, with unknown variations in methodology, sampling error, and population statistics happen to be similar. That isn't science, it's numerology, as a priori meaningful as the "observation " that the outcome of the last game the Redskins played during election years predicted the result of the Presidential Election from 1936 to 2004.

The only serious way to study the question at issue, how people's attitudes change over time, is to do a proper longitudinal study in which you ask the same people over and over again how they think, with careful statistical adjustment for how the composition of your study group changes over time, if it does. I think the National Opinion whatever center is just flinging numbers carelessly about, in a way that reminds me of Twain's famous comment that there are three kind of lies: lies, damn lies, and statistics.
6.19.2006 8:23pm
William (mail) (www):
Medis makes an interesting point. Gay marriage was brought to the forefront of national politics in 2004 with the presidential race and several state constitutional referenda banning gay marriage. Perhaps the spike in the polls was a reaction to the prevalence of the issue in that particular year, and the 2002 and 2006 data more accurately reflect Americans views on the issue. How many states have constitutional referenda on gay marriage this year? I know Virginia is (sadly) one of them.
6.19.2006 8:39pm
Medis:
Splunge,

First, of course Professor Volokh is "interpreting" the data (although only suggestively), but you have him offering all sorts of hypotheses about mechanism, which he did not do.

Second, that is a terrible analogy. I agree that this is a highly imprecise "cohort study"--and I don't think Professor Volokh would disagree either--but the Redskins example isn't even the same sort of claim.

Third, your original claim was that we shouldn't even entertain a hypothesis to this effect. But now you are just saying this isn't a "serious" study. I think those are clearly different issues: we obviously have to identify hypotheses worth testing before we can conduct "serious" studies, and thus it would be absurd to say we can't entertain a hypothesis until it has already been the subject of a "serious" study.

Finally, your entire original post was full of conjecture, intuition, and appeals to common sense in support of your propositions. It strikes me as quite odd, therefore, that you are now requiring Professor Volokh to engage in a "serious study" before even suggesting a contrary proposition.
6.19.2006 8:42pm
Clayton E. Cramer (mail) (www):

Medis makes an interesting point. Gay marriage was brought to the forefront of national politics in 2004 with the presidential race and several state constitutional referenda banning gay marriage. Perhaps the spike in the polls was a reaction to the prevalence of the issue in that particular year, and the 2002 and 2006 data more accurately reflect Americans views on the issue.
Why is the 2004 data less accurate than the 2002 and 2006 data? That the issue was brought to the forefront in 2004 doesn't mean that they the extra opposition that year wasn't real. It means that there are some people whose opposition to gay marriage is fired by being reminded of it on a regular basis.
6.20.2006 2:27am
Medis:
Clayton,

Or it may be that they were specifically worried about having gay marriage brought about suddenly by the courts or local officials. And it may be that this was a one-time effect, and now being "reminded" of it will not have such an effect.

That is part of why I think it will be interesting to see what happens this year. There is obviously another ongoing attempt to fire up opposition to gay marriage in preparation for the 2006 elections, and whether it works or not will help us test your hypothesis.
6.20.2006 9:03am