It started to resemble junk science in the worst way once the conclusions began to surface.

Second, the write up of the study is terrible. It’s just lists one finding after another, rather summing the key findings. So whoever released this needs to get some writing lessons.

Third, the reason he is generalizing– is because the study is based on a random sample survey. The upside of random sample surveys is that we can make some general statements about groups–hence the terms more likely or less likely. The downside of random sample surveys is that they are almost always geared toward the “dominant population.” In other words, if you are studying interracial families, relationships, or people, they do not have any of the questions you mentioned because the studies are looking at basic demographic characteristics that can fit neatly into survey boxes (The larger data set he is using was probably about health or education). The researcher didn’t gather the data himself; he looked at this random sample survey and tried to do the best he could do. That is a big limitation with sociology, demography, and almost all of the random sample surveys. If we want to know about identity as a process, surveys are completely useless, but if we want to know if biracial people are more likely to identify as biracial they live in the southeast, Midwest, etc, then this is the only type of data that can answer those questions.

Just a few points…..I know it’s much more about methodology, but it is a huge problem that sociologists looking at IR families and people face.

You really nailed it there, Jen. I don’t see how any of these data are really that helpful, enlightenining, or relevant, and I’d much rather read a few dozen interviews. Parents of multiracial children could use the reasons and rationale these other parents provide to inform their own discussion and decision on how to identify their child (before s/he can self-identify, of course.) And what a great resource that would be for the ND workshops!