General Social Survey participants were asked the following: "When you think of social and political issues, do you think of yourself mainly as a member of a particular ethnic, racial, or nationality group or do you think of yourself mainly as just an American?"
I consider this to be a key indicator of assimilation. If you think of yourself as something other than American when considering issues important to the country, you are not fully an American. You are at least partly something else.
So what predicts identifying as something else? I conducted a logistic regression analysis in order to answer this question. This technique tells you what matters after you have adjusted for the influence of other variables. Here are the coefficients for variables that might matter:
Factors predicting ethnocentricity
Church attendance .06
City size .00
Some of these are expected: immigrants and non-whites are less American. These are the strongest predictors in the models. But some factors might be a bit surprising. Older people are less ethnocentric. Gender doesn't matter. Education (the 3rd strongest effect) encourages ethnocentricity. It is the opposite of an assimilator. Church attendance is unimportant. So are liberalism and the size of the place where you live.
So the picture that emerges for the "partial American" is a young, educated, non-white immigrant. Not surprising. And the profile of a real American is a less educated, older, white native. God bless 'em.
UPDATE: I looked at region of the country, but nothing jumped out.