In The Perils of Diversity, Byron Roth challenges the multiculturist argument that minority groups must preserve their cultures to prop up self-esteem by citing research which failed to find group differences.
Let's look at the question with Add Health data. Students were asked how frequently do they feel they are just as good as anyone else. Answers ranged from "never/rarely" (0) to "most/all of the time" (3). Here are means:
Mean self-worth scores
Mexican Americans 1.61*
* significantly lower than whites
While blacks and whites--the two groups usually studied--don't differ significantly, Asians and Mexican Americans have lower averages.
Why? This is a sample of adolescents. Blacks and whites are more likely to be popular. This seems to be due to greater athleticism, extraversion, and more interesting personalities. I also suspect that whites and black males are perceived as more attractive. Self-esteem seems to be, in part, a reflection of reality. A Mexican American is more likely to have low self-worth because he doesn't compare well to others and lacks traits that are valued. It might be the case that whites and blacks get a self-esteem boost in schools with Mex-Am and Asian students. If I find time, I can test the hypothesis.