Friday, May 05, 2006

Why you gotta be so white?: Out-group Favoritism among Minority Groups


Everyone at Harvard probably knows one Asian kid who is more "white" than Bryce: government concentrator, preppy, has predominantly white friends… He would be a prime example of someone who exhibits out-group favoritism by "acting white".

Minority groups tend to attempt to integrate with the dominate racial group (white) or to disassociate with one's own minority group when incentives to do so are high and the costs of not doing so are also high. Incentives: the white racial group is observed to be the dominant social group, and members of the dominates social group are assumed to be "higher quality" people due to their being in the dominant social group. Thus, friendship with members of the dominate group increases the social capital of members of minority groups by more than a similar friendship with a member of the minority group. Thus members of the minority group have incentives to seek out friendships with the dominate group. Indeed, if the returns to ties to the dominate group are high, we expect that minority group members might try and strengthen their connections with the majority group by emphasizing their similarity with the dominate group and de-emphasizing, or be hostile to, things which suggest differences, like speaking their native language. On the flip side, the costs of not integrating and being socially marginalized can be high. Thus, we would expect that being brought up in a suburban/rural area, where residents are mostly of the dominate racial group, incentives are high for someone of the minority race to try to integrate into the dominant group by rejecting to speak their native language.

Hypothesis: A minority brought up in a rural area may have a greater possibility of rejecting to speak their native language.

Data/Empirical Strategy

My data will come from the 2001-2002 California Workforce Survey, a telephone survey collected by the Survey Research Center at the University of California, Berkeley. I will generate a variable including all those who are Asian and who spoke English at home. I will regress that variable on whether or not they lived in a rural area. To validate my hypothesis, there should be a positive correlation.

Alternatively, I am hoping to get data on the Prefrosh Minority program, and look at the effect of hometown characteristics on whether or not a minority chose to host another minority.

The control variables I can think of using are gender, age, and income.


The most important problem is simultaneity. Because this will be a correlation, a positive correlation can very well be because those who live in rural areas are those who have made a decision to assimilate and thus chose to live in rural areas. Then naturally they will not speak the native language at home.

Also, there are a lot of cases when it is the parents who decide whether their kids learned their native language as a kid. Asian languages are very hard to learn, thus even if someone wants to pick it up when they are older, it may be so hard that they give up easily. Thus, if the parents wanted to train their kid to speak English, they will speak the native language at home, and thus that will not have been an effect of living in a rural area, but an effect of their parents' wishes to assimilate.


The pre-frosh minority program idea is ok, but I have no idea how to get the data.

In general, I like the hypothesis. You have a nice clear, economic hypothesis.

The trick, as usual, is being able to test it convincingly. Ideally, we would find some exogenous shock to the costs or benefits of associating with the minority group (or majority group) and use this. The rural idea is ok, but you need to think hard about exactly what your results do and do not show, and discuss this in your paper.
I think your paper topic is very interesting. Being a "coconut" (brown on the outside, white on the inside) and growing up in the suburbs this is definitly something I can relate to. I do have a comple of concerns however. I think that you are understimating the effects that parents have on influencing whether or not a child speaks a language. I think the location (rural vs. urban) has very little influence in determining at a young age whether children choose to speak a language. Rather, the parents decision to speak their foreign language around the child probably has a much greater effect. One suggestion I have to overcome this is to test the differences in language learning between rural and urban kids. For example, we now focus on children who did not learn the language as young children but choose to learn the language later in life and analyze if the parents choice of location influenced this factor. This eliminates a large part of the early parental influence and yeilds a far more interesting result in my opinion.

I question your assumption that in suburban/rural areas, residents are of the dominate racial group. This should not be assumed and can actually be controlled for. If you can find data that would give you some community fixed effects on race, this instument would improve you esimates greatly. Even if your assumption is correct, this will help to capture some of the variability in the independent variable.
Alex makes a good point. The addhealth data contextual file includes information on pct white and pct black (although I am not sure if it has anything on other groups), but you could at least use the pct white variable to more precisely get at what you are interested in.

Further, I guess you could use some of the network variables which look at diversity in the schools' social network. To more precisely proxy for what you are looking for.
I think Varun makes a very valid point, but his suggestion is probably difficult for you to use since you're limited by the data you're using. Kids usually are not given the choice to speak or not speak a native language. I would think that that decision would be made by parents long before a child can speak anything. But, I don't think that necessarily defeats your hypothesis because assuming that a parent has good intentions for his or her child, then that parent would face many of the same incentives to prevent the child from speaking english. Thus, if the child speaks less of his or her native language at home, it still shows a desire of groups to integrate (the desire of the parent and not the child)
I think Varun makes a good point in emphasizing that parents generally influence their children's ability and desire to speak their native language more than the neighborhood in which they live. I'm not sure what the data look like, but there may be information about the parents that would enable you to control for this.
I also believe that Alex is correct in questioning whether the residents in suburban/rural areas are necessarily white. Perhaps you can use this to your advantage. Depending on the data available, maybe you can compare the differences between minority children who grow up in predominantly white communities versus those who grew up in a community that was predominantly of their own race. This might allow you to more easily control for parents' effects on their children.
If you can get the prefrosh data I think that would be awesome. I don't know if you can base turning away from ones racial identity solely on one's choice to not speak their native tongue. There may be other variables influencing this decision, like comfort with the language. I premdominantly speak English because I am much more comfortable with it than Polish because I live in America, and I don't think this has anything to do with the fact that I grew up in a certain area, particularly since I grew up in the city. I don't know is this is included in the data, but controlling for comfort with the language, or how advanced one is may make the results more definitive. I do think it is a very interesting idea, and like Bryce said, a very clear hypothesis. I don't know how possible this is, but it may also be interesting to look at what organizations minorities belong to in correlation to where they came from. For example, if they join a cultural group and correlate this with rural or urban upbringing. This data is probably much less available though and so it might be too late to look at this. Good luck with the paper!

I feel like you should try to find data on how old the kids were when they moved to the US. That very obviously correlates with how much they identify with American language and culture versus their own. Growing up in Houston, there were tons of asians at my school, but we all tended to reject speaking Chinese. We were all raised in America from the age of 2 or 3, so we identified more with American culture. When I moved to NYC for high school, I found that more people were comfortable speaking their "native" language. However, I think this has a lot to do with how your parents raised you and how important of a value your identity and culture is in your family. For example, I have noticed that Korean-Americans tend to speak Korean a lot more than other Asians, even if they were entirely raised in America. This difference seems to be an issue of cultural pride and identification. You should probably look into finding data for and controlling for cultural values, parental influence, identity, and age raised in America.
Post a Comment

Subscribe to Post Comments [Atom]

<< Home

This page is powered by Blogger. Isn't yours?

Subscribe to Posts [Atom]