Tuesday, April 8, 2008

Which Matters More: Race or Wealth?

I've wondered about this question for a while. I was under the impression that a rich white kid outperforms a rich black kid, but that a rich black kid outperforms a poor white kid. And then I heard a throwaway line in a documentary I saw in class last semester that said something to the effect of "if we control for wealth (including home equity) instead of income, the black-white gap disappears."

What? Really? Since then I've been trying to figure out where that information came from and if it were true or not. Finally, yesterday, a friend who's been working on a huge project on the achievement gap forwarded me an article (abstract) from the latest issue of the journal Child Development. In it, the authors use a large dataset to see if wealth mediates the achievement gap. Before I get to the findings, let me stop for a second and answer a question that might be in your head. Why would wealth matter more than income? In this study, Whites earn about twice as much as Blacks, but have about five the times the amount of wealth, on average. Think about it: one family that makes x dollars/year but has no money in the bank and rents an apt. will not have the same lifestyle as another family that makes x dollars/year but owns their own house and has a large portfolio of stocks and bonds.

So, what's the answer? Which one is more important? Well, it looks like . . . neither.

The authors have data on a lot more than just wealth and race, and when they include all their other variables neither race nor wealth are statistically significant predictors of achievement in most of their models (they have both reading and math scores for 3-5 and 6-12 year-olds). The variables that are most frequently statistically significant are parent's occupation and mother's cognitive ability. Depending on the model, parental education, grandparent's education, debt, stock ownership, and a variety of home environment measures are also statistically significant.

So, according to this dataset, there is no difference in achievement between blacks and whites who have parents with the same job, cognitive ability, and some other factors. I e-mailed the author to make sure I had this right before writing about it (I did). I also asked if race were still a statistically significant predictor when only controlling for wealth (I didn't see a table with only these variables in it) and I was told that it was.

So, there appear to be two main lessons out of this paper (for me, anyway).
  1. Race, in and of itself, doesn't seem to determine achievement level. Now, maybe it influences what kind of job somebody can get (for example), and that influences achievement level -- but if a parent is able to get that job then their kids should score about the same as other kids of parents with similar jobs regardless of their race.
  2. Though a better predictor of achievement than income, wealth does not fully mediate the effects of race on children. In other words, white children whose parents own a certain amount of property, stocks, etc. will out achieve (on average, of course) black children whose parents have the same level of wealth.
But -- and this is important -- no one study will ever definitively answer a large question like this one. In the literature review, they cite studies with findings that both agree and disagree with this one. Furthermore, the paper was an analysis of large dataset. These analyses provide compelling information on general trends, but can never answer questions as complicated as this one. Don't get me wrong, I'm glad the authors took the time to do this -- I think it's important -- but race, wealth, and other factors impact each individual differently. It sounds like the data they used are quite good, but about 70% of the variance in achievement is left unexplained in these models. In other words, the paper looks to be pretty solid, but only a body of literature (including both quantitative and qualitative analyses) can provide conclusive evidence.


RDT said...

I have heard that the strongest predictor a student achievement (as measured by standardized test scores) is the parents' educational level. Do you know if the study looked at that?

Corey Bunje Bower said...

It did -- it would take me a lot longer to fully describe the study (at which point it would no longer be a summary). Parental education is a statistically significant predictor of achievement in some of the models. But occupation and cognitive ability of the mother is a stronger predictor. Of course, there's probably a strong relationship between how educated your parents are, what kind of job they have, and how well they score on a test.

Nancy Flanagan said...

Interesting, especially the distinction between wealth and income. It strikes me as a way to attempt to quantify the cultural capital that kids bring to the table (with another quantifier being parent education). Researchers (especially in our "scientifically based" environment) seem to shy away from the sociological or anthropological aspects of how and why kids achieve.

I live in Michigan where autoworkers with HS diplomas often outearn academics and certain professionals (although that is changing, rapidly). Some of them have obtained trappings of "wealth"--second homes, investments. And some of them have used good incomes to make sure their children are well-educated. But not all. I think there are some things that cannot be reliably quantified.

Maybe Bourdieu is right--it's all about the social and cultural capital.