What is the causal link between parental income and education?

by on March 29, 2012 at 5:54 am in Education | Permalink

…the connection between income and student performance “is no less true in the Age of Obama than it was in the Age of Pericles.” But, he points out, most of the connection is not causal, but due to other factors.  He cites a study by Julia Isaacs and Katherine Magnuson (Brookings Institution, 2011), that examines an array of family characteristics – such as race, mother’s and father’s education, single parent or two-parent family, smoking during pregnancy – on school readiness and achievement.  The Brookings study finds that the distinctive impact of family income is just 6.4 percent of a standard deviation, generally regarded as a small effect.  In addition, Peterson calls attention to earlier research by Susan Mayer, former dean of the Harris School at the University of Chicago, which also found that the direct relationship between family income and education success for children varied between negligible and small.

Responding to Ladd’s claim that the gap in reading achievement between students from families in the lowest and highest income deciles is larger for those born in 2001 than for those born in earlier decades, Peterson points out that the achievement gap between income groups was growing at exactly the same time the federal government was rapidly expanding services to the poor – Medicaid, food stamps, Head Start, housing subsidies, and many other programs.

“A better case can be made that any increase in the achievement gap between high- and low-income groups is more the result of changing family structure than of inadequate medical services or preschool education,” Peterson says.  In 1969, 85 percent of children under the age of 18 were living with two married parents; by 2010, that percentage had declined to 65 percent.  The median income level of a single-parent family is just over $27,000 (using 1992 dollars), compared to more than $61,000 for a two-parent family; and the risk of dropping out of high school increases from 11 percent to 28 percent if a white student comes from a single-parent family instead of a two-parent family.  For blacks, the increment is from 17 percent to 30 percent, and for Hispanics, the risk rises from 25 percent to 49 percent.

That is from Harvard’s Kennedy School, from Paul Peterson.  Here is more, and for the pointer I thank Ilya Novak.

1 dearieme March 29, 2012 at 7:42 am

Nothing you’ve said implies anything about a causal link. It’s all mere correlation, isn’t it?

2 rjs March 29, 2012 at 10:56 am

yes, and less so than teachers or schools…

3 Andrew' March 29, 2012 at 7:48 am

Parents are good at making kids do stuff while they have them.

4 Bob Dobalina March 29, 2012 at 7:50 am

Bordering on crimethink, that paper.

5 Bruce B March 29, 2012 at 8:32 am

Disappointing that Peterson can quote precise statistics as to why the “other side” is wrong (income impact of 0.064 std deviation) but has no such hard numbers to back his argument. Surely the data that generated the family income data would also have information on single vs two parent status?

Not saying I think he’s wrong, he should just hold himself to the same standards he holds those he disagrees with.

6 Joseph Davidson March 29, 2012 at 8:36 am

A good source of dis-aggregated data would be certain immigrant groups who have high intellectual and social capital. The recent immigrant parents may not have financial resources, but their kids do well ( all statistical, not individual )

7 Uninformed Observer March 29, 2012 at 8:59 am

Why do we continue to pretend that the answer is not obvious?

“Gee, parents who display the markers of personal responsibility and hard work tend to have kids who display the markers of personal responsibility and hard work, and the inverse appears also to be true. Now why could that possibly be the case? Must be something to do with [income|race|marital status] or something else that can’t be attributed to the individual.”

8 joan March 29, 2012 at 9:26 am

A mother with two children is “poorer” than a married couple with one child with the same dollar income and a family with two earners is “poorer” than a family with one earner with the same dollar income.

9 Robert March 29, 2012 at 10:35 am

The school district I’m in claims it gets such good test scores because their schools are better.

I think it’s because those affluent parents send their kids to Kuman and other after school private tutoring franchises. these places are packed after school.

10 Andrew' March 29, 2012 at 11:28 am

It wouldn’t be as easy to claim that and free ride if the parents could choose schools.

We can now “choose” to go to the STEM versus artsy elementary school. We can even pay a premium to go out of the district. Something is happening, I’m not sure what, and I’m sure they’ll muck it up.

11 KM March 29, 2012 at 4:02 pm

And where is the evidence that family structure causally has an effect on childrens’ educational results? Most good studies I have seen tend to show that is largely a correlation without causation (and explained by other factors driving family structure).

12 GiT March 29, 2012 at 4:09 pm

“That is from Harvard’s Kennedy School, from Paul Peterson”‘

Sort of.

It’s also from a press release for the Hoover Institute.

13 GiT March 29, 2012 at 4:33 pm

““A better case can be made that any increase in the achievement gap between high- and low-income groups is more the result of changing family structure”

O rly?

Should one give this post a Straussian reading? Exoteric message: It’s all the fault of those single mothers. Esoteric message: Paul Peterson doesn’t tell us anything about causality?

Donna Ginther and Robert Pollak: http://apps.olin.wustl.edu/macarthur/papers/wp-pollak-fs-10-02.pdf

“Those who favor policies that promote marriage often cite stylized facts – – descriptive regressions that control for nothing other than family structure – – while those who oppose such policies respond by citing descriptive regressions that control for variables such as mother’s education and family size. Both sides use descriptive regressions that support their positions as political weapons, but descriptive regressions are only descriptive regressions – -summaries of correlations among endogenous variable”

“The lack of a consensus about the effect of family structure on children’s outcomes is striking. Research shows that living with a single parent or a step parent is correlated with poor outcomes for children. Biblarz and Raftery [1999] show that the correlations between family structure and children’s outcomes diminish as more family background variables are added to the specification. When researchers attempt to address the endogeneity of family structure, the estimated effect of family structure depend on the identification assumptions employed. The most consistent set of results are found when parental death is used as a quasi-natural experiment. These results suggest that family structure, per se, is not causal.”

” Descriptive regressions reveal somewhat different patterns. Descriptive regressions control not only for family structure, but also for variables such as mothers’ education and family size. With these controls, the effect of family structure falls substantially and often loses statistical significance. In particular, the effect of living in a single parent family is no longer statistically significant.”

14 dearieme March 29, 2012 at 5:12 pm

No mention of IQ, either child’s or parents’: you don’t need to be a IQ-worshipper to find that odd.

15 Eva March 29, 2012 at 6:07 pm

I was thinking we need to study the children of lottery winners, but there is already something like that: the literature on unconditional cash transfers and education.

16 Steve Sailer March 29, 2012 at 10:39 pm

The 1979 National Longitudinal Study of Youth, which gave the military’s IQ-like AFQT test to 12,000 15 to 23 year olds in 1980 is now tracking thousands of second generation participants as well as their 47-59 year-old mothers. So, the data is out there is anybody wants to look.

17 DK March 29, 2012 at 11:37 pm

When an IQ is not even mentioned while discussing connection between income and student performance, in a sane world it’s got to be judged as a sort of “academic malpractice” and the authors have to be banned from publishing for a long time. Pathetic.

18 athEIst March 30, 2012 at 4:21 pm

The median income level of a single-parent family is just over $27,000 (using 1992 dollars), compared to more than $61,000 for a two-parent family; .
This is not intuitively obvious, or don’t I just not get it?

19 athEIst March 30, 2012 at 4:23 pm

I mean “or do I just not get it”.

Any guess?

Comments on this entry are closed.

Previous post:

Next post: