Social-science research is sometimes criticized because its findings are deemed to be obvious. The implication is that researchers are wasting their time and other people's money. But one reason to do research is to determine which of many "obvious" notions are true and which are based on misconceptions. After all, isn't it obvious that the sun rotates around the earth?
Recently, I encountered a stunning example of social-science research that challenges strongly held beliefs. It seems obvious that an unmarried teenage girl who has a baby puts herself on the economic slow track. And, until recently, research seemed to support this idea. Teenage mothers were shown to earn much less money, drop out of school sooner, be less likely to be married, and be far more likely to depend on welfare than those who delay childbearing until at least age 20. Teen childbearing certainly looks like an economic disaster.
But looks can be deceiving. New research by Arline T. Geronimus of the University of Michigan and Sanders D. Korenman of Princeton University suggests that teenage motherhood may be less harmful than is commonly supposed. Why? Because economic failure might be a cause, rather than an effect, of teen childbearing. Teenage mothers are not selected for motherhood at random; they tend to come from the ranks of the disadvantaged. Compared with other women, teenage mothers are more likely to have come from single-parent families and to have had poor, uneducated parents. Since these attributes are leading indicators of economic failure, some teenage girls may look at their poor prospects in life and opt for early motherhood.
SIBLING STUDY. The notion that teenage motherhood might be a symptom of poverty rather than a cause is perhaps not as stunning as it first appears. After all, we know that people on diets are, on average, heavier than people who are not on diets; but we do not therefore conclude that dieting makes you obese.
When it comes to teen pregnancy, distinguishing cause from effect is difficult. In a laboratory, we would simply raise a bunch of gerbils under identical conditions, impregnate half of them as "teenagers," and see if they underperformed the others. But people are not raised in laboratories. So how can we separate the effects of teenage motherhood from the effects of poor family background?
Geronimus and Korenman came up with an ingenious solution. Suppose we compare teenage mothers with their sisters who delay childbearing until at
least age 20. That would isolate the pure effect of teen childbearing on economic success.
A nice idea. The hard part is finding suitable data. First, you need a data source that identifies sisters and tracks them over some years. Then, you need to find examples of teenage mothers whose sisters were not also teenage mothers. Such cases are relatively rare. In the main data source that Geronimus and Korenman use, there are only 51 such teenage mothers; an alternative source has 52, a regrettably slender data base. Nonetheless, the findings are worthy of note.
DROPOUT DIFFERENTIAL. How much does teen childbearing lower subsequent family income? Among all the young women in the main sample that Geronimus and Korenman analyze, average family income is 33% lower among those who were teenage mothers. A difference that large sounds like an economic catastrophe. But remember that teenage mothers often come from disadvantageous circumstances to begin with, so many were destined for failure anyway. If you control for socioeconomic background by looking only at the 51 families in which at least one sister was a teenage mother and at least one was not, the income gap narrows to just 16%.
Do teenage mothers drop out of school sooner? Looking at all the sisters in the Geronimus-Korenman sample, we find that 65% of the teenage mothers graduated from high school vs. 89% of those who first gave birth after their teens. That's a big difference. But once again, we must remember that many teenage mothers would have dropped out of school anyway. If we compare the 51 teenage mothers to their sisters who delayed motherhood, the difference in graduation rates disappears entirely. And the same is true of being on welfare. Among all sisters, 21% of teenage mothers are on welfare vs. only 5% of those who were not mothers--a huge difference. But among sisters who made different choices about teen childbearing, there is no difference at all.
The suggestion, then, is that much of the economic failure commonly associated with teenage motherhood really stems from disadvantaged social backgrounds. Having a baby per se may not be much of an economic handicap after all. This conclusion is so contrary to popular belief that Geronimus and Korenman tested it on a second source of data. Doing so yielded similar results for family income and welfare but somewhat different results for high-school graduation: Teenage mothers were less likely than their sisters to have graduated from high school.
All this suggests that teen childbearing causes less poverty than is commonly thought. What is obvious is not always true.