Freakonomics

In the past few decades, parenting has become its own science. There are “parenting experts” who publish books on the proper way to raise a child. Countless sociological and psychological studies are being conducted about the proper way to breast feed, the proper way for children to sleep, the proper way to punish children, etc. Like most so-called experts, parenting experts are good at sounding sure of themselves, even if their information is questionable. And like all experts, parenting experts are adept at inspiring fear in their audiences of parents—the fear of raising bad children.

One reason that parents are so easily convinced by parenting experts is that parents—and, in fact, all human beings—are bad at assessing risk. There are certain risks that scare people into changing their behavior—but these changes in behavior are often out of proportion with the risk itself. For example, one case of mad-cow disease in New Jersey prompted huge numbers of Americans to stop eating beef altogether. On average, people are far more frightened of planes than cars, even though cars are responsible for many more fatalities than planes. If one accounts for likelihood of death in a car versus likelihood of death in a plane, assuming equal time spent in both vehicles, then the overall likelihood of death is about the same.

Why are people frightened? One persuasive theory about fear is that people tend to be frightened of things that pose an immediate threat, rather than a far-off danger. For example, Congress is more likely to pass a bill fighting terrorism than a bill fighting heart disease, even though heart disease kills far more people every year than terrorism. Heart disease is a far-off problem; terrorism, according to the authors, is “happening now.”

The authors return to the question of parenting. When parents try to make their children safer, it usually involves buying some new product—a product which won’t necessarily protect the child at all. For example, the car seat is often touted as a vital way to protect children in car crashes. In reality, though, the real benefit of putting a child in a car seat is that the child sits in the back seat of the car, rather than riding shotgun; the car seat itself doesn’t do much to save the child’s life.

Another important aspect of the parenting debate is the question, “how much do parents really matter?” On one hand, it seems clear that bad parenting can play a major role in determining a child’s future—this is why, as described previously, unwanted children born to parents who might otherwise have gotten an abortion are more likely to commit crimes as adults. But on the other hand, it’s not clear how much good parents can prepare their children for success in adulthood. Numerous studies of twins who were separated at birth suggest that genetics is responsible for about fifty percent of a child’s personality and abilities. Further studies suggest that parent nurturing accounts for a surprisingly small amount of a child’s development. For example, the Colorado Adoption Project, a study that followed the lives of 245 infants put up for adoption, found no correlation between the personality of the child and the personality of the child’s adopted parents. There’s still a lot of controversy about the role of parenting in a child’s development, with notable proponents on both sides of the debate.

The authors ask us to consider two hypothetical children, one white, the other black. The white child is raised in Chicago by parents who spend a lot of time with him, reading with him and taking him to museums. When the white student proves himself to be good at math, his parents are proud. The black child is born in Florida, and his mother leaves him when he’s two years old. His father, who raises him, is an alcoholic, and sometimes beats him. The black child grows up selling drugs. It seems pretty clear that the white child is likely to have a successful life, while the black child is less likely. The question then becomes: to what extent can we attribute the two children’s situations to genetics, and to what extent can we attribute their situations to nurture?

While this chapter will not attempt a comprehensive theory of child rearing, it will try to measure the role of parenting in a child’s success. A good place to start is academic performance, often taken as a benchmark of a child’s talent, intelligence, and hard work. In the Chicago Public School (CPS) system in the 1980s, it was mandated that any incoming high school freshman could apply to attend any high school in his or her district. This created a chaotic situation in which hundreds of thousands of teenagers were trying to get into the schools that were perceived as being the best in the district, submitting their test results, grades, etc. The only way to be fair was to use a lottery system for the students who applied to schools with more applicants than availabilities. As a result of the lottery system, there were thousands of students with identical test scores and grades sent to different schools. Thus, the CPS affair is a great opportunity to test the causational relationship between high school education and academic success.

The CPS data leads to one conclusion: school choice barely matters at all in determining one’s academic success. When academically identical students (i.e., students who applied to attend an elite high school and had the same grades and test scores) were sent to different high schools, these students tended to have the same likelihood of graduating their high school and passing their federal-administered tests. Furthermore, students who claimed the right to attend another school in their school district weren’t any more likely to graduate or pass their federal tests than academically identical students who didn’t claim this right. In other words, a high school education seemed not to provide a measureable academic benefit for students.

However, there was one group of Chicago students who saw a dramatic change in their academic success as a result of the school choice laws: students of technical schools. Students who opted to transfer from traditional high schools to specialized trade schools tended to do better in their new academic environments. But apart from this small subgroup of students, the CPS school choice law seemed to have little to no success in improving students’ quality of education.

For a long time, studies have shown an income gap between black and white adults. But when one takes into account the differences between the test scores that black and white students achieved when they were in the 8th grad, the income gap virtually disappears. In other words, it would seem, educators can reduce the adult income gap by making sure that black middle school students test at the same level as white middle school students.

Why is there a “testing gap” between white and black middle school students in the first place? There have been many theories: poverty, genetics, racial bias, etc. Some sociologists, such as the Harvard professor Roland Fryer, have argued that there is an unfortunately strong social incentive for black students to do poorly in school, since a black student who does well academically runs the risk of being mocked by his peers for “being too white” or “selling out.”

In the 1990s, the Department of Education (DoE) undertook a study of childhood development from kindergarten to the fifth grade; this study has proven to have important results for anyone seeking to understand the testing gap between white and black students. The study measured the students’ academic performance and compared this data with such other factors as race, family structure, socioeconomic status, etc. In order to understand this monumental study, sociologists and economists have used regression analysis. Regression analysis seeks to isolate the relationship between certain specific factors and other factors: for instance, the relationship between a child’s third-grade math scores and the child’s parents’ level of education. By itself, regression analysis cannot prove causation; it can only show correlation. Interpreting this correlative data, however, sociologists can attempt to prove causation.

The Department of Education’s study from the 1990s yields many important results. First, the black-white testing gap disappears when economists control for factors like income level, parents’ educational level, and mother’s age at the birth of her first child. (When the authors talk about “controlling” for certain factors, they mean that they eliminate the influence of these other factors. To control for the influence of parents’ education on white and black students, economists could focus their attention on students whose parents have identical levels of education.) These results are encouraging because they mathematically refute the racist notion that black students are inherently worse than their white counterparts: on the contrary, they show that black students underperform because of environmental factors—factors than can be improved.

The Department of Education study suggests some ways for the federal government to reduce the black-white testing gap. However, it’s also a discouraging study, because it emphasizes the extent of the problem. There is an enormous disparity between the quality of different elementary schools in the U.S., and the quality of one’s elementary school education, in contrast to one’s high school education, would seem to have a dramatic influence on one’s academic success later on.

The Department of Education study isolated sixteen distinct factors that, one might think, play a major role in a child’s development. Eight of these sixteen factors have been shown to play a major role in the child’s development: 1) The child has highly educated parents; 2) The child’s parents have high socioeconomic status; 3) The child’s mother is thirty or older at the time of her first child’s birth; 4) The child had low birthweight; 5) The child’s parents speak English in the home; 6) The child is adopted; 7) The child’s parents are involved in the PTA (Parent Teacher Association); and 8) The child has many books in his house. The study also identified eight factors that are, somewhat surprisingly, not correlated with the child’s development: 1) The child’s family is intact; 2) The child’s parents recently moved into a better neighborhood; 3) The child’s mother didn't work between birth and kindergarten; 4) The child attended Head Start; 5) The child’s parents regularly take him to museums; 6) The child is regularly spanked; 7) The child frequently watches television; and 8) The child’s parents read to him nearly every day. The authors will now go through these sixteen factors, two at a time.

It matters that the child has educated parents, because families with lots of schooling tend to value education. Also, education correlates with IQ, and IQ is strongly hereditary. However, it doesn’t seem to matter greatly if the child’s family is intact or not. We’ve already seen that family structure (number of siblings, whether or not both parents are alive, etc.) seems to have little impact on a child’s personality, so perhaps it makes sense that family structure has little impact on academic success, either.

It matters if a child’s parents have high socioeconomic status, because status correlates strongly with education and IQ, and intelligent parents tend to have intelligent children. However, moving to a better neighborhood doesn’t necessarily improve a child’s academic success. Nicer houses don’t “improve math or reading scores any more than nicer sneakers make you jump higher.”

It matters if a child’s mother was thirty or older at the time of her first child’s birth, because women who are older when they have their children are likely to have had some advanced education or had significant career experience; therefore, she’ll want her children to have the same academic that she had. However, it doesn’t seem to matter if a mother stays home from work until her child is in kindergarten. The authors offer no explanation for this fact; it's just what the Department of Education data suggests.

It matters if the child has a low birthweight, perhaps because being born prematurely foreshadows poor parenting (since often, babies are born prematurely because the mother drinks or smokes). However, there seems to be no correlation between academic success and attending Head Start, the federal preschool program. Many of the people who work for Head Start preschools don’t have college degrees of any kind, and the job pays poorly. Therefore, Head Start preschool programs are unlikely to attract talented teachers, or correlate with student success in the long run.

It matters that the child’s parents speak English at home. This isn’t surprising, since language ability improves the more one uses the language. However, museum visits don’t correlate with academic success at all.

It matters if a child is adopted; indeed, there is a strong negative correlation between adoption and school test scores. In part, this might be because on average, mothers who offer up their children for adoption have lower IQs than mothers who keep their children. (The authors acknowledge that this might be a distasteful line of thinking.) However, there’s no correlation between spanking a child and its academic success.

It matters if a child’s parents are involved in the PTA (Parent Teacher Association). This is probably because the kinds of parents who attend PTA meetings tend to be educated and therefore motivated to help their children succeed. However, there is no proven correlation between TV watching and academic performance. There’s a strong bias against television when it comes to academic issues—and yet there are plenty of cases when watching TV can be educational.

Finally, it matters that the child has many books in his or her home. However, there is no proven correlation between actually reading to a child every day and the child’s academic success. This seems exceptionally strange, since one would imagine that that reason why books correlate with academic success is, at least in part, because reading books makes you smarter. One possible explanation is that families with lots of books in the house tend to be well educated, hard working, and have higher IQs.

The authors return to the list of the sixteen factors that do and don’t correlate with academic performance. One important thing to note is that the eight factors that do correlate tend to describe things that parents are, (hard-working educated, career-oriented), while the eight factors that don’t correlate tend to describe things that parents do (read to children, go to museums, use corporal punishment).

The authors now return to the nature-nurture debate. The Department of Education study could support the conclusion that parents’ genetic makeup makes a far bigger impact on a child’s development than any specific things that the parents do with their children. An overbearing parent who thinks that he can spank or teach his child into academic success is a little like a foolish politican who thinks he can use money to buy an election—as we already saw, money correlates with electoral success, but it can’t really change whether or not people like that politician.

The Department of Education’s study isn’t the only study of parental influence on child development. For example, one study analyzed adopted children in the U.S. and Britain. The study found that parents who adopt children tend to be smarter, better educated, and more highly paid than the child’s biological parents. While the foster parents’ education and money seemed to have little influence on the child’s early academic success, the study concluded that adoptees were more likely to attend college and get a well-paying job later in life. So adoptive parents, we might conclude, do play an important role in their adopted children’s long-term development.