Levitt makes an argument that we do not use data enough to come to the conclusions that we come to. We rely on our pre-determined beliefs and biases to draw conclusions, and rarely seek out objective information. Nowhere is this more apparent than in the management of offspring: parenting.
In a house that contains both a gun and a swimming pool, children are roughly 100 times more likely to die in the swimming pool than at the hands of a gun. But our beliefs about the dangers of guns ensures the focus is on the gun rather than the more apparent danger, drowning.
Levitt uses economic tools (namely, regression analysis) to figure out the role of various parent initiatives. A regression analysis is a statistical tool that allows the scientist to control several variables from within a mountain of data, thus allowing us to determine which variables have an effect on the output.
From this analysis, Levitt was able to determine that genetics plays a much larger role than it is given credit for. Furthermore, when holding all other variables constant (e.g. incomes, school quality etc.), a child's race does not determine its financial future. Specifically, these are the factors that are strongly correlated to a child's test score performance in the US:
1) Educated parents
2) Parents have high socioeconomic status
3) Mother was 30 or older at the time of her child's birth
4) Low birth weight (negative factor)
5) Parents speak English in the home
6) Child is adopted (negative factor)
7) Parents are involved with the PTA
8) Many books in the home
According to the data, the following factors do not matter:
1) Family is intact
2) Family recently moved to a better neighbourhood
3) Mother didn't work between child's birth and kindergarten
4) Attended Head Start
5) Is regularly taken to museums
6) Is spanked regularly
7) Watches television frequently
8) Is read to frequently
It's important to note that a regression analysis can only show correlation, not causation. That is, the first eight factors above may not cause higher test scores, as there could be another factor that causes both one of these eight factors and the higher test scores to occur simultaneously, or the causation could be reversed (e.g. higher test scores could cause a higher interest in books, leading to more books in the house).