Search Results

The probability of treatment increases with income in the U.S., but not in Canada. However, in models of outcomes, interactions between income and ADHD scores are statistically insignificant in the U.S. (except for delinquency), while in Canada these interactions indicate that higher income is protective. The U.S. results are consistent with a growing psychological literature which suggests that conventional treatments for ADHD improve behavior, but have inconsistent effects on cognitive performance.

We use data from the Canadian National Longitudinal Survey of Children and Youth (NLSCY) and from the American NLSY. The NLSCY is a national longitudinal data set which surveyed children ages 0-11 and their families beginning in 1994.8 Follow up surveys were conducted in 1996 and 1998. The initial sample consisted of approximately 22,831 children in 1994. We restrict our sample to those children who were between the ages of 4 and 11 in 1994, and who were surveyed in both 1994 and 1998. We keep only those children who were given the hyperactivity screener in 1994 which yields a sample of just under 4000 children. For our analyses that use math and reading test scores we have a smaller sample (not all children's test scores were recorded and we discuss this further below) of approximately 2200. We use the NLSCY data to ask how hyperactivity in 1994 affects treatment in 1994 and outcomes in 1998.

Also: The NBER Working Paper No. 10435 was updated in July 2004: http://www.econ.ucla.edu/people/papers/currie/more/mental.pdf

We find that ADHD has large negative effects on test scores and schooling attainment and the effects are much worse than those of physical health problems. The results are qualitatively similar in the U.S. and Canada, and are robust to many changes in specification. The test scores of higher income children suffer as much from ADHD as those of lower income children, though high income children are less likely to be retained in grade. Surprisingly, there appears to be little effect of income on the probability of treatment conditional on hyperactivity scores. A third finding is that even children with relatively low levels of symptoms suffer negative effects. The severity of the effects and the pervasiveness of the symptoms suggest that efforts to find better ways to teach the relatively small number of children diagnosed with ADHD could have a larger payoff in terms of improving the academic outcomes of large numbers of children with milder symptoms.