In additional analyses, we evaluated whether the effects of these pollutants were modified by participants' characteristics: sex, birth month (January vs. July), and maternal smoking status (nonsmoker vs. smoker). The p-values for interaction less than 0.05 (two-sided) were considered significant.