We have taken a number of steps to minimize the impact of the study's primary limitations, the likely bias in our prevalence estimates resulting from differential attrition, and the reliance on maternal questionnaires rather than direct clinical assessments. With regard to attrition bias, it is important to note that any cross-sectional study in a school-age population inherently represents a biased sample of children compared with their original “birth cohort,” although in most cases the attrition rate is unmeasured. Here, because ALSPAC is a longitudinal study, we were able to examine the factors related to attrition and take note of potential bias, turning an inherent weakness into a strength of the study. Although the identified attrition bias might reduce the generalizability of our findings to some degree, the factors that we identified as being associated with attrition (female gender, nonwhite ethnicity, lower maternal age, and lower socio-economic status) are likely to predict nonparticipation in most epidemiologic studies. We may also have missed some subjects who had chronic tics that abated before age 13, the time point for which we have the
