GSMS began in 1993 with most recent data collected in 2005. Each cohort was observed annually up to age 16 with the exception that the youngest cohort was not assessed at age 13 due to funding constraints. All participants were also assessed at 19 and 21. Each study member aged 9 at initial assessment has therefore contributed up to 9 observations to the dataset; the age 11 cohort have contributed up to 8 observations and the age 13 cohort have contributed up to 6 observations. The study contains 8806 observations with 82% mean individual wave participation rate (range 94%-74%). Interviews were conducted in the home or a convenient location and timed, as far as possible, to coincide with the participants’ birthdays. There is little evidence that symptom prevalence was influenced by assessment wave, cohort or differential drop out (Costello et al., 2003). Parents and children provided signed informed consent/assent and the study received ethical approval from the Duke Institutional Review Board.
