A: There are a lot of understudied populations out in the world. For ADHD, we now know that like just about every other neurodevelopmental disorder that has its origins in childhood, boys outnumber girls. That's true of autism, true of learning problems, true of conduct problems, true of ADHD. But in clinics, there may be 10 to one boys to girls. Why? Boys get referred at higher rates because they're more disruptive and because we think stereotypically that only boys can have these conditions. In the real world, it still may be three to one boys to girls, but if we only relied on clinical samples out there in the community, we'd be misled. So in some of my original studies of ADHD we got sometimes a 15 to one ratio of boys to girls. There would never be enough girls to really justify having them in the summer camp study because they'd get picked on by 50 or so boys.

So one day my team and I realized that we simply didn't know enough about girls with ADHD. We decided rather than just balancing half and half, we'd just do a study of girls in their own right. We said, rather than automatically jumping to, what's the parametric contrast of males and females with this condition, we decided to study a big sample of girls, as representative as possible, and find out what the within gender processes are so we can generate hypotheses for later studies of sex differences.

A: After getting the grant, we then had the task of recruiting a sample of well over 200 girls, with and without ADHD. We did not want to do a straight epidemiological survey. We wanted to conduct the kinds of research summer camps I'd successfully done with boys for many years where we could have classrooms, playgrounds, peer sociometrics, neuropsych testing, and observations at lunch. We needed to have naturalistic, ecologically valid settings so we could look at natural social behavior, not just what happens in a clinic or in a lab. To do that would require sampling from clinical settings and from the community, not just a straight epidemiological representative sampling.

So we used a broad recruitment strategy, everything from newspaper ads to talks at schools to self help groups to pediatric to tertiary mental health centers. The initial contact was used to rule out boys and non-elementary school age kids. Another screening layer were rating scales from parents and teachers, but with thresholds intentionally set a bit low to make sure that we didn't prematurely exclude girls for whom the norms may not be as appropriate. Then when we get to the next gate of the multi-gated procedure, bringing families in. This involved lengthier assessments, insisting on sex-neutral criteria to make sure that the girls who finally entered did meet full diagnostic criteria.

In the end, we actually got more girls than we thought we would need for our design, because we beat the bushes and recruited everywhere we could, had flexible initial gates and then at the final gate made sure that the diagnostic criteria were met.

Recruitment takes time and work. Giving talks at self-help groups; getting to know the local medical and psychological community; running programs that are successful, all of these things helped recruitment and helped maintain the sample.