Thanks for the usual smart analysis. 1. Good point about people who know someone with fatigue being more likely to take part in the study. For accuracy, 43,000 people answered the phone and 20,000 were excluded as had no kids or were a business. Of the remaining eligible 23,000, 5600 took part (24%). So there is scope for selection bias as a result of people knowing that it was a fatigue study, but not as much scope as your starting figure of 148,000 suggests. And I am not sure how this compares with previous studies. 2. Interesting point about a 83% having PEM — I will ask about that. 3. I think this is probably the best adolescent prevalence study out there. It uses a community sample which, despite the bias, is likely much better than trying to go through physicians , through instance, where you depend on physicians spotting it. And it involves a medical evaluation, including a physical exam, which was carried out by Dr Ben Katz, one of the top Paediatric consultant specialising in ME/CFS in the US. It also included a psychological screen and fatigue questionnaire. 4. Although I am also surprised about the very low rate of previous diagnosis, I don't have any reason to doubt it. See above for how diagnosis was carried out. Also, previous studies (one by Jason, won by the CDC also found extremely high rates of non-diagnosis in prevalence studies). Don't forget this is the US where access to medical care is patchy at best. And they have a big problem with physicians being willing to take the illness seriously. I plan to blog about it is next week, once I've had the chance to harvest the wisdom from this thread .