Some issues with this study Anirudh Kumar, CIPM 15 August 2011 As an Indian, I cannot help wondering why Pondicherry was chosen as representative of India at all. The Election commission of India has some sophisticated statisticians who could have helped you with location of a more representative site. Although there is the usual mea culpa lip service of limitations, I find there are too many glaring ones both in methods and interpretation of results. I only scanned the paper so if I misread any, I am sorry but I do not think I did. Here are my reservations: 1. Your response rate for Pondicherry was a whopping 98.6%, and that is surprising even among the low-middle group which had higher rates of response. One could perhaps reason that in rich countries people have survey, interview and focus group fatigue. 2. But how do you explain this: you find that low-income families in India are at a higher risk (14 times?) of life-time MDE but you seem to think that high-income individuals are more prone to MDE. In fact, high income individuals seem to come out with lower levels of propensity on many of your measures. 3. Since I know the context a little bit, I cannot but wonder if the low income group understood the questions and responded candidly after the usual probing, clarifying etc required of interviewers. a. Interpretation of what has been clarified. b. Do not forget that contemporary Indian culture glorifies melancholia. Take a look at the soaps on Indian TV, a large part of the various Bolly etc-woods output, contemporary writers who routinely put out tragedies. After all, middle aged women to date swoon over a rich fool who dies under a tree penniless and drunk. There is no doubt that other cultures also have their "tragic" characters, but in the contemporary Indian context, it is taken to the extreme. It is more likely to be so with the low-middle income households with less exposure to the cosmopolitan tastes cultivated by the higher income bracket. The upshot is that being depressed becomes something heroicly romantic (both mean the same but to give some emphasis), and pride in being depressed is accentuated. No wonder you found the results. 4. It would have been useful and more credible if your group reported on the composition of the final sample (usable responses) in terms of demographics particularly for the lower-middle income countries. These cultures are typically conservative with high income brackets do not really like the idea of self-disclosure to a stranger (vulnerability issue) even if s/he flashes a WHO id card. In that case, one has to assume that the sample came from patients or referrals of those who were treated or were visiting the involved investigators in some way or the other. A fuller disclosure would remove these suspicions. 5. Finally, although Psychiatry- granted it is my opinion- has since risen from the couch for a glass of water and prozac, it still tip-toes that line between pseudo- and hard-science. Interviews and surveys may not be the best way to establish its scientific character. I do not doubt that India has its share of depressed people and even manic depressives. However, to conclude or infer about proportions- it might make for a nice publication contributing towards tenure or increment- is erroneous. Since I know the context, I made a few observations however ignorant I might be. I wonder how many others are easily misled by the findings (I came across these findings not because of professional interest but because I found the report in a popular magazine). It is useful to remember what Dr. Ioannides has been reporting on medical science and statistics over the last few years. Competing interests None, I am not even marginally connected to this field of inquiry.