A bigger study is required to confirm the hypothesis Suvash Shrestha, None 16 May 2012 I read the article with great interest and felt it addressed very important and relevant topic. However, I would like to point out few things about the study that I believe have reduced the credibility of the findings. 1. The sample selection poses a huge problem. Not only the sample size is small but it is also biased. The findings or the opinions are of the residents in one center among many in New York, leave aside the other hospitals in other states. Residents in community hospitals in other places may have different opinion about the fellowship. 2. The number of residents in "Interested in fellowship" and "Not interested in fellowship" is highly unequal with 37 out of 45 (82%) in the "Interested in fellowship" group. The fact that these two groups are highly unequal makes the comparison between these two groups highly biased and unrealistic. 3. As the authors have also mentioned, currently the research respondents are residents, they are yet to join fellowships, and whatever answers they gave are just ideas and not necessarily what they are going to do in future. This difference in thought and action also reduces the significance of the findings. In my opinion, a retrospective case control study which would enroll doctors doing fellowships at the moment and doctors not doing fellowships and look into their past examination reports would better answer the hypothesis. 4. As the authors have also mentioned, the association should not mean causal relation. The relation could be both ways - academically sound residents go for sub speciality fellowship, while academically poor residents are left behind or residents with interest for fellowships work harder so they score high in the exams. Competing interests None.