... as opposed to one-off performance. Indian Express reported recently on an internal study conducted by the IITs.
Apparently, this study found that good academic performance in IITs correlated better with the students' academic performance in school than with the JEE rank.
It doesn't surprise me at all; more importantly, if this finding is confirmed by studies on larger samples (for example, a state-wide study in Karnataka, or an expanded study in IITs including the data from several decades), it would undermine our system of selecting students based on some one-off exams (offered just once a year!).
I am glad this study has been done. I would like more such studies on the IITs and IITans. There is a great gold mine of information sitting in the vaults of IITs -- if tapped by a skilled sociologist, we will know a great deal more than what we know about the 'creamiest of the cream' of science-and-tech-oriented students in India.
Wouldn't you like to know -- I mean, really know -- about other correlations among things like their school records, JEE ranks, their background, their parents' income and education levels, parents's occupation, urban-rural divide, whether they went through coaching classes, proportion of women and dalits (among JEE takers and in the incoming class), where the students go, and on and on. As I said, all this information is just waiting for a major sociological study. Since such studies have not come to light, it means either the IITs are not interested in parting with the information, or the sociologists are not interested in studies of this kind.
I don't believe the problem lies with sociologists' lack of interest. The argument is quite simple. The Indian society (and, more recently, the entire world) has shown an enormous interest in knowing more about these great institutions and the geeky subculture that thrives there. Anything that you say or write about IITs will sell, and sell enormously. Just ask Sandipan Deb and Chetan Bhagat.