I have read the paper linked by anonymous reader on my last post. For the interested the abstract is here and the full paper is here.
I have tried to review the paper critically. I loathe to admit it, but I don't understand the statistical methods used in this paper entirely. However, I'm always sceptical when it comes to media, and this will be no exception so I will talk about the issues that bother me.
Looking at the wiki page for the regression method to be used correctly, I find that
"1)The sample must be representative of the population for the inference prediction."
The sample used was equal number of applications from the dalit, muslim and upper class population which violates this basic assumption.(population ratios are 12% muslims, upperclass 25-44% , dalits 24%)
I need not go further but let me add, the study might still have been valid if the "12 applications" sent by the authors of this paper to each job were the only applications they received.
But by their own admission, these were jobs at "entry level" to which any graduate could have access to and advertised widely(a fair assumption).
It doesn't need a genius to figure out that this study falls flat when the number of graduates churned out by our education system each year are considered. Also consider that the number of responses they received (total) was less than 10%.
Even if the "correlation" claimed by this study were to be true, it doesn't imply causation.
On the other hand as some people have pointed out in here, it might be a case of market adaptation.
However, I do realize that we all have stereotypes and bias as demonstrated over at IAT.
I just think that reservations only help in promoting these stereotypes and exacerbating the situation.
Lastly about the concept of introducing reservations in private institutions, it is infringing on personal rights.
PS: on the same note watch this brilliant interview