Thanks for your detailed reading and response.

My essay is about the claim that West Bengal is on a trend moving toward a Muslim majority state, not about explaining how the BJP mobilised its support. So your last paragraph is not relevant to it, and I will not comment on it.

Thanks for sharing the Piketty files on the 1941 census. It is a lengthy document and rather than going through it I will accept your assertion that the pre-Partition percentage of Muslim population in what is currently West Bengal was different from what I stated in my essay (I got that number from an indirect source). However, given the churn that Partition caused, this number is not the clinching argument. What is more relevant is the trend shown by the seven census operations from 1951 till 2011. This shows that Muslims achieving numerical majority across the state is statistically improbable and there is no data the underlies the claim of a trend toward a Muslim majority.

Finally, at no point did I state that first-person accounts are to be disbelieved. I have no doubt that the kind of incidents you describe did happen. There is no clear line dividing good and bad people or political parties, and there is enough villainy from all sides. The point I sought to make about first-person accounts is that they are insufficient for forming conclusions on overall trends given the well-established phenomenon of confirmation bias. I came to know about this phenomenon from the book by Nobel Laureate Daniel Kahneman, "Thinking Fast and Slow," and have since read about it from many other sources, including the Wikipedia link I cited.