We read UP wrong: Pollsters on their exit poll projections

Incidentally, the BJP also lost the Ayodhya (Faizabad) seat to SP’s Awadesh Prasad.

India Today-Axis My India had predicted between 361 and 401 seats to NDA, and 131-166 seats for the INDIA bloc. Talking to The Indian Express, Pradeep Gupta, Chairman & MD, Axis My India, said, “The NDA alliance seems to be getting close to 295 seats, 60 less than our projected range, largely because we could not get three big states – Uttar Pradesh, Maharashtra and West Bengal – right,” he said, adding, “we could not fathom which direction the Dalit votes were going. Even in the case of West Bengal and Maharashtra, we missed the mark by 14 and 10 seats, respectively.”

Bhawesh Jha, Director of CNX, whose India TV-CNX poll had predicted 371 to 401 for the NDA, says they were yet to ascertain why they got UP wrong, even as all other states were close to the mark.

While Congress leader Rahul Gandhi had dismissed exit polls as “Modi fantasy polls”, on Tuesday, as the results came out, CPM general secretary Sitaram Yechury alleged that exit polls were done to influence the stock market and help those who wanted to make money.

“Exit poll was done to influence the share market and those who wanted to make money in the share market yesterday made money. Now, the reality is coming in front of you. Now let the full results come, after that, we will tell you…INDIA alliance is doing well,” he said.

However, pollsters and psephologists continued to defend themselves, saying they got UP wrong and failed to read the quiet swing in the crucial state. Yashwant Deshmukh of C Voter said, “In terms of state-wise projection, it was only UP that we got wrong. We had already projected that West Bengal and Maharashtra will go down the drain, but somehow in the case of UP, the projection and results were very different.”

ABP News – Cvoter had predicted 353-383 seats for NDA, and 152-182 for INDIA alliance.

Deshmukh added that their methodology has been the same over the years – it gives you the vote share and then you extrapolate the seatshare.  “Maybe, a lot of polarisation happened after the last Assembly election. Also, the baseline for our calculation was the 2019 numbers when SP and BSP contested together. So maybe this time, we didn’t have the correct baseline for around 40 seats, and that could have been the reason mathematically, it didn’t come together. However, we got the vote share for UP more or less right,” he added.

Pradeep Bhandari of Jan Ki Baat, which gave 362-392 to the ruling alliance, as opposed to 141-161 for the Opposition alliance, said, “As far as I understand, projections have been off the mark this time because in UP, we failed to understand the anti-incumbency and also the caste factor… it was a Black Swan event. After all, it is based on scientific analysis of data, and the impact of caste factor in UP could not be ascertained to this extent.”