• UNNI RAJEN SHANKER: Most opinion polls for the Gujarat elections were wrong. In USA, polls for the Democratic primary in New Hampshire were way off the mark. Here, we have two of India’s best-known pollsters. They will talk about the methodology of pollsters, how they get it wrong, how they can get it right.

RANJIT CHIB: I would draw a distinction between pre-polls and exit polls. I’ve been doing pre-polls and believe that they are more accurate than exit polls. Pre-polls are sometimes published by the media. Sometimes they are done for political parties, so they are not published and the public is unaware of our findings. Pre-polls conducted by us for the media since 1989 have been fairly accurate.

• COOMI KAPOOR: You got the Gujarat election results right a month before the elections.

RANJIT CHIB: Yes, we knew the Congress would gain a few seats and the BJP would lose some. All the other polls were showing a close contest. In the Uttar Pradesh elections it was a similar situation: 10 days before polling, I predicted that Mulayam Singh Yadav would not get even 100 seats. Others, including Yogendra Yadav, had the Samajwadi Party marginally ahead of the BSP.

If our pre-polls are more accurate than others, it’s due to our methodology, which we developed in 1989. We have a lie factor, a default factor, and a third factor that takes into account people who claim to have voted in the last election but did not vote, or people who say they will vote but do not. So we do not take voting intentions at face value.

... contd.