AWRY AGAIN

How opinion polls always get it wrong with Kejriwal in Delhi

Quartz india
Quartz india

This post has been corrected.

It unfolded as it was expected to.

But nobody, not even the Aam Aadmi Party’s (AAP) internal survey or India’s best exit polls, predicted the landslide victory that the party managed on Feb. 10 in the Delhi legislative assembly elections.

And it is not the first time that exit polls have inaccurately predicted AAP’s victory margin in Delhi.

In November 2013, just a year after the party was launched, most exit polls had suggested a bleak outlook for the AAP. Eventually, though, it pulled off a surprise to emerge as the second biggest party and even managed to rule Delhi for 49 days with external support from the Congress.

This time, too, the pollsters have tripped. The party is leading in as many as 67 out of 70 seats as of 1.15 PM and, in all likelihood, will end up winning more than 60 seats.

The only exit poll that managed to get even close to the actual results in both the elections is Today’s Chanakya, which predicted that the AAP would win 31 seats in 2013 and 48 seats in 2015.

“Caste plays a huge role and we have managed to create a sample size that represents these factors,” a spokesperson for Today’s Chanakya told Quartz. “In the US, they use a smaller sample size to predict polls. But in India, the sample size needs to be representative with caste and religion and in a city like Delhi, these factors continue to play a huge role while voting.”

These five charts show how exit polls have managed to get it wrong in the Delhi elections over the past two years. The upper limit of predictions made by pollsters were selected as their final forecast.

Todays chanakya
India today
News Nation

Correction: A previous version of this story incorrectly showed that Today’s Chanakya had predicted that the BJP would win 48 seats in the 2015 assembly polls. The forecast was actually for the AAP.

home our picks popular latest obsessions search