Indian election: How accurate are exit polls?

Voting for the 2024 Lok Sabha elections ended yesterday with the conclusion of the seventh phase. Before the final results scheduled for June 4, exit polls show the BJP-led coalition is set to win the election by a big margin.

Most exit polls projected the ruling National Democratic Alliance (NDA) could win a two-thirds majority in the 543-member lower house of parliament, where 272 is needed for a simple majority. A two-thirds majority will allow the government to usher in far-reaching amendments in the constitution.

The exit polls for the 2024 Lok Sabha elections were due to be broadcast from 6:30pm yesterday, 30 minutes after the polling ended.

The actual results for all 543 parliamentary seats will be announced by the Election Commission of India on June 4.

Warning: The exit polls do not always get it right. They have been way off the mark in the past.

Here’s a look at a summary of some data on exit polls projections published in various Indian news outlets and the actual results in Lok Sabha elections in 2014 and 2019.

In 2014, predictions for the NDA, BJP-led coalition, were not very accurate. On average, the predictions were off by 55 seats. Some predictions were even further off. Predictions for the UPA, Congress-led coalition, were also quite inaccurate. The average error was just over 50 seats, but the errors varied widely.

In 2019, predictions improved for both NDA and UPA. For the NDA, the average error dropped to about 27 seats, with less variation in errors compared to 2014. For the UPA, the predictions were off by 22 seats on average, with fewer large errors.

In 2014, predictions were not very accurate for both NDA and UPA. By 2019, predictions improved a lot. The errors were much smaller.

This shows that polling methods and data collection got better over time. Polls in 2019 were closer to the actual results.