New Delhi: Most exit polls had given Congress a decisive edge over the ruling BJP in the Haryana Assembly elections. However, not for the first time, results of exit polls proved to be far from reality.
On Tuesday, when votes were counted, BJP sealed a hat-trick of wins in the northern state with 48 seats in the 90-member Assembly.
After polling ended on October 5, Dainik Bhaskar had predicted Congress would win 44-54 seats while giving BJP 15-29 seats. The C-Voter-India Today poll had Congress ahead with 50-58 seats, and gave 20-28 seats to the BJP. Republic Bharat-Matrize had forecast an even bigger win for Congress, projecting 55-62 seats against 18-24 for BJP.
The Red Mike-Datansh exit poll also predicted a Congress victory with 50-55 seats, while giving the BJP 20-25. Dhruv Research saw Congress winning 50-64 seats and the BJP 22-32. Peoples’ Pulse claimed Congress would form the government with 49-60 seats.
All these predictions turned out to be inaccurate.
So did the Jammu & Kashmir results. While most exit polls predicted a hung assembly, National Conference alone grabbed 42 seats out of 90 and its ally Congress got 6, limiting BJP to 29.
Even in the Lok Sabha elections, almost all exit polls got the results wrong. While most polls predicted an absolute majority for BJP (above 272) and 330-350 seats for NDA, the ruling party actually stopped at 240 and formed the government with the help of allies as NDA parties totalled 293.
No exit poll had seen the Opposition INDIA bloc getting 243 seats, including 99 for Congress.
So why are the pollsters getting it so wrong?
As for Haryana, among top factors leading to the inaccuracy in exit polls was the fact that they got overwhelmed by Jat opposition to the BJP. They overlooked the silent consolidation of non-Jat communities like Ahirs, Brahmins and Punjabis.
Voters not revealing their actual preferences and deliberately providing misinformation has emerged as another key reason for exit polls to make wrong predictions.
Some analysts have also pointed out that limited sample selection and wrong methodology make exit polls less reliable.
Instead of exit polls, many agencies conducted post-poll surveys, interviewing voters at home or markets instead of just outside polling booths. These surveys failed to factor in BJP workers’ last-minute efforts and mobilisation.
Also, misinterpretation of vote share into seat prediction led to errors. Despite close vote shares between BJP and Congress (39-40%), seat conversions were misjudged, with BJP winning more due to better dominance in competitive areas.
In exit poll surveys, phone interviews and errors in reporting field observations further impacted the accuracy of forecasts, which were being made based on a small number of respondents. Under-representation of critical data such as caste, socio-economic status, and gender contributed to potential inaccuracies.