The Election Commission on Monday asked for a police case to be initiated against the management of Dainik Jagran, a Hindi newspaper, for publishing results of what appeared to be an exit poll conducted after the first phase of Assembly elections in Uttar Pradesh. The EC called for a First Information Report to be filed against the paper, saying it appeared to have violated rules that prohibit conducting or publishing any exit polls until all phases of the current elections conclude.
“It is in clear violation of Section 126 A B of the Representation of the People Act and wilful disobedience of lawful directions of the Commission,” a spokesperson for the EC said.
Dainik Jagran, which has one of the largest readership of any newspaper in the country, had published what it called “public feedback” following the first phase of the Uttar Pradesh elections. The report said it received its numbers from an organisation called Resource Development International, which used a randomly selected sample of 5,700 voters across constituencies that went to polls on February 11.
That report is no longer online – only leading to an error page on the Jagran website – but The Wire’s summary of the survey said it predicted the Bharatiya Janata Party as being ahead, with the Bahujan Samaj Party second and the Samajwadi Party-Congress alliance third.
What are exit polls?
Exit polls generally involve asking citizens about their voting preferences on election day itself, as a way of understanding voter behaviour and to predict outcomes. The ‘exit’ refers to polling done on people who have just exited voting booths. They are different from opinion polls, which usually take the form of surveys conducted before elections.
In 2008, Parliament amended the Representation of the People Act to explicitly ban the conducting and dissemination of exit polls during the period laid out by the EC, usually until the very last day of voting for any elections taking place simultaneously.
Why are exit polls banned?
The official reason given for the amendment was that when elections are staggered over several phases, such as the month-long process in Uttar Pradesh, “the telecast of exit polls after each phase of polling affects the outcome in the subsequent phase of elections. It is believed that such telecast of exit polls affect the turnout of voters’ also.”
In other words, the authorities were concerned that publication of exit polls after the earlier phases could alter the way voters in later phases.
Conventional wisdom suggests two kinds of ways this may alter voting patterns: The bandwagon effect, also generally referred to in Hindi as hawa, where people change their voting preferences based on the party that appears to be successful early on, and the underdog effect, when political parties work harder to persuade voters if the polls suggest they are falling behind. Both of these could then alter voter turnout.
News organisation have in general been opposed to bans on exit polls, arguing that they are an unfair restriction on free speech. The debate is usually clouded by the fact that many polling organisation in India have been shown to use unreliable methodologies, and on occasion, to even alter their survey results for political aims.
Do exit polls really have that much influence?
Not much research has been able to determine just how much of an influence exit polls have had on the results of elections, in part because voters rely on many factors to make up their minds. Former Election Commissioner SY Qureishi recently wrote that in his opinion it doesn’t matter if there isn’t proof that it could wing an election since, “even a single voter, cheated into believing that ‘X’ is winning and falling in line (because of the bandwagon effect) is bad enough. That can alter the result.”
That said, a paper by economist Manasa Patnam published in 2012, looked specifically at this issue and concluded that exit polls do have the potential to alter voter behaviour, with significant enough effects that the results are affected.
The paper concludes that voters polling in later phases do react to exit poll results, in a limited manner, especially if those polls came from earlier phases in their own states. While the extent to which voters update their preferences varies among parties, “positive surprises in party performance, in the form of expected gains and losses, shift the vote share distribution substantially.”
What does this mean for the UP elections?
Using a model, Patnam suggests early exit polls showing the Congress-led alliance doing much better than expected in the 2004 General Elections gave it almost 15% greater vote share than if the polls had not been published.
The equivalent in the current elections would be to suggest that a positive surprise about the BJP being ahead in UP, where predictions have tended to lean towards the SP-Congress combine of late, would benefit the saffron party in subsequent phases.
This is exactly what MK Venu argues, saying the potentially illegal survey has “greatly helped the BJP by publishing the exit poll for western UP illegally and prematurely, thus creating a positive ‘hawa’ for Narendra Modi in the remaining phases of the election in central and eastern UP.”