JOIN USCritics say poll panel move is an insult to voter’s intellect, no proof to say surveys sway electors.
In 2002, C fore conducted exit poll for Doordarshan, the only government channel in India and also the most viewed in those days, for elections in Uttar Pradesh, Uttarakhand, Punjab and Manipur. The election was in four phases and for every phase we conducted exit poll. Our predictions were on the dot. The seats and vote share we predicted for different parties in different phases were very close to the actual results much in contrast with other channels. The Doordarshan officials were very happy.
But to my surprise , few days after the election results were out, Doordarshan took a decision not to conduct any election survey. The decision was taken because some political leaders of a party believed that they lost elections due to exit poll results. They conveniently ignored the fact other channels had shown that the party was leading in two states in their exit polls even though they were losing. How come voters were influenced by one poll and not by another? Instead of working on their governance, it has become convenient for some politicians to look for scapegoats somewhere else.
When the results of second phase of UP elections were being telecast, I got a call from a senior politician from the state. He was very angry with our forecast. When I asked for the reason, he did not say voters were being influenced with our predictions, he was angry because he believed that predictions were bringing down the morale of his party workers who were managing the polling booths. He angrily remarked that party workers were not bringing enough voters to booths. I had heard of “bandwagon effect” and “underdog effect” of opinion polls but this reasoning was new for me.
For starters, bandwagon effect is believed to occur when some voters eventually vote for the winning party as declared by surveys while the underdog effect refers to the tendency for potential voters to be attracted to the trailing party as projected by opinion polls. To test the impact of polls on voters we conducted a post-poll survey in Punjab. We spoke to those voters whose database we had and who had in the pre-poll survey that we had conducted a month back , said that they have yet not decided whom they will vote for.
This was the segment that was supposed to be most influenced by opinion polls. None of the about 300 voters that we spoke to, said they were influenced by opinion polls. In fact, their voting was evenly split between Congress and Shrimoni Akali Dal(SAD), although marginally more number of them had voted for SAD, the party we had predicted to be losing. We also spoke to workers of a particular party in UP whose leader had spoken to me.
The party workers told us that their morale was down because they did not get sufficient funds as was expected and also because of lack of enthusiasm among voters due to bad governance in the state. Again none of the party workers that we interviewed said that they were in any way affected by opinion polls.
No evidence
It’s mostly politicians who want opinion polls to be banned although there is no concrete evidence that they in any way influence public opinion and voting pattern. Polls, if properly conducted, only mirror ground reality and the reality changes fast in election season. Pre-poll surveys predict vote share and seats to be won at that particular time. Things can change at the time of polling.
Theoretically, exit polls or post-poll surveys should be most accurate. When elections are held in phases, telecast of exit poll results is allowed only after the polling process of all phases is over. In case more than one state is going to polls, then when the voting process ends in all states the exit poll results can be telecast. Now, the new proposal is to ban pre-poll surveys also from the time the Election Commission announces the date of polls till the voting process is over. This is not justified.
The dates are usually announced more than a month before the actual voting takes place. Take the case of current Assembly elections in five states, the EC announced the dates on October 4 and the final elections – for Delhi - will be on December 4, after a full two months!
Banning pre-poll surveys would also mean that we are undermining maturity level of our voters. With my experience of 14 years of conducting surveys, I can say that average Indian voter is too mature to be swayed by election surveys.
If opinion polls are banned then rumours will replace them. In Delhi there has been an SMS campaign by one political party claiming that they are ahead of all parties as per a survey done.They also put banners behind autorickshaws giving figures of a survey conducted by them. I agree that there are black sheep among pollsters who don’t properly conduct polls but no reputed agency will deliberately indulge in false predictions as their credibility is at stake. Besides skill of psephologist , errors can creep in at the stage of selection of sample, data collection from the sample and analysis of sample.
Media should evaluate performance of polling agencies after the results are out and blacklist those who consistently go wrong in their predictions. We at C fore have displayed our accuracy rate of 95 per cent for exit polls on our website.
Surveys should be seen as important aspect of democracy as they gauge public mood, can serve the purpose of understanding voter concerns and starting point for debate on public issues.
(The author is chief executive of C fore, a multidisciplinary research organisation specialising in election surveys)
Related Stories
When SC forced EC to retreat
Banning opinion poll would be travesty of justice, voter not ignorant