Follow Us:
Friday, December 03, 2021

The mythical ‘science’ of predicting elections

If you go by the record of our pollsters in the last five years,you will find they got it wrong as often as they got it right.

Written by Coomi Kapoor | New Delhi |
May 1, 2009 4:21:52 pm

If you go by the record of our pollsters in the last five years,you will find they got it wrong as often as they got it right. The most famous example of a collective goof-up is,of course,the 2004 parliamentary elections. Every psephologist foresaw an NDA victory and the BJP far ahead of the Congress. Some virtually wrote the Congress’s obituary,claiming that the party could get less than 100 seats.

As we now know,the Congress party in fact ended up ahead of the BJP. In other professions,such major miscalculations would have forced people to shut shop,at least for a season or two,and appear suitably contrite.

But our psephologists moved merrily on to forecasting the next election without a pause. One pollster kept making the point on television that he had got it less wrong than everyone else.

Another questionable tactic of Indian pollsters is to keep drastically altering their guess estimates about poll outcomes in the six months leading to an election. Whereas the general belief is that unless a major scandal erupts,the majority of people do not change their minds in the three months before a poll,no matter how the campaigns of the political parties pan out.

So why do we adopt such a forgiving and tolerant attitude towards pollsters in our country? For one,there are a limited number of major players and you don’t really have much freedom of choice. If you fire your pollster,there is a good chance you may not get another well known face. Besides,there are always advertisers ready to bankroll surveys on the state of the nation and the likely poll outcome results,and enough credulous people to take the findings seriously.

In any case,our psephologists are among the glibbest talkers in the media business. They can talk rings around you about the swing factor,the switch factor,the margin of error,the index of popularity,the index of unpopularity,the index of opposition unity,the index of opposition disunity,the silent voter,an unspoken undercurrent and so on.

The “margin of error” phrase is a particularly useful safety belt. Most pollsters put this margin of error at a high five per cent on either side. Another helpful ploy is to predict a bandwidth rather than a single figure. So if you predicted between 100 and 115 seats for party A and 95 to 110 seats for party B,you can still claim to be correct if A gets 100 and B gets 110,even though you predicted that A would be the frontrunner and it actually ended up being the loser! I once saw a pollster patting himself on the back for just this scenario.

And if all else goes wrong,you can always point out that you had mentioned there was a five per cent undecided voter,and all of them swung in one direction at the last minute leading to the miscalculation.

Of late,a new trend has emerged. Give a detailed discourse on a state going to the polls. Provide percentages of what people think of leader A and leader B and issue C and issue D,but avoiding putting your money where your mouth is by projecting a seat tally. Without giving specific figures,but simply using jargon,you can claim you predicted every possible outcome.

The other day while switching TV channels,I heard a commentator explaining that a particular party in a particular state was on the edge of big gains. Then he qualified this by explaining that factionalism could hold back the projected rise in the party’s vote share. He then added the rider that one must not forget that minor issues can in the circumstances lead to major vote swings. I am a little dense and was at a loss to comprehend what he wanted to say,even though he had all sorts of graphs and charts to point to.

I may be over-cynical,but I get the impression in fact that organisations and political parties select pollsters whose own political predilections are in sync with theirs. Or else pollsters smartly fine-tune their surveys to suit whoever is footing the bill. The boss after all likes to hear good news,not Cassandra-like prophecies.

When irate politicians question the data of poll predictions,which vary drastically from their feedback at the grassroots,television anchors and interviewers explain patronizingly that poll prediction is a science,based on statistical data and there is no chance to fudge. But a well known pollster once admitted in an interview that he alters his final reading from the “raw data” he receives because he has to add his own input in the light of certain perspectives.

For instance,a pollster has to take into account “the fear factor”. This means that dalits may be too afraid to disclose their vote preference openly to caste Hindus. Or in a state like Gujarat,voters might feel insecure over openly voicing their resentment against the government. Some would dub this “perspective” “doctoring” — but heck,who can argue with “scientists” who have the disarming ability to throw statistics of all sorts in your face?

This election,the pollsters uniformly predict the Congress will better the BJP in the numbers game. This fact alone makes me uneasy. Judging by the previous record of pollsters,this really means there is a 50 per cent chance that they could be wrong.

📣 The Indian Express is now on Telegram. Click here to join our channel (@indianexpress) and stay updated with the latest headlines

For all the latest Opinion News, download Indian Express App.

  • Newsguard
  • The Indian Express website has been rated GREEN for its credibility and trustworthiness by Newsguard, a global service that rates news sources for their journalistic standards.
  • Newsguard
0 Comment(s) *
* The moderation of comments is automated and not cleared manually by