Psephology, the study of elections and voting, came into its own in India, after journalist and economist Prannoy Roy, writing for a weekly magazine, correctly predicted the outcome of the 1984 General Elections. Roy did the same for the 1989 General Elections ~ in a television programme commissioned by Doordarshan. Roy’s accurate election forecasts set off a trend of predicting election results, with every journalist worth his salt, joining the bandwagon. Sadly, unlike Roy, most of the newly minted ‘psephologists’ had little training in statistics or statistical methods, and in the guise of forecasting election results, most were trying to create a tempo in favour of certain candidates.
Taking note of this deleterious trend, the Election Commission prohibited publication of exit polls etc. till the last vote was cast. It is quite another matter that election forecasts are freely available on You Tube and social media, even while elections are on. Aired in the interval between casting and counting of votes; exit polls provided both entertainment, and political insights to viewers. Most surveys got the election re – sults right, though with minor discrepancies, and thus prepared the public and candidates for the results that were to follow.
The first time that opinion polls got election results terribly wrong was for the Delhi Assembly Elections 2015; almost everyone predicted a majority for BJP, which ended up with only three seats, in an assembly of seventy legislators. The issue of biased opinion polls drew the public’s attention after the 2024 General Elections; contrary to what surveys had predicted, the NDA did not end up with a big majority, and again contrary to predictions, opposition parties were not decimated, but increased their tally considerably.
Catastrophic events followed the airing of exit polls, the Sensex rose dizzyingly, falling around six per cent during counting ~ wiping out lakhs of crores of investors’ money. The fall in Sensex after publication of election results can be easily understood; elections are watershed events in India, which bring normal life to a grinding halt. Right from the day any kind of elections ~ Municipal, State or Parliamentary ~ are announced, Government offices cease to work. During elections all conversations and activities invariably veer around to election results; everyone becomes an authority on who will win and why. It seems that newspapers, magazines and TV have no other news to propagate. As a result of endless miniwars on WhatsApp and Twitter, elections make foes out of friends, and vice-versa. Despite such a high degree of informed public involvement, it is tragic that all parties view the voter only as a member of a particular caste or community.
Disgusting caste and religionbased analyses of voter profiles are commonplace ~ even in respected journals and TV shows. By implication, the Indian voter is supposed to vote on the basis of his caste and community. Lack of an alternative identity forces voters to tacitly accept such compartmentalisation, which is strengthened by unabashed appeals to vote on caste and communal lines. However, the innate strength of our democracy falsifies such ill-founded assumptions, and ensures that each election throws up different results, even though composition of the electorate remains unchanged. The underlying method of statistical analysis is to draw a random sample from a population, and use it to make generalized observations about the entire population, which can be an estimate of the way the electorate would vote.
Researchers often devise a questionnaire, and gauge the electorate’s voting preference by their responses. Given India’s great diversity, issues differ from place to place, which makes accurate formulation of issues, almost impossible. Also, a 1 per cent sample size for predicting the way 90 crore electors would vote requires 90 lakh people to be quizzed ~ almost hundred times more than the largest voter survey. Then, there are many types of statistical biases which the researcher should avoid. For example, the researcher should take an unbiased random sample, which is a difficult task, because persons with definite political views will voluntarily offer themselves for sampling, leaving out the vast majority of voters.
Another kind of bias is the ‘omitted variable’ bias ~ two variables are correlated doesn’t mean one caused the other ~ like depending on the Phalaudi Satta Bazar betting to predict election results. This can be equated to a famous example of analytical error: “…Robert Matthews discovered an extremely high correlation between the number of storks in various European countries and the human birth rates in those countries. Using Holland as an example, where only four pairs of storks were living in 1980, the birth rate was less than 200,000 per year; while Turkey, with a shocking 25,000 pairs of storks had a birth rate of 1.5 million per year. The correlation between the two variables was an extremely significant 0.62. Matthews used this example ~ drawing from the myth that storks deliver new born babies ~ to illustrate that correlation doesn’t imply causation. The high correlation between the two variables doesn’t imply that a high stork population causes an increase in birth rate.
Rather, there’s a third variable at play: geographic area. Large countries have more people living in them ~ hence higher birth rates and a higher stork population” (Jenny Gutbezahl, Harvard Business School). Despite such obvious limitations, ersatz psepho logists have no qualms in predicting the election outcome to the last seat. Once elections are over, these soothsayers go into hibernation ~ till the next election. However, opinion polls in India, go awry mainly because of ‘Experimenter Expectations,’ that is most analysts, motivated by chai-pani, want the results to come out in a particular way; a sting operation on a reputed poll agency in 2014, had the company’s business development manager saying that they would charge a separate fee for manipulations, labelling them as ‘margin of error.’
It does not help that despite the Election Commission’s admonitions, black money is liberally used in elections; in another conversation a senior executive of one of the companies, was heard saying that very little payment was made in white ~ for a survey worth Rs.4 crore only Rs.12 lakh was received by cheque. Another factor is that the Indian voter does not take elections very seriously because elections do not change anything ~ the same persons are elected every time, though representing different parties, and after being elected all political parties follow a similar agenda. Sadly, no political party tries to gauge the aspirations of voters through voter surveys ~ voters may be more concerned about their daily problems, than constructing a temple or statue somewhere, or hammering Pakistan.
Election manifestos are replete with extravagant promises which are forgotten as soon as the elections get over. Over time, voters have come to treat political manifestos not as letters of intent but as works of palliative fiction. Sadly, there is no mechanism in law to hold a political party liable for not fulfilling the promises it had made for winning elections. Moreover, most Indian politicians have no record of public service; entering politics as they do only on the strength of their criminal antecedents, or political lineage. Once elected, most MPs and MLAs do nothing more than interfere in local administration, and vote according to the whip issued by their party leaders. The interests of their constituents, for which they can at best spend their MPLADS funds, are soon forgotten.
Visuals of MLAs and MPs beating each other or public officials, often with shoes, or watching porn in the assembly have only added to the notoriety of our public representatives. Then, we had cases where the lawmakers cast their votes for money, or asked motivated questions in return for cash. The public is rightly wary of their representatives, whom they regard as a necessary evil. The difference in the level of private facilities in any field, as compared to public facilities in the same field shows the extent to which public representatives have failed to improve the lives of the common man.
Ever-increasing cash and liquor seizures during elections show that rather than waiting for a particular political party to come to power, and fulfil their promises, voters prefer instant gratification through cash or liquor ~ which queers the pitch for election surveys. However, the main reason why surveys fail to predict election results accurately, and which applies equally to electors, their representatives and polling agencies, was stated thus by the German “Iron” Chancellor Otto Von Bismarck: “People never lie so much, as after a hunt, during a war or before an election.”
(The writer is a retired Principal Chief Commissioner of Income Tax)