Report No. 255
(ii) Potential of opinion polls to influence voters
8.9.1. The central justification for the regulation of opinion polls is the preservation of the sanctity and integrity of the electoral process. Concerns on this count arise on the grounds that opinion polls are able to influence electoral behaviour and distort electoral outcomes.
8.9.2. This influence on electoral behaviour can take two forms, assuming that opinion polls are properly conducted.343 Both are predicated on the understanding that the dissemination of social research will alter the original situation so that it is impossible to accurately predict outcomes. First, there is the possibility that a bandwagon or contagion effect could result.
This refers to the case where information predicting the victory of a candidate could lead to votes being switched in his or her favour and away from other candidates. Second, there is the underdog effect. This refers to voters switching to favour candidates not predicted to win, so that the prediction or appearance of success undermines the actual outcome.
343. Herbert A. Simon, Bandwagon and Underdog Effects and the Possibility of Election Predictions, 18(3) The Public Opinion Quarterly 245 (Autumn, 1954)
8.9.3. There is no clear empirical evidence to precisely demonstrate the degree to which these effects play out among the Indian electorate, or even to establish that such effects do operate. Even in jurisdictions (such as the United States344, Canada345, Germany346 and the United Kingdom347) where studies have been undertaken, there is no authoritative understanding on how much influence opinion polls have on electoral behaviour.
344. See, e.g., Richard Henshel and Willian Johnston, The Emergence of Bandwagon Effects: A Theory, 28 The Sociological Quarterly, 493 (1987), noting that evidence concerning the existence of a bandwagon effect in US polls has been mixed.
345. See, e.g., Do Polls influence the Vote? In Capturing Campaign Effects 263-279 (Henry E. Brady and Richard Johnston eds.), finding that polls did influence the vote in the 1988 elections in Canada
346. See, e.g., Rüdiger Schmitt-Beck, Mass Media, The Electorate, And The Bandwagon. A Study Of Communication Effects On Vote Choice In Germany, 8(3) INT J Public Opin Res 266 (1996), finding that opinion polls in Germany do not appear to mislead voters)
347. Catherine Marsh, Back on the Bandwagon: The Effect of Opinion Polls on Public Opinion, 15(1) British Journal Of Political Science 51 (1985)
8.9.4. On the other hand, an argument against opinion polls has been that information from opinion polls confuses voters, or as said by the EC, 'would be a deleterious intrusion into the mind of the voter'348. It is also believed that information from opinion polls may affect voters' perceptions of the chances that various parties may have of winning and consequently, by affecting voters' expectations about the outcome of an election, polls may affect the vote.349
However, certain studies by political scientists suggest that holding a lead in an opinion poll generally earns an electoral candidate no more than a 4%-5% lead among undecided voters.350 In fact, opinion polls published extremely close to the day of polling do not affect public opinion to a large extent because only a small percentage of voters remain undecided by then.351 However, this is not to say that opinion polls do not influence the voters at all, only that the margin of voters actually influenced may remain unclear.
348. Election Commission of India, 'Guidelines for Publication and Dissemination of Results of Opinion Polls/Exit Polls', Order No. ECI/MCS/98/01, 20th January 1998
<http://eci.nic.in/archive/instruction/recent/media/pnxitpoll_FINAL.html>
349. Andre Blais, Elisabeth Gidengil and Neil Nevitte, 'Do Polls Influence the Vote?', Capturing Campaign Effects, at 263,
https://www.press.umich.edu/pdf/0472099213-ch11.pdf>
350. Stephen Ansolabehere and Shanto Iyengar, 'Of horseshoes and horse races: Experimental Studies of the impact of poll results on electoral behaviour', 11(4) Political Communication 413, 417 (1994)
351. Id.