manas

manas
SAJO'S

WILL&VERDICT
OPINION POLLS AND PSEPHOLOGY
In order to bring as much transparency as possible to the electoral process, the media are encouraged and provided with facilities to cover elections, although subject to maintaining the secrecy of the vote. Media persons are given special passes to enter polling stations to cover the poll process and the counting halls during the actual counting of votes. Media are also free to conduct Opinion Polls and Exit Polls.1 By a recent set of guidelines issued, the Election Commission has stipulated that the results of Opinion Polls cannot be published between two days before the start of polling and after the close of poll in any of the constituencies. Results of Exit Polls can only be published or made otherwise known only after half an hour of the end of polling hours on the last day of voting. However, Opinion Polls are not necessarily right always.
In Opinion Polls the terms ‘right’ and ‘wrong’ are really insufficient to explain how accurate they were. An Opinion Poll might be wrong in naming the winner, yet be within one percent of true division of votes between the winner and runner up. In 1991 during Kerala assembly elections, the results were close to the PASC – Kalakaumudi# Poll in predicting the division of votes between two major political fronts-LDF and UDF- but went wrong in choosing the winner2 . During 1996, Assembly elections Mathrubhoomi,one of the leading Malayalam newspapers ,conducted an Opinion Poll3 by administering the responses from the readers. Despite the brilliant analysis the poll findings were nowhere near the division of votes or seats. Moreover it picked the wrong winner.
The outcome of Kerala elections might be determined by the difference of an extremely small number of votes, but a big difference of seats. By citing the failed predictions a section of the media reaffirmed their belief that the survey machinery has not reached, and probably never would reach the degree of accuracy necessary to forecast the outcome an election, which is decided by fewer votes than the margin of error under which a typical survey operates. For any election survey the margin of error averages around two per cent.
An important learning for the survey researchers from the past experience is that an accurate prediction of a Kerala election is a near impossible task; whether it is in terms votes or seats. The reason for this is plain. Mere five to ten votes in a polling booth could change the results of 10 to 20 of the total 140 seats in the Kerala Assembly. Measuring political sentiments of this level is a challenge indeed; but opinion researchers achieved this feat in 1987 during the Kerala assembly elections. PASC-KALA KAUMUDI poll prediction was very close to the election results, not only in seats but in terms of popular votes also.Opinion polls have been successful in forecasting election results in majority of cases. Notable failures were also there in the forecasts. In Britain, all the poll forecasts in 1992 elections reported Labor victory but owing to the last swing the Conservatives won. India Today - MARG opinion polls, in most cases were close to the mark in predicting the winner seats in the national elections in India, however, went wrong in the state elections of North India in 1994. The MARG poll also went wrong in predicting the region wise results in 1996 general elections; but surprisingly the all India figures came close to the results.
In most cases the major Opinion Polls did not show large deviations from the actual division of votes but nevertheless picked the wrong winner. Predicting elections will always be chancy because of the possibility of last minute swings, turn out problems, voter’s empathy to the looser, and an understanding among the trailing parties after the opinion poll . Just as it can be said with certainty that opinion polls will be highly accurate in the vast majority of elections, so with the same certainty it can be said that on occasions they can go wrong. But when they go wrong, they still should be reasonably close to the truth, if the best practices are followed.
In Kerala with her multiparty coalition structure, it is not easy to forecast the seats and percentage of votes to individual coalition partners. This is because the link between votes and seats are much more variable than in a two party system. Hence the design of the Opinion Poll needs to take into account all the special features of the election that it intends to predict.
Another important requirement of theOpinion Poll is to mark the swing zones# in each district and projection should be done along that swing zones using the data collected. In the parliamentary elections, Kerala as a whole can be taken as homogeneous swing zone but for the assembly election it differs regionally. Even if the elections to assembly and parliament are taking place simultaneously the swing pattern might differ.
Now a new branch of election studies Psephology4 has become very popular especially through television and other media. Psephology derives from classical Greek Psephos, the piece of potteries on which certain votes, mainly about the banishment of those seen as dangerous to state were inscribed. The scientific study of election and forecast – the modern Psephology recently attained a respectable place in social science in its more than six decades of research experience.
From 1930’s on, the spread of Opinion Polls conducted by both commercial and academic practitioners continue at an accelerated pace worldwide. In the recently held American Presidential elections, the same drama that emerged during the counting was foreseen by a number of campaign tracking Opinion Polls include the Gallup . The Labor electoral victory in Britain (1996), Socialist’s gain in Poland (1997) and the ANC’s landslide second win in South Africa (1999) were predicted accurately by Opinion Polls. A hung parliament after the 1996 general elections in India with regional parties holding the key and the BJP as the emerging force in the subsequent elections of ‘98 and ‘99 were predicted by many leading Poll agencies
In India, many Poll organisations now engage in Opinion Polls on important national and international topics. Some agencies in India are working for election surveys and marketing research from the early 1980’s. Of them, CSDS, MODE, MARG are the leading ones.
Now opinion polling has become integral part of modern democracy. Hence media began to provide more space and time to Polls. Regardless of their accuracy in predictions the Opinion Polls serve a purpose of identifying a trend, mood of the voters and their likely behavior. An Opinion Poll reveals the mood of the electorates only at the time of administering the survey.
Some people believe that the Opinion Polls can influence pubic opinion. the Do Opinion Polls actually influence voters?The simplest answer to this question is that if Polls had no significant influence on voters, it is not necessary for the Opinion Polls to be correct. But in many elections, they have been remarkably accurate but also failed sometimes. This itself is a proof that there is no consistent, significant, influence of Opinion Polls on voters.
However it is said that Opinion Polls might influence people in three ways. The first argument is that voters are likely to support whomever the Polls show to be in the lead, so that they have the satisfaction of voting for the winning side. This is known as the ‘bandwagon effect’, as people jump on the bandwagon of the leading party. However, it is also possible that people will react against the party in front, particularly if they are along way in a big lead as they do not wish to see a government with a very large majority. This is called a backlash effect. The final theory has the same effect as the second, as the entire opposite of the bandwagon effect, but for different reasons. This theory argues that people are inclined to support whomever is trailing – an underdog effect, much as is often seen in the field of sports.
The recent civic elections in Kerala that held in two phases with substantial gap, show that even the actual election results of one region hardly had any effect on the other region that go to poll later. If knowing the result merely of an Opinion Poll had an effect on people’s voting behaviour, then one would expect knowing the result of the actual election should have an even grater effect.
.
1 eci website
# Kala kaumudi,a news weekly in Malayalam publishing by rhe Kerala Kaumudi Group of publications,Trivandrum commissioned an NGO, Progressive Artists and Scientists’ Center(PASC) to conduct a Public Opinion Survey during the 1987 Assembly elections.
2Two surveys were conducted. The first one became irrelevant since the election was postponed due the assassination of Rajeev Gandhi.
3 Many media organizations happened to have this mistake by selecting samples from their readers. That will never represent the entire population.
# Constituencies that have same number of votes changed one party to another in successive elections.
4 Dr. David Butler and Dr. Prannoy Roy have popularised the discipline in India.