With the Tamil Nadu elections around the corner, the number of election surveys has been proliferating. For the state, this is the first election with a real multiplicity of 24x7 Tamil news channels, and each is attempting to distinguish itself based on the attention-grabbing numbers coming out of poll surveys.
Frustratingly for viewers, with each new survey comes a whole new set of margins on vote shares and seats won. No one survey seems to agree or align with the numbers of any other.
Tamil channel News7, in association with the daily Dinamalar, for instance, predicted that the AIADMK would win about 87 seats, DMK 141 seats, DMDK one seat, PMK two seats, BJP one seat and others about two seats. These predictions were purportedly the results of a survey of 2.34 lakh respondents, with 1000 samples in each constituency, carried out in a span of six days.
Another survey done by a firm run by Father Gasper Raj, a close acquaintance of Kanimozhi, predicted that DMK will win between 127 and 139 seats, the AIADMK between 81 and 90 seats and others would get 14-17 seats. This survey was reportedly carried out among 12,000 respondents in eight regions in the state over a time period of 10 months.
Contrary to these, a Times Now C-survey in the first week of April predicted an AIADMK win with 130 seats, with DMK getting 70 seats, the BJP scoring a zero and others claiming about 34 seats. The vote share of the AIADMK was predicted to be 39 percent, of the DMK 32 percent, of the BJP four percent and of others 25 percent.
Tamil channel Puthaiya Thalaimurai too showed a huge victory for AIADMK, predicting 66 seats for the DMK, and 164 for the AIADMK, while the others were expected to get four seats.
Then there was Cauvery News, in which a total of 3562 people were surveyed across Tamil Nadu — 2676 men and 886 women. With these 3563 samples, the survey predicted that the DMK would get 125 seats, the AIADMK would get 79 seats, and others 30 seats. It also predicted vote shares for the various parties at 32.9 per cent for AIADMK, 37.9 percent for DMK, 12.10 per cent for DMK, 6.74 per cent for PMK, BJP 3.31 percent and others will get 7.05 percent.
An opinion poll conducted by People Studies of Loyola predicts a vote share of 42.7 per cent for DMK, 36.6 per cent for AIADMK, 7.9per cent for DMDK-PWF alliance and 3 per cent for PMK in the Assembly polls. The survey was conducted in 177 assembly constituencies and 8,064 voter respondents.
For political analysts in the state many of these surveys are equally unconvincing.
Political analyst S Murari, for instance, says that surveys become immediately suspect when they do not clearly state their methods up front. “Most of the surveys do not appear to be scientific. The methodology with which the survey is conducted is not disclosed. In one or two cases, even the agencies which conducted the survey were not disclosed,” he said.
Ramu Manivannan, another analyst, also points out that while most surveys tried to justify their surveys in terms of sufficiency of numbers, many of them suffer from sampling issues. “One common thing in all the surveys is that the methodology is not defined well. The sample size is also very less. The variation between age groups, the urban and rural distinction, all should be included while defining the sample,” he said.
In such a case, Murari says a survey of thousands cannot possibly properly represent crores of voters. “When Tamil Nadu state has more than five crore voters, how can agencies just take a few thousand people and conduct the survey?” he asked.
Senior political columnist TN Gopalan pointed out that it was not merely a question of some of these surveys eventually being proved right or wrong. “These surveys can also influence the voters to vote for a particular party. But we are a democracy so everyone has the right to have an opinion,” he said.
However, he added, the variance in the results of the various surveys may not merely be the result of faulty methods. "Some Tamil channels have vested interests,” he said, adding that the results of many surveys are at variance with ground level trends that journalists are reporting. “It is dicey. Some of these surveys are politically motivated.
Ramu adds that he places far more faith in the perceptions of voters than in the surveys. “The perceptions of the people are much more realistic than surveys.”