UPDATE: Firstpost has updated the story based on clarifications received.
Thanks to the lack of homogeneity and the sheer number of voters in India, poll predictions have always been a difficult task.
So when two research agencies poll the same state during the same time and arrive at two different predictions, it’s not very surprising.
One will be closer to being correct and the other will be further from it and that brings us to the moot question – why conduct these polls at all, when the chances of being embarrassingly off-target are so high as far as the research companies are concerned?
For the media houses who have commissioned the polls, they can get off the hook by simply saying that the failure was that of the pollsters. But with the chances of being wrong so high, why make the investments in such polls at all?
Take the current instance, where the two poll results differ from each other significantly, so at least one of the research agencies will suffer significant embarrassment – perhaps even both of them.
AC Nielsen and ORG India Private Limited conducted polls in Uttar Pradesh for Star News and India Today respectively. The poll for Star News was conducted between 16 and 30 November, while the poll for India Today was done between 13-23 November.
The Star News-AC Nielsen poll threw up the following results, reports Moneycontrol.com. “The opinion poll was conducted in 202 constituencies of the state in November and has predicted 117 seats for BSP and 132 for SP in the 403-member house. The survey predicted 68 seats for Congress and 65 for BJP. It also predicted that Rahul Gandhi's "aggressive campaign" in the state "seems to be striking the right note with the voters there". The opinion poll in 202 constituencies had a sample size of 36,712 people.”
Now take a look at what the India Today-ORG India Private Limited poll says. “The BSP's 2007 vote share of 30.4 percent, which translated into 206 seats in the 403-member assembly giving Mayawati a clear majority, is likely to shrink to 25 percent. The SP's vote share could remain almost unchanged at 25 percent although it may get more seats in what is likely to be a hung house. There is a marginal improvement in BJP's vote percentage from 16.9 to 18. The corresponding figures for the Congress have also increased from 8.6 percent in 2007 to 16 percent, but the survey says the party is unlikely to improve its number of seats substantially from the present 22. More details here.
Who do we believe? One or the other? Or neither?
Neither – because neither of these polls matter, since there have been many impactful developments since these polls were conducted. Mayawati’s proposal to break up the state, the impact of the Lokpal debates, the decision by Anna Hazare not to campaign in UP, the various sops and quotas announced by various parties, the Election Commission’s decision to wrap statues of Mayawati and the BSP’s election symbol, the Elephant, and so on.
Neither poll has a snowball’s chance in hell of being close to accurate – so why undertake the exercise at all?
UPDATE: Firstpost has received communication from Nielsen as under:
“Be advised that the company that has conducted the opinion poll for Aaj Tak/India Today is one ORG India Private Limited which is not related to Nielsen in any manner whatsoever. The organization formerly known as ORG-MARG merged with AC Nielsen in 2000 and the merged entity is now referred to collectively as ‘Nielsen’. The company that conducted the poll for Aaj Tak/India Today is not the one that merged with Nielsen and is not part of Nielsen group, nor is it associated with it in any way.”
Based on this clarification, we would like to state that, since we had never heard of ORG India, we had presumed that ‘ORG’ referred to by Aaj Tak/India Today was the well-known and reputed agency ORG-MARG. The presumption is, as is made clear by the communication received from Nielsen, wrong. Consequently, all references in the article to both research agencies being owned by the same company are wrong.