The Hindu : National / Elections 2004 : Making sense of the opinion polls

Online edition of India's National Newspaper
Monday, Apr 19, 2004

National

National - Elections 2004

Making sense of the opinion polls

There is a near consensus among the surveys on the outcome of the elections in 12 of the 18 big States, says Yogendra Yadav

It's been raining polls in the last few days. Between Thursday and Saturday last week, as many as five new nationwide forecasts have been released by the media. It is hard even for professionals to keep track of this flood of polls and forecasts. It is understandable if an ordinary citizen feels dazed and somewhat confused. Let us therefore try to take stock of all the polls and try to look at the big picture along with the relevant details.

The big story that emerges from the latest round of polls is that the NDA juggernaut is now beginning to slow down. If one takes a simple average of all the five latest polls (NDTV-Indian Express, India Today-Bhaskar, The Week, Outlook and Star News) the NDA's tally works out to 271 seats, just below the majority mark in the Lok Sabha. That would bring no comfort to the NDA that was expected to win between 330 and 340 seats in the first India Today forecast in February. Since then, fresh polls have tended to bring NDA tally down, even if marginally. It is a difficult business to espy a trend across polls taken by different agencies with different sample sizes and designs. But now there is more robust evidence to suggest a decline in the NDA's fortunes. The update poll conducted by NDTV-Indian Express in the first phase constituencies shows a clear swing of more than one percentage point cutting across nine States. In terms of seats the NDA had lost 14 seats out of the 140 that are to go to polls in the first phase. If the trend is projected on to the remaining constituencies of these nine States, the loss could be about 25 seats. Something is clearly happening here.

Politically, the relevant question is, are we looking at a trend that is likely to continue further or has the situation stabilised? If it is a systematic trend, this is very bad news for the NDA. If its tally were to keep dropping in the next three weeks, even if marginally, it could fall below 250. There is another way of reading the situation. What we are witnessing is a one-time downward adjustment that happens when an abstract choice made before the candidates are announced is translated into a concrete choice in a real life political contest. On this reading, the NDA will not continue to lose in the next few weeks but will stabilise around the current projections. As yet we do not have sufficient evidence to say which of these readings is true. It is crucial to keep a close watch on the opinion trends in the next weeks.

As yet the big picture of the national verdict centres round the two possibilities mentioned earlier in this column. The first possibility is that the NDA secures a clear majority on its own. It looked the most likely outcome a few weeks ago and remains a possible outcome even now. The second possibility, that of the NDA falling short of a majority but still forming the Government by securing support from non-partners, now appears to be more likely than it did a few weeks ago. The third possibility, that of the non-NDA parties coming together to form the Government, was an outside possibility right from the beginning but receded into the background once the Congress-BSP alliance failed to happen. It still remains an outside possibility, outside the range of the most likely, after the latest round of polls. The Congress and its allies are still expected to fall short of the NDA tally by nearly 100 seats. If about 50 seats of the Left are added to this tally, the combined opposition would still need to garner additional 70 MPs. That would be a tall order. The worst, and real scenario, for the NDA would be that it can form the Government, but not without securing the support of a major party like the SP, which can then extract a political price.

Which of the possibilities turns into reality would depend on how the contest evolves at the State level. A new and welcome feature of the opinion polling in this election is that some of the major forecasts have given a clear State-by-State break-up of seats for the big States. Earlier the pre-election surveys tended to employ a small sample (typically about 15,000 to 20,000) that did not allow them to give any break-up below the four broad zones in the country. In an era where politics is lived and articulated at the State level, this was quite unsatisfactory for political analysis. This time, the NDTV-Indian Express led the move to give State-wise break-ups and the second round of India Today-Bhaskar poll and the Star News poll followed it up, though the latter has done it on the basis of a small sample of 12,000. The Outlook poll gives the break up for the three State assembly elections. If one puts all these trends together, one gets a fairly clear picture for most of the major States. There is a near consensus among the various polls in as many as 12 out of the 18 big states (with 4 seats or more). Let us first look at these.

Rajasthan, M.P., Chhattisgarh: All the polls point to a BJP sweep in these three States where the BJP won four months ago in the Assembly elections. The victory in the Assembly elections did not indicate a sweep, at least in Rajasthan and Chhattisgarh. The BJP had done very well here in the 1999 Lok Sabha elections too, but it did not sweep these States. This time the indications are that the BJP has gained votes since the Assembly elections in all these States and may pick up most of the few seats that it does not already hold. The Congress may have to struggle to pick up a seat in Chhattisgarh, the State where the BJP has recently benefited from incorporating an NCP faction. In M.P., the Congress bastions like Chhindwara do not appear safe, though polling is still two to three weeks away. The Congress may do marginally better in Rajasthan.

Gujarat: One could club it with the three States above that are showing signs of BJP sweep. No forecast has given the BJP less than 20 seats out of 26 in the State. But the NDTV-Indian Express poll found 59 per cent respondents agreeing with the suggestion that those not voting for Modi's party may be scared to say so. If the polls are not distorted by the fear factor, the BJP could sweep the State and add to its already high tally.

Haryana: Although polling is three weeks away and the effect of the candidates is yet to be measured, it is clear that neither the BJP nor its estranged ally, the ruling INLD led by Chautala, can look forward to the kind of windfall they had last time when they won all the seats. There is a distinct anger against the incumbent party and the Congress is set to benefit from the three or four cornered contests. The estimates of the Congress seats vary between 5 to 7, out of the 10 seats.

Karnataka: All the polls agree that the BJP has an upper hand in the Lok Sabha elections and can well surpass the previous best of 18 that it secured in 1998 with the help of its allies. The latest round of surveys indicate that the party may have hit a peak and may drop a little in the final run-up to the polls. The signals for the Assembly elections are conflicting and range from a slight edge to the Congress in a hung Assembly, to forecasts of a clear majority for the Congress to predictions of a clear victory for the BJP. I would wait for exit polls, or better still, for the actual results of the Assembly elections.

Tamil Nadu: Every poll points in one direction here: the DMK-led alliance has an advantage over the AIADMK-BJP combine. The only question is about the number of seats this will translate into. Will the DMK-led alliance win about two-thirds seats, as many polls are forecasting? Or are the pollsters playing it safe here, and we will finally see a traditional Tamil Nadu style complete sweep?

Orissa: It is clear that the BJD-BJP alliance is not affected by anti-incumbency. The NDTV-Indian Express survey does report a small swing away from the ruling combine, but it is still a long way before this swing can cause serious damage. The alliance looks set to retain its majority in the Assembly as well, for there are no signs of any ticket splitting here.

West Bengal: There is little doubt that the Left Front is set to continue its dominance in the State and secure about three-fourths of the seats. There are no signs of any significant revival of the Congress. The only question is whether the Trinamool-BJP alliance can retain its impressive haul of 10 seats in the last parliamentary elections. The polls differ on whether it will retain its seats or lose some to the Left Front. Perhaps the confusion is because the fate of the two partners is somewhat different: the BJP may gain while Mamata's outfit may not.

Uttar Pradesh: As of now, there is a surprising degree of unanimity among the polls about U.P. All the polls show the SP and the BJP in a joint lead, followed way behind by the BSP and the Congress. But it remains to be seen how this trend translates into seats. Two warnings may be useful here. First, it may be too early to make confident seat forecasts in the triangular or four-cornered contests that U.P. is likely to witness. Second, in recent years, the opinion polls have tended to underestimate the BSP and overestimate the BJP in the State.

Uttaranchal and Himachal Pradesh: Although the Congress is in power in both States, the two polls that give a clear forecast for these (India Today and Star News) suggest that the advantage lies with the BJP. It needs to be remembered that it is difficult to predict smaller States with the kind of small samples that polls normally use.

The second category has some States where all the polls are not quite saying the same thing. The signals coming from the various polls are conflicting or changing or simply unclear. Let us take a look at these States to see if the uncertainties can be telescoped into a range.

Andhra Pradesh: The mother of all uncertainties in this election, the State has the pollsters, politicians and the public quite confused by now. There are at least three sources of uncertainty here. One, the trends have changed significantly over the last six months and is changing in the last weeks of elections, a point brought out very well by the Outlook State Assembly poll. Two, in the TRS we are dealing with a new political outfit on the upswing, a psephological nightmare all over the world. Three, there is a possibility that the State might witness a split verdict on regional lines: Telangana and the rest of Andhra Pradesh may vote as if they are two different States. The NDTV-Indian Express poll points in this direction. The cumulative effect of all this is that we have all kinds of forecasts: a near sweep for TDP-BJP or the TDP-BJP just ahead of the Congress alliance or the Congress alliance ahead of the NDA. All one can say is that it is not going to be a cakewalk for the NDA and that there is no indication of a Congress sweep in all the regions of the State. One can also say that there are no signs of a serious difference in the Lok Sabha and Assembly polls. Exit polls might reduce this confusion. I would wait for May 12, the day Andhra Pradesh counts its vote, a day ahead of the rest of the country.

Bihar: Another psephological nightmare in the best of times, the State has seen quite contradictory forecasts being made about it. One set of forecasts suggests that the BJP and the JD (U) will continue their dominance and win anything between 25 and 30 seats out of the 40 in the State, while the latest NDTV-Indian Express forecast puts the RJD alliance firmly ahead of the NDA. There are difference about the direction of the change as well: Star News finds votes moving towards the NDA in the last few weeks, while the NDTV-Indian Express poll reports the opposite. This time the problem in Bihar is not that of a multi-cornered contest and therefore one can expect some help from the exit polls.

Jharkhand: Pollsters agree that this time the going will be tougher for the ruling BJP than it was in the last elections. The difference is on whether the Congress-JMM alliance will inflict minor or major damage on the BJP, and how many seats the BJP will win.

Maharashtra: Although the differences among the pollsters here have got more publicity than anywhere else, a close look shows that the differences are not as sharp as in Bihar. Everyone, including the confidential poll released by the NCP, agrees that the Congress-NCP alliance will bring it major gains but that some of this would be lost by the unpopularity of the Shinde Government. The differences between the forecasts are about whether the gains will be completely wiped out by the Government's unpopularity or not. The most generous estimates for the Congress-NCP put it around 30, while the least generous place it around 20.

Assam: No one is predicting a repeat of the Congress sweep, but there are very different estimates of how much of a surge has the BJP made in the state. Some forecasts see it gaining only three to four seats, while other forecasts have the BJP overtaking Congress for the first time. The NDTV-Indian Express poll suggests a late surge by the AGP that may damage the BJP. Given the difficulties of carrying out fieldwork in the State, it is risky to rely too much on any forecast.

Punjab: Prevailing political wisdom and some polls suggest that the Congress is going to face its Waterloo in this election, while at least one poll suggests otherwise. Elections are still three weeks away and I would wait for the second round of pre-election opinion polls.

Delhi: The polls and political estimates have exhausted almost all the combinations of how seven seats can be split between the Congress and the BJP, from 5 seats to the Congress to 6 seats for the BJP. But none of these polls has captured the situation after the candidates were announced. Given the media attention focussed on the national capital, one can expect better polls in the next three weeks.

That leaves 25 Lok Sabha seats where the opinion polls offer next to nothing. These include the 11 seats in the Northeastern hill states (including Sikkim), the 6 Union Territories besides Delhi, two seats in Goa and six seats in Jammu and Kashmir. Although India Today and Star News report polls from Jammu and Kashmir, these are pioneering efforts and it may be premature to depend on these indications from a State where survey fieldwork is quite challenging. In these States one has to go by political commonsense and reporting from the ground to gain a sense of what might happen.

That, incidentally, is not to be discounted in the other States as well. Years of observing elections and opinion polls has convinced me that opinion polls are perhaps the best indicator of the big picture. These are better indicators, at any rate, than gossip in the cocktail circles or press clubs. At the same time, polls cannot guide us in anticipating the precise details of the number of seats or the exact outcome in a specific constituency or a region. It is crucial to keep these warnings in mind while reading meaning into the opinion polls.

Printer friendly page
Send this article to Friends by E-Mail

National