Opinion Polls add to the suspense in crucial #BiharPolls

After dates to Bihar polls have been announced, debate has heated up as to who will win these crucial polls wherein career / prospects / prestige of many leaders like Nitish, Lalu, Modi, Amit Shah are at stake. While Janata alliance has the arithmetic, elections are not all about arithmetic but also about chemistry (my favourite quote). This was repeated and referred to by C Voter founder Yashwant Deshmukh when he released his pre-poll survey results this week “This election is a contest between arithmetic and chemistry.”

Needless to say, NDA faces seat sharing differences, not yet finalized and their alliance chemistry can’t be taken for granted. Advantage for BJP is that it is in power at the center and can accommodate allies at the center to placate disgruntled partners. Kushwaha and Paswan are cabinet ministers already and BJP may give cabinet post to Manjhi after state polls in return for agreeing to fight on less number of seats.

As soon as the dates were announced, two opinion poll survey results were published – India Today / Cicero and India TV / C Voter. Both the polls show that there is neck to neck fight, but both of them predict different winners. India Today projects a NDA win while India TV predicts a Janata win (both simple majority). The difference in vote share between the two alliances is 2%-3% in both the polls.

For the first time in Bihar state elections history two alliances are expected to get >=40% vote share. The last time any party / alliance got 40%+ vote share was in 1977 when Janata Party defeated Congress in Lok Sabha as well as many states including Bihar.


Heartening news for NDA from both polls is that while it is expected to increase its vote share to 40%-42% from 39% in Lok Sabha, Janata led alliance is seen losing vote share (down from 46% in Lok Sabha to 40%-42% in the 2 polls). So the ARITHMETIC is not working clearly.

I am not a big fan of opinion polls (don’t really understand how such a small sample can actually accurately predict the results). Nevertheless they have become an essential component of our electoral process and their calls have sometimes been right and sometimes wrong.

Bihar has 6.4 crores of voters , out of which 3.6 crore voters cast their vote in Lok Sabha elections in May 2014. India Today sample size is 0.017% and India TV sample size is 0.03% of total votes cast in Lok Sabha. May be a bit too small for a big diverse state like Bihar. To note the average sample size of Delhi opinion polls held in beginning of year was 0.078%, more than double the sample size employed in Bihar. In terms of representation, not only does India TV have more respondents but it also covers all constituencies (while India Today covers only 1/3rd) which is a big plus for me.

There is a huge difference between preferred CM candidate – Nitish leads the pack but is top choice of only 29% (India Today) vs 53% in C Voter poll. One of the reasons could be multiple choices were offered to respondents in Cicero vs limited options in C Voter. The 2nd choice Sushil Modi gets similar votes in both polls.

There is a big contradiction in C Voter poll, while 53% respondents prefer Nitish as CM candidate, 52% wanted a change in Chief Minister and 56% respondents said that the present elections will be a vote for change. This puzzling situation was also evident in opinion polls for Delhi at the beginning of the year when initial polls showed Arvind Kejriwal as most preferred CM while predicting BJP to win the polls.

A back of the hand calculation shows data consistency of India Today polls. Urban population is 11.3% of Bihar population. 48% urban & 41% rural population expected to vote for NDA translates into overall vote share of 42% as shown in the polls.

India TV has predicted seats in a range: Janata 116-132, NDA 94-110 and Others 13-21. You have to apply mid-point theorems to understand the result. Of course it increases the polling agency chances of calling the seats right. However, it does leave a scope for many permutations and combinations.

One such possibility is Janata 116, NDA 110 and Others 17. Though agree that momentum will be with Janata as largest pre-poll alliance, given the acrimony between Nitish-Modi, BJP may not give up easily. If it is able to garner support from 12 of the 17 MLAs, it will be able to form the government.

One last thing, the latest track record of these polling agencies. In the last assembly elections, India Today – Cicerco got it right when they projected Aam Aadmi Party win in Delhi. India TV – C Voter got it wrong as they projected BJP to win.

So which of these polls will be correct? India TV – C Voter which has a big sample size and covered all constituents or India Today – Cicero which got it right last time in Delhi. Only time will tell.

One common theme (despite both polls putting their neck out and projecting winners) is that the elections are damn close. Anything can happen and it will all boil down to each seat, which party in the alliance gets the ticket, number of independent candidates, rebel candidates, seamless transfer of votes across alliance partners and the smaller parties / spoilers affect which alliance more. Toh picture abhi baaki hai doston…


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s