by a Thinker, Sailor, Blogger, Irreverent Guy from Madras

Delhi: Why exit polls can be wrong


Within 12 hours from now, the results of the 5 state elections of 2013 would be out.  The opinion and exit polls till today have been forecasting a rout for the Congress party in 4 states, with a lone win at Mizoram.

Of the 4 other states where Congress is forecast to lose and BJP is to win - Madhya Pradesh, Rajasthan, Chhattisgarh and Delhi - the last one is of interest.  Not because it is the National Capital Region or that the current Congress CM of Delhi Sheila Dikshit would have made 4 in a row, but because of an untested factor - the Aam Aadmi Party (AAP) and how much support it has actually generated.

The exit polls widely vary, and are not consistent.  Inconsistent not because they cannot forecast an outright majority for BJP, but because they widely vary on the number of seats the AAP will win.

Nothing sums up the situation better than the cartoon by Sudhir Tailang on Deccan Chronicle on 6 December.

Here is it.
sudhir-tailang-exit-polls-delhi-2013

Now there is no doubt Delhi has seen the highest voter turnout of around 67% which is 10% more than its usual.  Now, does this higher turnout automatically mean a vote for AAP and against Congress and BJP?

Let us just analyze.
Sitting in Chennai it appears that all the pollsters have assumed that an extra 10% vote will be for AAP and not for BJP or even for Congress.  They also seem to have assumed a 10% swing against the Congress in addition.

Now, that may be or not be.  But to assume that all those 20% have gone for AAP is too much.  In the elections before in 2008, the vote share (seats) has been approximately Congress 40% (43), BJP 36% (23), BSP 14% (2) and Others (2) the rest.  The point to note here is there are 12 reserved seats in Delhi and the BSP with a 14 % vote share, had been able to score in only 2, even though it claims to be a party of Dalits.

Now let us assume all of the additional 10% votes are for AAP.  Let us also assume there is another 10% swing overall *for* the AAP, and away from both the Congress and the BJP.

So even with that the AAP will end up with a vote share of 20%, spread all over Delhi and without any ‘special’ conclaves, earmarked for that party.  Seeing that in 2008 the BSP, a proclaimed Dalit party with a 14 % vote share, with 12 Reserved constituencies could garner only 2 seats, how do the exit polls give so much to AAP in 2013?

Next is the vexing question of the swing.  Here there is confusion.  Assuming there is a 10 % vote (from 2008 votes) *for* the AAP as above, how much of that 10% is anti-Congress and how much is anti-BJP?

If we assume a generous 70-30 split, then the Congress vote share would be down to 33%.  The BJP’s would be down to the same 33%.  So it is still a toss up.

For my 2 paisa worth, I think the Delhi elections in 2013 would be won with narrow leads, perhaps in the low hundreds, and the party which mobilized its sympathizers better on the voting day would see it through.

Of one thing I am sure. It is going to be exciting.  It is going to be tense.  And it is going to give a lot of surprises - even to me!  Remember, I wrote about it first.

No comments:

Post a Comment

Support - Donate

Your Blog is

Donate thro ECWID

Contact Form