Pollsters fail the ‘wisdom of the crowd’


V Pattabhi Ram

In his 2004 book, The Wisdom of Crowds James Surowiecki argues that a random group of independently deciding persons is likely to make predictions better than any single member belonging to that group could have done. Amongst the many anecdotes to buttress his argument, he talks of how the crowd at a county fair accurately guessed the weight of an ox when they averaged the guesses of its individual members (the average was closer to the ox’s true weight than the estimates of most crowd members).

I had a glimpse of this when I was a jury member in an Inter B-School Stock Market Contest. The participants received access to the NSE terminal, gleaned it between 945 am to 11 am, at the end of which they predicted the NIFTY points at the closing bell, namely 345 pm. Yes, they were making their prediction 5 hours in advance. The winner would be the one whose prediction came closest to the actual closing points. If you averaged the several predictions of the 27 teams, the number turned out to be the winner!

Armed with this information, I proceeded to check if collectively the exit pollsters came with an outcome closest to the actual result. I picked the case of UP where the BJP swept the polls, Punjab where the Congress had an early Holi, and Goa that saw a hung assembly.

Let’s begin with Goa.

India TV C Voter 15 18 2 5 40
ABP CSDS 13 19 6 2 40
India News XMRC 10 15 8 7 40
Wisdom of Crowd (AVERAGE) 13 17 5 5 40
ACTUAL 17 13 0 10 40

The ‘crowd’ got the numbers regarding the major alliances right, but they got the names wrong! Also, notice that in the case of each party there was at least one psephologist whose prediction was closer than that of the crowd.

Up next is Uttar Pradesh.


India News XMRC 185 120 90 8 403
ABP Lokniti 170 163 66 4 403
India TV C Voter 161 141 87 14 403
Chanakya 285 88 27 3 403
AXIS 260 100 35 8 403
Wisdom of Crowd (AVERAGE) 212 122 61 8 403
ACTUAL 325 56 19 3 403

Again the ‘crowd’ got the numbers horribly wrong. Chanakya was the best of the bad lot, with their predictions coming closest (howsoever far they were) to the actual!

Finally Punjab.

Chanakya 9 54 54 0 117
AXIS 6 65 45 1 117
India News XMRC 7 55 55 0 117
India TV C Voter 9 45 62 1 117
ABP CSDS 23 51 41 2 117
Wisdom of Crowd (AVERAGE) 11 54 51 1 117
ACTUAL 18 77 20 2 117

Ha, when it came to Punjab, the wisdom of crowd again went for a toss.

My sense is that pollsters can say whatever they like, but their strike rates in the past few years have been pathetic. There was a time, many years ago when Prannoy Roy predicted to a seat the victory of the DMK in the late 1980s and almost to a seat the Congress tally in 1989. That possibly hooked the country to the idea of exit polls. But with over the years, the pollsters making a hash of things, intentionally or otherwise, the idea of poll of polls emerged. And this time around, that too has been hit for a six.

Exit polls can engage us with some tamasha. They would have been valid if your results were to be announced several days after the conclusion of the final phase of polling. Otherwise, it might just as well be that we wait for the real stuff. And in that sense one is terribly happy with the Election Commission for banning opinion polls. Given the divisive polity that we today live in, given the short memory of the public, these opinion polls can have a slant and influence voters. Those who said that voters must have information while they vote should remember that you don’t need information on whom the other person is going to vote for. It’s like copying in the exams.




About Pattabhi Ram

A chartered accountant by profession, a writer by passion and a teacher by accidental choice.
This entry was posted in POLITICS and tagged , , , , , , . Bookmark the permalink.

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