Remember Al Gore? There is a loose parallel between how some Indian TV channels mucked up the Bihar counting day coverage last Sunday with the embarrassing on-air calls that American TV networks made during the 2000 ‘Bush vs Gore’ presidential polls.
Announcing the results on November 7, 2000, channels first called Florida in favour of Gore, then retracted to call it instead for George W. Bush, and finally after Gore had conceded the presidency, retracted again on the Bush call. NBC’s T im Brokaw said at the time, “We don’t just have egg on our face. We have an omelette.”
Bihar is not Florida, but there is no question that many Indian networks too have ended with an omelette on their face. Amit Shah and Narendra Modi may have lost their 2014 zing in Bihar, but some channels also lost much of their credibility.
There are two different issues here. The first is NDTV wrongly calling the election for the BJP as early as 9 am on counting day. The second is the credibility of exit polls and if they really are more accurate than the proverbial blind-folded monkey throwing darts, as Princeton economist Burton Malkiel once argued about Wall Street, where he claimed random monkey shots “could select a portfolio that would do just as well as one carefully selected by experts”. Is it the same for exit polls?
For calling the election wrong, NDTV co-founder Prannoy Roy, the father of Indian psephology who first made his sterling reputation as a pollster by going against conventional wisdom to predict a sweeping Indira Gandhi return in 1980, has already apologized (We Apologise for Bihar Results Confusion, Says Prannoy Roy). It is a cruel reminder of how fickle a mistress live TV can be even though NDTV was right that pollsters all over the world sometimes get it wrong.
As they did in Britain where the British Polling Council ordered an independent enquiry into how no poll could predict David Cameron’s victory earlier this year. Or how no Israeli poll predicted Benjamin Netanyahu’s latest poll victory. Or how the Greeks and Turks recently got their election ‘results’ wrong as well.
But NDTV’s big blooper was not that their exit poll was wrong. It was that it called the counting trend wrong. The channel blames “incorrect” data. The News Broadcast Standards Authority has reportedly asked Nielsen, the agency that supplies a common feed to all subscriber news channels, for an explanation. The fact is that whatever the data may have been, calling the election so early in the morning, when only postal ballots are counted first, is inexplicable.
In the American case too, the channels had blamed polling agency VNS. Whatever the reason may have been in Delhi’s newsrooms this time, the outcome is the same as the one found by a three-member committee appointed by CNN to investigate its mistaken 2000 election call: “a news disaster”.
While all national TV channels do get election feeds from a common agency, most of them also supplement this with their own information from the ground through agencies, stringers and constituency respondents. These updates are manually punched in to add to the automatic feed. On Sunday, some did this faster and more accurately than others.
CNN-IBN, which was getting updates through its strong ETV network of reporters in every district of Bihar, was fastest on this and called the election right at 10.03 am, as did ABP News and News X. The Times of India’s website, which relied on C-voter also got this right.
In that crazy hour before news channels and experts changed their numbers and analyses in favour of Lalu-Nitish, even NDTV India’s own reporters began talking of ETV numbers on their channel, even as these differed hugely from NDTV’s own on-screen data. As one morning tweeter summed up: “Amit Shah: Watching NDTV. Feeling happy. Nitish Kumar: Watching CNN-IBN. Feeling happy. Rahul Gandhi: Watching Comedy Central. Feeling happy.”
The second issue is the credibility of exit polls. The bizarre defence given by Today’s Chanakya, the only poll that predicted a landslide BJP victory, will do the tribe no favours. The organisation that proudly declares on its website that “we are having the arts of winning elections [sic]” claims that a simple computer coding error meant that the alliances got interchanged in its prediction. In other words, it got the numbers right, only the names of the winners and losers got inter-changed! This is surely a world-first in the history of excuses.
Similarly, CNN-IBN would be ruing its decision not to broadcast the only exit poll that accurately predicted the landslide for the Mahagathbandhan. The agency it hired, Axis, was bang on in its forecast of 50-70 seats for the NDA and 169-183 seats for the Grand Alliance. It was the most extreme example of ‘herding’, where channels choose to hedge their bets by disbelieving their own numbers and going instead with what they may be hearing in the ‘Delhi circuit’.
The word psephology comes from the Greek word ‘psephos’ which means pebbles. The ancient Greeks threw pebbles into urns in their elections. Perhaps, David Butler, the social scientist who in the 1940s invented modern psephology, was right about his advice to politicians on opinion polls: “Believe and disbelieve them all.”
Psephology remains an inexact science, an art-form, even an esoteric game. For news channels though last Sunday, Bihar exposed the need to be more transparent about the robustness of their data-gathering systems.
(A slightly different version of this piece first appeared in ‘Poke Me’, The Economic Times, 13 November 2015).