Thursday, November 7, 2013

Congress right in shunning "fraud" opinion poll discussions on TV

Everyone knows these opinion polls are a sham. They are more “entertainment” than “science”. They are almost always off the mark, and that too, by a huge margin. In the rare instance they are not, it is simply because of the generosity of probabilities; something has to come good, sometime! The fact is that most of these polls are conducted by news channels, which are so strapped for funds, and so desperate to increase their viewership, that they have all the motivation they need to make the research literally “quick and dirty”!

After all, look at the poll forecasts made in 1999, 2004 and 2009. How many of them came even close to being right? None! (My favorite is “The Mood of the Nation” poll that India Today conducts… has got it exactly wrong in each of the above years! That too when its sample size is much higher than what TV channels deploy these days!). And this, at the “aggregate” level where statistical errors tend to “average” themselves out. At the state level, it becomes more complicated as sample sizes drop dramatically. At a constituency level, it is almost impossible to predict outcomes. And imagine the complication in this entire business considering that seats are contested at the constituency level, not state or national. Everything depends on the spread of the lead. If a party leads by just 1% at the state level, and this lead is uniform in all the constituencies, it will win ALL the seats in the state. However, if the lead is concentrated in a few constituencies, that 1% lead may actually make the party second in the state. That’s why there is a usually a small gap in voting % for the top two parties, but a much larger seats difference between them.

Why do media outlets do such a shoddy job? For one, there is no law to regulate such polls. The only thing that matters is that it should “appear” to be OK. That is why a sample size of 2-3000 is considered adequate for a full state. Then there is the economics to be considered. A sample size of this order makes the cost of the research – maybe around Rs 20 lacs or so “affordable”. Besides, I suspect TV channels actually prefer the smaller sample sizes…..the “froth” they generate makes for more sensationalist points to make on-air! For eg., the AAP has suddenly increased its chances in Delhi….that’s great TV content, truth be damned!

Any researcher will advise that the closer the contest, the more the requirement of sample size. In most states going to the polls now, the gap in voting % appears to be tiny (In Bastar for example, the win margins are as low as 100-200 votes in many seats). A larger sample size is also required when there is more heterogeneity. It’s well established that state elections are fought on micro-local issues, and these change as often as the dialect changes – every few kms. Yet, our polls scarcely recognize this fact.

Further, there is another reason the Congress should boycott poll discussions. Just look at the kind of panels these channels put together for the deb.tes Here is how it typically goes. The news anchor usually has a strong anti-establishment viewpoint, because being neutral is considered too “soft”. The other panelists are ALL “involved” parties. In the case of journos, MOST are aligned with one party or the other. It’s a matter of easy observation that the likes of Swapan Dasgupta, Ashok Malik, Minhaz Merchant, Madhu Kishwar, Prabhu Chawla and scores more are all aligned with the BJP. Then there are others called “analysts” – professors of prestigious institutes, Presidents of “policy thinktanks”, even noted economists and other “wonks” who are almost all aligned one way or the other. Then there is the other common trick that the anchor deploys. He packs in politicians from all parties without bothering about their strengths in Parliament. A classic composition would be one politician each from the Congress, BJP, SAD, Shiv Sena, AIADMK and BJD. See what I mean?! This is the way the anchor creates bias. Intentionally. Why should the Congress participate in such biased jokes of debates?

In the UK recently, I noticed how Sky News set up a panel. It had one politician each from the Tories, Lib Dems and Labor. That’s 2 for the government and 1 for the opposition, reflecting genuine public voting. Then there was a journalist who was truly non-partisan. And there was a public representative who offered an additional viewpoint. And remember this – a debate requires an environment where each panelist gets to speak his or her point of view. In India, even the anchor doesn’t allow anyone to speak beyond a few seconds – especially if his views clash with the anchor’s!

Incidentally, when the BJP was ruling, it also opposed opinion polls. It also accused media of bias. It also boycotted individual channels. But today, it is taking a high ground! So much for the BJP’s commitment to consistency in values!

One other point. Does our media ask itself how often it gets it wrong on all its conspiracy theories? How many times has it said that the Congress used the CBI to pressurize the SP and BSP? And yet how vitriolic is the attack both these parties pile on the Congress every now and then? It spun the theory that the Raghuram Rajan panel set up to define states as “poor states” afresh was motivated with the desire to win the JD(U) over. Now we hear that the Planning Commission has rejected that report. We heard the media mock the Tendulkar line of poverty. But later, every economist worth his name, including those who write for media, jumped in to defend it. This makes it clear that media is more interested in froth, and less in the truth.

The real truth is that boycotting panel discussions may appear to be the wrong decision…..but in reality it is the right one. These panel discussions are fraud shows; the research findings complete “nonsense”. Such fraud research is whipped up to create even more froth….giving life to a lie that should never have been allowed to be created in the first place….

No comments:

Post a Comment