The mythical 'science' of predicting elections

Coomi Kapoor Tags : psephologists, pollsters, india election analysis Posted: Friday , May 01, 2009 at 1621 hrs

Related Articles
Most Read Articles

The buck for 26/11 Help the little ones live Imagining Indira Stop the obsession with Pakistan!Poverty of research in India How Quattrocchi walked free

Vajpayee a ‘pseudo moderate’? A canard In celluloid Bollywood actress Isha Koppikar weds in Mumbai What’s indispensable?Brabourne pitch; a 50:50 wicket De Dana Dan Tata plans Nano hybrid cars: Report

Discussion
Blogs

"poked by pitchforks... - By Charles Unfair - By Anon Traveller - By citizen Is iaf craeless - By Dr. Ajay Brabourne stadium co... - By Indian Hate campaign launch... - By Abdulmajeed Khan Advocate What is the credibil... - By Vijay

Sanskrit Festivals Ex-servicemen discriminated Ex-servicemen discriminated Disc Replacement Content

This election, the pollsters uniformly predict the Congress will better the BJP in the numbers game. This fact alone makes me uneasy. Judging by the previous record of pollsters, this really means there is a 50 per cent chance that they could be wrong.

Single Page Format

Previous1 2 3 4

Comments

Post comment

View all Comments (13) |

TRP, TRP, TRP!!!By: S. Jain | 05-May-2009 Reply | Forward It is obvious that the author has had no formal training in advanced statistical methods. There are two clues for this conclusion: (i) The concept of “margin of error” is an in-built prediction error for any statistical exercise of this nature (remember that the poll prediction exercise is based on the responses of a sample consisting of, maybe, 15,000 prospective voters from a total voter population of about 750,000,000 i.e. nearly 1:50,000!); (ii) the author has obfuscated her argument on the science behind this statistics by subtly (brazenly?) suggesting that “doctoring” is an accepted ploy within the profession: no, it is not, and such doctored statistical exercises are, ab initio, void, whatever any dubious proponent of such fudging might say. Besides misunderstanding the inherent limitations of poll predictions, the author also needs to understand why the psephologists’ tribe still thrives on mainstream TV – it’s the bottom-line, dear, eyeballs, TRP, baki sab bakwas, literally.

Which side are you on? By: Ramesh Sahni | 05-May-2009 Reply | Forward Looks like you are one those pseudo-seculars who dot the media and bash the BJP and its constitutents as to the damage they could do if they came to power. Getting back to the point, we cant have a Third Front or Congress come into power this time.......!

Prediction for UPBy: satya | 05-May-2009 Reply | Forward I remember during UP elections, many pollsters went totally wrong in their prediction. No one really predicted Mayawati would get a majority. However, later ndtv did a detailed analysis on why it went wrong and the analysis was very interesting.

The mythical 'science' of predicting electionsBy: Raghbendra Jha | 04-May-2009 Reply | Forward Pollsters routinely get their predictions wrong. They should be held accountable for these errors since they set up certain expectations upon which people act and then these expectations are belied. There are real economic costs to such mistakes. Media outlets also have a responsibility to stop airing such unmitigated nonsense. But, of course, accountability in public life is moribund in India and, tragically, in the media as well. Physician, heal thyself!

holier than thou!!By: Pratik Bhandari | 03-May-2009 Reply | Forward I agree with this article. But I am inclined to believe that the NDTV prediction is better than the others. Again, cant say I have any scientific measure of authenticity of prediction either.

Close
Please Login

User ID	*
Password	*

	Forgot Password? click here