If imitation is the best form of flattery, then what the Thuglak magazine did in its issue dated 07 May 2014, must be the best compliments ever paid to this blog. It is a shame that they did not acknowledge the original idea.
Remember the spreadsheet comparison of 3 respected Tamil magazines/daily that was done during the 2011 Tamil Nadu Assembly elections in the Fort series? The point of that exercise in 2011 was to show how confused the election scene was. The spreadsheet plot showed us that none of the surveys could actually predict the results.
3 years down the line, the situation is no different. Tamil Nadu confounded the psephologists for the last four elections (including 2011) and is likely to do so again, in 2014.
The main reason for such a confused scene is the Indian voter in general, and the Tamil Nadu voter in particular, are wary of telegraphing her voting intentions.
The second major reason is the concerns and hopes of the voter across the state are not uniform on similar issues or even on the same issue. For e.g., the whole of Tamil Nadu want Kerala to accept in finality the Supreme Court judgement in Mullaiperiyar Dam issue, and would object to a proposed review petition by Kerala. But (most of) the same Tamil people want the Government of TN to file a review petition against the Supreme Court judgement on Jallikattu ban, delivered the same day.
The third problem is the defects in the attempted surveys themselves. For a reliable and accurate forecast, the sample size needs to be adequate, the sampling rate needs to be frequent, and data collection should be concise. I am afraid that most of the surveys fall short on all 3 points.
To get back to the point, the Thuglak compiled the data from 5 or 6 magazines/dailies and published it in a spread sheet, along with their own projection for 31 seats. They did not attempt any further analysis on them. Worse still they appear to have grabbed slightly older data from the others. On their issues dated 27 April (in my hand on 21 April), both Junior Vikatan and Kumudam Reporter published their forecasts for 40 seats - predicting not only the victors but also the 2nd and 3rd place losers.
Now, tallying up forecasts (seat-by-seat) in a spreadsheet is a novel mad.madrasi idea. Though I cannot claim a copyright on such an idea, or even suggest that Thuglak copied it from me, it would have been better if they had asked me to do it - would have done it for free.
Anyway the botch-up provoked me into running my own spreadsheet analysis. That is on the next post. But for now, have a look at what Thuglak did.
Remember the spreadsheet comparison of 3 respected Tamil magazines/daily that was done during the 2011 Tamil Nadu Assembly elections in the Fort series? The point of that exercise in 2011 was to show how confused the election scene was. The spreadsheet plot showed us that none of the surveys could actually predict the results.
3 years down the line, the situation is no different. Tamil Nadu confounded the psephologists for the last four elections (including 2011) and is likely to do so again, in 2014.
The main reason for such a confused scene is the Indian voter in general, and the Tamil Nadu voter in particular, are wary of telegraphing her voting intentions.
The second major reason is the concerns and hopes of the voter across the state are not uniform on similar issues or even on the same issue. For e.g., the whole of Tamil Nadu want Kerala to accept in finality the Supreme Court judgement in Mullaiperiyar Dam issue, and would object to a proposed review petition by Kerala. But (most of) the same Tamil people want the Government of TN to file a review petition against the Supreme Court judgement on Jallikattu ban, delivered the same day.
The third problem is the defects in the attempted surveys themselves. For a reliable and accurate forecast, the sample size needs to be adequate, the sampling rate needs to be frequent, and data collection should be concise. I am afraid that most of the surveys fall short on all 3 points.
To get back to the point, the Thuglak compiled the data from 5 or 6 magazines/dailies and published it in a spread sheet, along with their own projection for 31 seats. They did not attempt any further analysis on them. Worse still they appear to have grabbed slightly older data from the others. On their issues dated 27 April (in my hand on 21 April), both Junior Vikatan and Kumudam Reporter published their forecasts for 40 seats - predicting not only the victors but also the 2nd and 3rd place losers.
Now, tallying up forecasts (seat-by-seat) in a spreadsheet is a novel mad.madrasi idea. Though I cannot claim a copyright on such an idea, or even suggest that Thuglak copied it from me, it would have been better if they had asked me to do it - would have done it for free.
Anyway the botch-up provoked me into running my own spreadsheet analysis. That is on the next post. But for now, have a look at what Thuglak did.
No comments:
Post a Comment