Exit polls are a tricky business
On the disparity in exit poll figures on various channels and their implications.
A lot of people don’t believe exit polls to be accurate predictions of the actual results. Critics point to the 2004 and 2009 elections as examples where polls underestimated the performance of the UPA coalition.
Vozag, a quantitative research firm did some number crunching of the recent exit poll data and here are some key takeaways.
1. 282 seats were the median projected seats across all polling firms with Chanakya at 354 seats and Times Now at 249 seats. A difference of 105 seats which both sides can point to as examples of unreliability.
Projected number of seats for NDA by polling firms
2. Even excluding Chanakya from the analysis, there is a difference of 60 seats between the best and worst estimates for the six biggest states for NDA/BJP. That is another arrow in the quiver for exit poll critics.
- Sum of Big 6 Best Polling for NDA= 199
- Sum of Big 6 Worst Polling for NDA= 139
3. The lowest estimate of 249 seats for the NDA/BJP by Times Now is driven by very low projections for BJP in Rajasthan and Madhya Pradesh, which most people on-ground expect BJP to sweep.
Times Now versus Median Estimate in 6 Big States for BJP/NDA
4. People are calling Chanakya’s estimates a “crazy” outlier but they have a very good track record ,especially in the recently concluded assembly elections in November 2013. Take them lightly at your own risk.
Estimate of Chanakya versus Actuals versus Other Polling Firms – Nov 2013
1) A 1% positive change in Maharashtra between the last opinion poll and the current exit poll shows a gain of 8 seats for the NDA
2) Similarly, a 2% positive change in Karnataka leads to a 4 seat difference for the BJP.