This is neither an exit poll nor an opinion poll. It was not done by taking any sample or meeting any voters on the ground. We have taken the data from the past, observed the trends of voting and percentage of votes at the first level. Then we looked at the candidates contesting, which party they belonged to before, as the second level. Then the third level is the campaigning done by top leaders of BJP and Congress and how they were able to swing the voters. All these are based on my observations and from the news reports.
The methodology adopted for each state is different and it is as follows
1. Telangana: We have only one election ie 2014 data which is relevant as the state was carved out from AP in 2014 and the political equations before were very different and hence irrelevant. Here we looked at every constituency and the votes polled for TRS and compared it with INC-TDP. We have not just taken a mathematical value but we based it on how many votes of TDP can be perceptually transferred to the Congress. For every constituency it is different. We have also taken into the note of the fact that the desertions that happened in TDP, INC and YSRCP and all of them went to TRS and how it can impact on the ground level. Many were given tickets in this elections also. So, we looked at the winnability of a candidate from that angle also. and projected individually for all the 119 seats.
2. Chhattisgarh: Here also we have gone constituency wise and we looked at the
3. Madhya Pradesh: Here we have gone region wise. We have divided Madhya Pradesh into 5 regions of Malwa, Bhopal, Mahakaushal, Vindhya Pradesh and Chambal and we have subdivided these 5 regions based on the 52 districts in MP. We looked at the data of 2008 and 2013 elections and have assumed and made projections on the fact that if there is a uniform 7% swing against BJP in this election from the 45% vote they got in 2013 what will happen to each region. We have averaged 2008 and 2013 and on that average, we took a 7% swing to project for the 230 seats.
4. Rajasthan: This is the toughest as no party got elected twice. so, taking 2008 and 2013 won’t be sufficient. So we went back and took 1998, 2003, 2008 and 2013 data which has 2 congress wins and 2 BJP wins and projected for 2018. We have divided Rajasthan into 7 regions of Bikaner, Shekhawati, Hadoti, Vagad, Mewar, Marwar and Dhundhar. There were many factors like Rajput anger on BJP, Anger against Vasundhara but positive perception of Modi, traditional BJP and INC strongholds that were considered. Like MP, for Rajasthan also we have gone for district wise predictions for all the 7 regions covering the 200 seats.
All this was done for an educational purpose and by combining analytics with ground level incidents and not taking any voter opinion. So, the accuracy entirely depends on how well we were able to capture the mood of the past voter and mixed with the present political dynamics and project the voting trends and converting them into seats.