Delhi Polls - Why the Media backed 'Opinion' Polls are Misleading?

It is as if AAP has got hold of some 'Bramhastra'. The party which was predicted to go into an oblivion post Delhi Assembly elections 2015, not 2 months ago, is now suddenly being hailed as a sure shot winner by its all time Cheer Leaders, the Indian mainstream Media. What exactly changed in the last 2 months? What trick did BJP miss? I would try to analyze how much of it is true and trustworthy, but before all of this a short note on who this election is really fought between.

Battle Royale

It will be quite incorrect to say this fight is between BJP and AAP (not because Congress is a non entity here). It would also be incorrect to say that this fight is between 2 individuals Bedi vs Kejriwal or Modi vs AAP or Kejriwal vs Shah etc. In real terms, BJP had lost the fight even before it began. BJP had no answer to AAP's carpet bombing. Thousands of energized AAP volunteers swarmed the city and visited every potential household, they assessed, might vote for them. In contrast BJP was relying on Modi rallies and eventually paid a good price for sitting idle. However their supporters were not silent and campaigned aggressively both in Social media and on Ground. On the other hand, AAP, despite having a strong campaign was not taken well by the voters. They had help, from the Media. Media played one giant role in negating the perception of 'Bhagoda' about Kejriwal. Moments back an 'independent' journalist in ABP news said AAP completed all its promises (free bijli, free water) except Lokpal and when it could not pass Lokpal, it left the govt. due to moral reasons. Now how true is that? Did AAP not leave for Kejriwal's Prime Ministerial ambitions? So in a way it would not be wrong to say that in effect, Delhi elections are between Indian Media and BJP supporters with AAP and BJP acting as the face of these two entities.

Lutyens dominated Media supports AAP as it is their last resort to undermine their nemesis, Modi. Not for no reasons have these opinion polls sprung up, hailing Kejriwal to be heading for a sweeping victory. Let's analyze these polls from an electoral viewpoint.

Story the numbers tell

How did the parties fair in last 2 polls
AS 2013 - BJP 33%, AAP 30%, Congress 25%
LS 2014 - BJP 47%, AAP 33%, Congress 15%

To naked eyes, it may appear that both BJP and AAP gained at the expense of Congress and Others, with BJP gaining more in the Lok Sabha elections, compared to the Assembly. However this is not true. Various sampling experiments reveal that AAP lost a SIGNIFICANT chunk of voters to BJP, especially the educated and/or middle class voters who were fed up with the AAP theatrics. This could be as high as 10% of the voters base. Some of these would also include voters who voted Kejriwal for CM but wanted Modi as PM. So the new figures should have been BJP 43%, AAP 20% and Congress 25% but the story does not end here. BJP would also gain some votes from Congress (~2% due to National Anti Incumbency) and Others (~2% due to BJP being a National party). On the other hand, many Congress floater voters (who were/are anti BJP) would not have voted AAP in Assembly thinking it to be a wastage would have considered AAP as a more viable Anti BJP alternative in Lok Sabha, so as high as 8% of them would migrate to AAP. AAP would also gain 5% from the 'Others' section as historically, these voters were neither fond of BJP nor Congress. So rightfully the number read 47%, 33% and 15%. So it is conclusive that AAP did lose a huge chunk of votes in the Lok Sabha.

The Loyal Voter Base

Each party has a dedicated votebase.
BJP - If one analyzes BJP's voteshare across elections, it has hovered around 33% to 52% in last few terms, and never went below 33% (at the peak of AAP wave) in the last two and a half decades decade. So it is safe to assume, BJP has a loyal vote base of 30% in Delhi

Congress - Congress has degrown since 2009 Lok Sabha elections in Delhi. Their low of 25% in Assembly 2013 was soon eclipsed by a new low of 15% in Lok Sabha 2014. Despite that fact that Congress was a loser in Assembly and a sure shot loser in 2014, 15% still held on to it. However some of these may realize that voting a losing entity is going to be futile and except for a few scattered pockets, Congress' story is all but over this time, so atleast a 1/3rd of the voters, the educated and the less emotional one would desert Congress, pegged at around 5%. So the loyal votebase for Congress is 10%

AAP - There are very few elections to analyze AAP's loyal base but the 2 elections DO give a story. As mentioned earlier, a huge chunk (as high as 10%) deserted AAP in Lok Sabha. However 20% of them still hung on to the party, despite its failures, theatrics and over ambitiousness. So AAP loyal vote base forms of 20%

Others - Small parties like BSP, INLD, SP etc have their pockets of influence in the city. Add to it the independents. The others vote share was around 5% in Lok Sabha and 12% in Assembly, which is understandable given that voters tend to polarize towards known parties in Lok Sabha elections than in Assembly. Also due to the higher number of seats in Assembly (70 to 7), the number of Independents with localized influence increase in an Assembly election. Hence the others would have a Loyal vote base of 10%.

Total Loyal Vote Base = 30% + 10% + 20% + 10% = 70%

The Floater Voter Base

These are the voters who would decide the fate of this election or any election for that matter. In 2013 and 2014 and even before that, these voters have shifted their allegiances across parties. But who are these voters and who are they likely to vote this time and for what reason? Let us try to further segregate these voters into segments and try to analyze which way they would vote.

1. Young Urban women (3%) - Aged between 18 to 35, they are approximately 3% in number. This does not mean that there are just 5% women in Delhi between the age of 18 and 35. This means that those women who do not have any party loyalty are approximately 3%
Who did they vote in 2013? - Possibly AAP
Who will they vote in 2014? - Possibly BJP
Who will they vote in 2015? - Possibly BJP
Driving Reason - Face of Kiran Bedi and issues related to women
Is there a swing compared to 2014 - No

2. Non Anti BJP Minorities (6%) - This group comprises mostly of voters who are not hostile to BJP and can vote to any party which they perceive would raise there issues
Who did they vote in 2013? AAP, BJP, Congress in that order
Who did they vote in 2014? BJP, AAP, Congress in that order
Who will they vote in 2015? AAP BJP Congress in that order
Driving Reason - BJP govt. did some work for them bu not enough to be perceived
Is there a swing compared to 2014 - Yes

3 Rural Settlers population, working class (5%) - This is the group which comprises of rural class (excluding above 2 categories)
Who did they vote in 2013? Congress, BJP, AAP in that order
Who did they vote in 2014? BJP, Congress, AAP in that order
Who will they vote in 2015? BJP, AAP, Congress in that order
Driving Reason - BJP's promise of rehab and growing pro poor image
Is there a swing compared to 2014 - Yes for AAP, No for BJP

4 Poor and weaker sections (11%) - The group of people who represent the poorest of the poor who lack even the basic facilities.
Who did they vote in 2013? AAP, Congress, BJP in that order
Who did they vote in 2014? AAP, BJP, Congress in that order
Who will they vote in 2015? AAP, BJP, Congress in that order
Driving Reason - AAP has been successful in convincing this set of voters and hence will get a major chunk of it
Is there a swing compared to 2014 - No

5 Others (5%) - The remaining population who are mostly in the outer areas of Delhi.
Who did they vote in 2013? BJP, Congress, AAP in that order
Who did they vote in 2014? BJP, Congress, AAP in that order
Who will they vote in 2015? BJP AAP, Congress in that order
Driving Reason - AAP was limited to the city region in 2013 but has now penetrated into the outskirts
Is there a swing compared to 2014 - Yes

The above 5 categories are mutually exclusive among themselves and one is not present in the 2nd. However these 5 categories are not exhaustive and would also be present in the 'loyal' voter base of the respective parties.

The partywise takeway of floater base would be

BJP - 1.5%+2%+2.5%+3%+2.5% = 11.5%
Congress - 0%+1%+0.5%+0.5%+0.5% = 2.5%
AAP - 1.5%+3%+2%+7.5%+2% = 16%

AAP would walk away with the majority of floater vote base as per my analysis, BJP a close 2nd and Congress not in the picture. The total voteshare would hence look like
BJP - 30+11.5 = 41.5%
Congress - 10+2.5 = 12.5%
AAP - 20+16 = 36%
Others - 10%

At a this vote share (BJP 41.5%, AAP, 36%, Congress 12.5%, Others 10%), the seats tally should hover around
BJP - 50-54
AAP - 12-16
Congress - 1-5
Others - 0-2

Hence the Media run Opinion Polls with the story of sudden surge of Kejriwal should be taken with a sack of salt.

To conclude, this is just an individual analysis without backing of any segregated sample survey. So it is possible that the analysis might not hold true during the counting day and results are quite different from what is stated here. However if logic is to be believed, the results should not be much different from what is stated here. What happens eventually will be known that day, for Indian voters have a knack of giving surprising results.