Now the Wait Begins

The long drawn Indian General Elections for the 16th Lok Sabha finally drew to a close today. The nail biting wait for the final results starts today to end on Friday. In the meantime, every news channel in India is happily belting out their combinations of numbers and guessing out loud which set of parties will win. Strangely, many western media including the CNN and the BBC insist that these numbers are extremely arbitrary and rarely accurate.

So, I did some background reading on exit polls and when people did get it closely accurate enough. Several poll pundits did a decent job of sampling the polls during the 90s and then a strange phenomenon happened. These fairly accurate polls started to become very strangely off kilter and the blame was laid at the door of lack of better sampling populations. Actually, with increased prevalence of cell phones, I would think that (intuitively of course) it should be the other way- there needs to be greater accuracy in predicting results as more data samples can be collected effortlessly.

Then I looked at what other parameters changed from the 90s to the 2000s when every poll indicated a BJP led victory only to be smashed by a "surprising" Congress led victory- and the margins of error were considerable but not so considerable to be questioned of fraud. This is an interesting website that has some clear data on the past 4 major survey conducted. Read the paragraph right after the 2009 polls that states how the Congress knew they would win.

So what changed? Electronic voting machines were introduced in India for the 2004 elections. Suddenly, the polls did not reflect the survey numbers but more widespread sampling occurred with widespread cell phone and computer use. This brings us to the question of what other parameters effect poll results. This is an interesting wiki with lots of info on EVMs in India and their vulnerability towards fraud. Dr. Subramaniam Swamy went to court to demand greater transparency, getting the courts to recommend the VVPAT system ie the Voter Verification Paper Audit Trail. In this election, only 8 of the 543 constituencies will have it in place. I wonder why sensitive constituencies like the Amethi seat contested by Rahul Gandhi and Varanasi contested by Modi and Kejriwal did not get this VVPAT system put in place to ensure fair polling.

Another interesting factor is the polling percentage: Here is the Election Commission of India's figures, charts for the same. 2009 saw less than 60% cast their votes. The elections of the first decade of the 21st century saw lower voter turn out than this year's elections. From my rudimentary understanding of the EVMs, fraud is reduced when more number of people show up and actually cast their ballot. In several urban areas, including my own brother and his wife, they, who live in a proper house, paying property taxes and their electricity and water bills for over a decade, still do not have their votes as their address does not "exist" on the electoral lists!

Now, these machines are "in storage" for 3 whole days. Wonder how many polls will be proved "wrong" yet again and a "surprising" win to an unexpected party be announced on Friday. Until yesterday, AAP was almost "written off"- many impromptu interviews with the common man on the streets in and around Varanasi showed a disdain towards Kejriwal as a person who abdicated his responsibilities in Delhi and came there to contest a seat. Then today, there is a "surge" in reports that AAP may actually win the seat in Varanasi- just media reports with no "common man" interviewed off the streets. It will be very interesting to not just see what the "pattern" of votes casted in Varanasi was, but also see "when" they were cast in each machine and if the number of votes tally with the number of people entering the booth to cast their votes. Of course if it was the case like this Simpson's video, we are doomed to more false predictions! Given the highly partisan attitude exhibited by the ECI during the past few weeks, they have certainly lost my vote of confidence.