A School Experiment: Do Exit Polls Fail?

 

 

Nowadays, during my morning walks, a circle of seniors like us gathers in the park. While discussing the Bengal election surveys, everyone was of one opinion—most exit polls fail.

One of us was a schoolteacher. He said, "I once conducted an experiment in class to explain why exit polls fail." When the Haryana election results turned out to be the opposite of the exit poll predictions, the students had questioned it. The teacher decided to give them a live demonstration.

The Experiment: There were 42 students sitting in four rows. The teacher called two students forward. One was asked to create a table on the blackboard with headings: 'Exit Poll', 'Election Result', 'Party A', and 'Party B'. The second student prepared 40 ballot papers from sheets provided by the teacher. The students recorded their votes, folded them, and handed them to the teacher.

The teacher then asked a student to pick 10 ballots at random—this was the 'Exit Poll'. This sample represented a significant 20% of the total votes. A student opened these ballots and wrote the figures on the board: 4 votes for 'A' and 6 votes for 'B'. According to the exit poll, 'Party B' was going to win with a two-thirds majority.

The Shocking Reality: Then, all the ballots were counted. The actual result was shocking: 'Party A' received 22 votes (55%) and 'Party B' received 18 votes (45%). Despite a massive 20% sample size, the exit poll was completely wrong!

The Takeaway: Now think about this—most states in our country have crores of voters (often exceeding 50 million). If an exit poll agency collects data from 50,000 to 200,000 voters, it amounts to less than a quarter of a percent (0.25%) of the total turnout. No matter how scientific the data collection is, the probability of an exit poll being accurate on such a tiny sample is negligible.

In reality, exit poll agencies merely try to gauge the 'vibe' and make guesses based on that. You might as well call them 'astrologers reading horoscopes'! The students were satisfied. They realized that exit polls are merely a form of entertainment. Occasionally, an agency's prediction might come true by sheer coincidence, but beyond that, it holds no real significance.