Gujarat: Citizen Forecasts versus the Exit Polls

As some of you know, we started a unique experiment during the Gujarat election. We persuaded some 256 people to share their forecasts of the Gujarat election. Out of the 256 people, some 185 offered seat forecasts for each of the parties. Today, we compare their seat forecasts with the Exit Poll Forecasts.

The gap between Crowdwisdom forecasts and Exit polls is hardly minimal. Does this mean crowd's wisdom is greater than exit polls? The answer is, we have a promising start. We will know this a little more definitely over the next few months.

How are the predictions distributed?

There is a difference between how Crowdwisdom's forecasts are distributed versus how exit polls are distributed. Inspite of exit polls being accurate only about 35-40% of the times, 67% of them are loaded between 90 and 130. The Crowdwisdom forecasts on the other hand are spread a little wider. But that could be a problem as well because the extremes are occupied by partisans whose forecasts may (may not as well) be determined by their biases. We will know over time. One thing is certain, given the distribution, I suspect a minimum of 20% of Crowdforecasters will get it right. I shall announce those names on Monday night.

As social media hands over more power to the citizens, it is time we harness this power more constructively. Thank you all for participating in our first event. More such events will be announced soon. Thanks!

Data and Political Geeks: I will make data available after Monday if you intend to write on this topic. Thanks