Thursday 18 April 2013

Polls: the Russians were right

Well, I thought we would clearly reduce the difference with officialdom but I didn't think it could get that close. Before the election day Juan Nagel sent me a link about a Russian research centre that had produced forecasts based on social media activity. I read the reference article and some others I could find (here and here and here, in Russian, a tiny bit here in English), I found the whole idea interesting but I just couldn't believe those numbers. I thought: well, Venezuelans abroad and those in urban centres tend to be more active in the web and be more against the government, so we might be over-represented, I should know, I am Venezuelan. But lo and behold: the Russians were right and I was wrong.
More than counting birds

The Russians - a group of social scientists from ВЦИОМ (Vtsiom), the VOX Populi Fund and the company PalitrumLab - gave as forecast  around 50.7% for Maduro and 48.1% for Capriles, which was much closer than anyone had predicted in Venezuela. Their analysis was based on data collected from 23 March to 11 April.

I want to see how PalitrumLab managed to get the analysis, but I suppose they have their company secrets. Still, it would be kind of cool to get more information from whatever is available. They had apparently used the same method to predict results for the French presidential election.


1 comment:

  1. This may be valid. We can see a difference from urban to rural voting patterns that can mirror some of this predictions. on the other hand those rural isolated outpost are the more likely ones to be targets for fraud. The statistical modeling would be very interesting to study.

    ReplyDelete

1) Try to be constructive and creative. The main goal of this blog is not to bash but to propose ideas and, when needed, to denounce
2) Do not use offensive language
3) Bear in mind that your comments can be edited or deleted at the blogger's sole discretion
4) If your comment would link back to a site promoting hatred of ethnic groups, nations, religions or the like, don't bother commenting here.
5) Read point 4 again