The Sociological Research Center releases its pre-election survey for Andalusia this Friday, where citizens will vote on May 17. With 109 seats at stake, the poll will offer a detailed estimate of the parliamentary distribution just as the official campaign is about to begin. The data will serve as a thermometer of voting intentions.
How vote estimation is modeled in multiparty systems 📊
The CIS applies logistic regression models and post-stratification techniques to adjust raw data to Andalusian demographics. Variables such as vote recall, direct intention, and partisan sympathy are cross-referenced, using algorithms that correct the overrepresentation of certain profiles. The typical margin of error is around 2.5% for regional estimates, although the dispersion of votes across 109 seats requires fine statistical treatment.
The survey that will make some happy and others cry ðŸŽ
At this point, the parties already have their victory headlines and excuses for failure ready. The CIS will publish its poll and, as always, some will call it prophecy and others, worthless paper. The fun will come when analysts compare the estimated seats with the real ones and say they already knew it. Or not.