CEPR - Center for Economic and Policy Research

Multimedia

En Español

Em Português

Other Languages

Home Publications Reports Adjusting for Polling Biases in Venezuela’s 2012 Presidential Election

Adjusting for Polling Biases in Venezuela’s 2012 Presidential Election

October 2012, David Rosnick

This paper uses polling data from the various polling firms in Venezuela over the years 2004-2010 to adjust the most recent polling data for polling firm bias in past elections. The regression results show that the average lead of President Hugo Chávez over challenger Henrique Capriles moves from 11.7 percentage points (unadjusted) to a 13.7 percent (adjusted) lead. This would give Capriles a 5.7 percent probability of winning the election.

The adjustment employs a constrained regression model to take into account the bias of participation of the firm in each of the election years that make up the data set (2004, 2006, 2007, 2009, and 2010). The results are robust to excluding any one of the years from the constraint, with the probability of an opposition win ranging from 3.1 to 9 percent, depending on which year is not part of the constraint.

Report - PDFpdf_small | Flashflash_small

Press Release

Report Images (Click for larger versions)

venezuela-polling-2012-10-fig1.pngvenezuela-polling-2012-10-fig2.pngvenezuela-polling-2012-10-fig3.pngvenezuela-polling-2012-10-fig4.png

venezuela-polling-2012-10-fig5.pngvenezuela-polling-2012-10-fig6.png

 

CEPR.net
donate_new
Combined Federal Campaign #79613