Monday, September 12, 2011

Confusing Poll Numbers?

Some surveys conducted prior to this weekend's election in Guatemala reported two sets of number. In the first and most frequently reported one, analysts and commentators, including myself, reported support for the candidates in terms of what percentage of votes they received among all candidates on the ballot excluding null, blank and spoiled ballots.

For example, in Siglo XXI's most recent Encuesta IV (the one that lead people to ask why Perez's support dropped significantly during the final month) Perez came in at 44.8%, Baldizon at 22.5%, and Suger at 18.4%. However, these are the results when null votes, didn't vote, and blank vote are excluded.

When you include the the 39% of respondents who didn't vote (18.8%), spoiled their ballots (1.9%), or submitted a blank ballot (8.1%) in the poll, the outcome for the first election was more in doubt than previously recognized. Here, Perez counted 31.9%, Baldizon 16%, and Suger 13.1% support.

If you start from these last numbers and then try to figure out how the first round went the way that it did, it's not as difficult to figure out.

No comments:

Post a Comment