Is it possible to manipulate a poll? - Macleans.ca

Is it possible to manipulate a poll?

by

Dave Climenhaga suggests the Greens are trying to game the polls in Calgary Centre.

So it is significant – though impossible to criticize – that a Green Party organizer emailed committed supporters a note headed “There is another poll tonight – be sure to pick up,” not long before the latest survey. “Word from Chris Turner’s Head Quarters is that another poll is being conducted at this very moment,” said the email from Green Party Volunteer Co-ordinator Natalie Odd to committed Turner supporters. “Please be sure to pick up any calls your receive this evening!”

The emails were followed up with phone calls to supporters, although the pollster actually appears to have called a day later than the party expected. In addition to such emails and calls, Mr. Turner’s supporters posted similar messages on Facebook and some people distributed the call-display number the polling company was using.

This bit of gamesmanship seems to involve two assumptions: that it’s possible to manipulate a poll and that a good showing in a poll can precipitate a good showing on election day. The sample sizes used so far in Calgary Centre have been relatively small, but I’m not sure what the relative odds are that something like this could be pulled off. I can imagine that poll numbers could influence turnout and the result, but what are the odds that alerting supporters to the possibility of a poll would result in enough people responding to a survey who normally wouldn’t to significantly impact the results of that poll? I invite any and all mathematicians in the crowd to sort that out.