Political Literacy: How to read the polls
With election campaigns gearing up or set to kick off in several provinces across Canada, people will be taking the time to consider who's poised to secure office, what their policies are, and how to best utilize their voice at the polling station.
Part of that involves understanding the polls; not only do they show who's leading in a given campaign, but which issues are paramount in an electoral race and how each candidate best reflects or fits the associated policies.
In this first article focusing on political literacy, I'll be discussing how people can understand the plethora of polls released ahead of a given election day.
The general picture
It is worth noting that polls do not represent an entire population. Rather, the findings of a poll are based on a far smaller sample of people, usually ranging from 1,000 to 1,500 participants.
Polls or news articles summarizing a poll will show the percent support that a given party/candidate has. Alongside those measures, there will likely be a +/- value; this represents the margin of error that accompanies the findings. This represents the range in which a polls results might vary based on the wider picture; this is relevant in considering the fact that polls do not sample an entire population.
In the case of polls for a provincial/federal election, the percent support for each party will then be converted to a projected seat count. This considers the support for each party, demographic data, as well as results from the last election. For example, CBC's poll tracker for the most recent federal election built their projections using a proportional swing method, which is defined as the projection of election results based on relative changes in support from one election to the next (in this case from 2021 to 2025). The method is transposed for all ridings and all measures of party support, but it does consider individual factors.
Sometimes, polls will go deeper and break down whether a party is ahead or behind in a given riding, and what the projected margin of victory looks like.
Polls and associated news pieces can also indicate or focus on which issues will be at the front of mind of voters during a given election, sometimes considering the findings alongside the way the electorate is feeling. For example, a recent Ipsos poll indicated that 85% of voters in the upcoming Toronto municipal election are concered about being able to afford a house in the area. In addition, almost 70% of respondents said that the issue of housing would influence which way they vote in the upcoming election.
Can polls end up being wrong?
Absolutely. Though not necessarily all too frequently, polls can sometimes miss the mark when it comes to final election results.
In fact, this happened just recently in the case of an Alberta by-election. Specifically, the provincial NDP pulled off a major upset in a by-election in the riding of Calgary-Shaw, one that typically goes towards the right side of the political spectrum. That same riding saw the UCP, the current governing party in Alberta, hold a 20% polling lead in that riding back in June. Extend the data back to mid-April, and one can see that the UCP was polling 25% ahead of the NDP in that riding.
Comments