Election Polls: Do They Really Tell Us Anything?

As speculation concerning whether or not the Liberals and NDP will topple Harper’s Conservatives in March by voting “non-confidence” in the Government’s proposed budget—a move which would then trigger an election—polls claiming to represent citizen party preference are simultaneously beginning to crop up on news programs and in newspapers nation-wide. 

In one such random poll, executed by EKOS research (released to CBC), statistics suggest that the Conservatives currently have a substantial lead over the Liberals in terms of support, the party up almost 12.5 points. When asked whether these numbers were worrisome, Liberal leader Michael Ignatieff responded, “Polls don’t matter.” Though such an answer is perhaps expected from a leader whose party isn’t ahead in said poll, Prime Minister Harper also brushed off the numbers (which were in his favor!), saying “Polls go up, polls go down.” Harper and Ignatieff’s indifference to polls made me wonder: Do political polls concerning party preference accurately represent the political opinions of Canadians? Furthermore, do they provide Canadians with a true picture of how fellow citizens feel about the government or opposition? 

Personally, I don’t believe public opinion is as tangible or uniform as most might expect. Some consider public opinion to be the opinion of the majority or an aggregation of individual opinions, while others consider public opinion to be media or elite opinion. Certain scholars even argue that public opinion is a fiction, constructed by the media or government to conform society to elite decisions. Regardless of this debate, it should be mentioned that actually measuring public opinion (whether it be fictitious or not) presents its own problems. According to how the information is collected (face-to-face interview, over the phone, internet, mail-back survey, Facebook, etc.), some people may not provide pollsters with their true answers (problem of anonymity, coercion, leading questions) or may change their opinions after the polls have been conducted. Additionally, experts suggest that most “undecided” voters don’t even pick who they are voting for until the day of the vote. These inconsistencies—coupled with the highs and lows of elections (from one week to the next, citizens may change their minds or be swayed by events, advertising, etc.)—seem to render polls fairly useless, especially if their utility is based upon predicting election outcomes or events.

All in all, I think polls attempting to gauge or measure public opinion or party preference can be quite irrelevant (especially before an election has even been called) and the variety of results to be telling of how inaccurate polls can really be (riddled with sampling biases, errors, etc.). If political leaders are skeptical of polls, shouldn’t we be, too?

If the above poll is trying to suggest that the Conservatives are going to win the next Canadian election, I think such an assertion it is way too premature.