Skip to main content
public editor

A recent Globe and Mail survey asked reader if they believed Stephen Harper’s Canada is better, worse or unchanged.The Canadian Press

I've had a few complaints about a recent survey of Globe and Mail readers' views on federal politics asking if they believed Stephen Harper's Canada is better, worse or unchanged.

So what's good and what's bad about this particular survey?

It's good that readers are being engaged on what will be the political story of the year. The survey attracted 138,000 votes, so it's obviously a topic of great interest.

What's the bad? It's not a "poll." The headline on it says "poll," but it is very far from a science-based opinion poll and it hurts the credibility of the real and important polls that news organizations publish, especially during the campaign period.

These things are, in the words of a tweet by Ipsos pollster John Wright, "mob audience feedback with no scientific backbone."

While it attracted 138,000 votes, it is far from clear how many readers gave their views because people could vote multiple times, and often what happens with these questions (on many different topics such as politics or policy issues such as assisted dying), the surveys are taken over by interest groups whose members vote repeatedly and skew the numbers beyond recognition.

One reader told me: "I wish to complain about your poll, which has no credibility whatsoever as it allows multiple voting. The number of votes and huge gap in numbers makes it clear that not only is it being manipulated but it's likely autobots are being used. If you want to have a valid poll, then make it secure."

So what is the value in a survey of readers that is open to manipulation? Very limited, I would say. If the purpose is to engage readers in a conversation, it needs to be set up as the readers' views with clear caveats. (The current one does say the results are unscientific, but it needs more explanation of how the survey actually works.) It could be called a reader forum or reader feedback, which would be more accurate.

Or here is an even better idea: In an election year when real polls are being used, The Globe should consider stopping these reader feedback surveys about important matters such as politics or federal policy.

A survey that asks if you think the Toronto Maple Leafs have a chance of making the playoffs or what you plan to do for Valentine's Day are good fodder for comments and conversation. But greater care needs to be taken around weighty subjects.

Here's another blog I wrote about the importance of real opinion polls during this election year.