According to The Globe and Mail, some pollsters are considering new methodology that would produce two sets of numbers – a traditional survey of eligible voters and a second set of numbers that “gives less weight to groups least likely to vote, such as youth”.

In effect, senior citizens and other groups that are most likely to vote would be given added weight, giving pollsters a better chance of correctly predicting results. Ekos tested such a system before the Ontario election and found that the numbers provided by the “most likely to vote” category were more accurate.

Not everyone is keen on such an idea, however. Nick Nanos of Nanos Research argued that there was no need to release two different sets of numbers because young people are “quite upfront” about telling pollsters that they do not plan to vote. Simply asking people how likely they are to vote would address the problem, he said.