06/18/2014 08:25 EDT | Updated 08/18/2014 05:59 EDT

How 'Likely Voter Models' Leave Out My Generation

How do you make eligible voters disappear completely?

Last week's Ontario election saw top Canadian polling firms trying something different. In an effort to more closely predict the final election result, many made predictions using their 'likely voter model'.

To avoid making inaccurate conclusions about the election, these likely voter models are employed to get a closer read, based on those who will actually vote. Likely voter procedures typically involve asking poll respondents a variety of questions about their interest in the coming election, their past voting behavior, and their intention to vote.

In his recap of the Ontario election, Éric Grenier concludes that polling is "not yet capable of estimating likely turnout with more consistent accuracy than their estimates of support among the entire population." He suggests that, in the future, "should perhaps rely solely on those eligible numbers, until the likely voter models consistently prove their worth."

I'm glad. As many pollsters conceded after this election, their likely voter models underperformed compared to their eligible voter models. We can assume that among those publicly released polls in this election, the majority has not been able to represent natural fallout of the true voter population.

A separate, but equally important issue that arises with the use of these models is the inherent discrimination in excluding eligible voters from the discussion. Likely voters tend to be older, more educated, and have higher incomes than unlikely voters. They are also more likely to be men.

Through the use of weighting and scoring systems, prediction models effectively exclude groups who are unlikely to vote from the conversation surrounding election outcomes. In an increasingly market-oriented party system this is a problem for those left out of the conversation.

The picture that is left disproportionately excludes younger voters. Where youth voter turnout rates are consistently low already, now those fighting for the transactional consumer-voter can easily disregard any consideration of these silent citizens.

With Elections Ontario results indicating an increase in voter turnout overall (indicating that 52.1 per cent of the province's 9.2 million eligible voters cast their ballots), we could probably assume that the proportion of Ontario's younger voters was lower than the average.

Elections Ontario does not produce demographic voter turnout data but federal trends show that among 18 to 24 year olds from Ontario, only an estimated 38 per cent voted in the last federal general election. As that election had a total vote estimate of 57.6 per cent for Ontarians we could probably assume that this was even lower for our provincial election.

In some cases, likely voter models go out of their way to exclude based on age groups that are typically less likely to vote, thereby awarding fewer points to an individual who is younger, making them more likely to be removed from the final results of the model.

When all is said and done, this leaves a much smaller proportion of the total number of likely voters who are between the ages of 18 and 29, compared to the true number of eligible voters. Therefore the models place higher value on the opinions and intentions of older Canadians, who are wealthier, more educated, and more likely to vote right wing.

Why is this a problem for my generation?

As Dr. Paul Kershaw discusses in his work, this generation is not only being left out of the conversation, but we are largely being left out of the policy picture. The latest study of Ontario election platforms from Generation Squeeze finds that the four major party platforms reveal a trend across party lines. They state "the Ontario Liberals, NDP, Greens and Progressive Conservatives all prioritized adding or protecting spending on retirees with more urgency than they propose adapting for younger Ontarians."

This comes as little surprise when voter turnout studies consistently show low results among young people. Naturally, candidates draft campaign platforms to suit likely voters. My concern lies in the potential effect of removing these groups from the conversation altogether. If young people do not see their views, priorities or issues emerging in campaign conversation, this will add a significant barrier to the already daunting task of engaging youth participation.

While supporters such as civic action associations and government work in fostering youth participation, youth are still participating at a lower rate than the average Canadian. As we have seen in past elections, had young voters turned up to vote at the same rate as the rest of the general population we would have seen a different outcome.

It is up to young voters to have our interests and issues be represented at the table. With that in mind, a united group of mobilized youth supporters would be well served by polling data that shows the difference younger Canada could make if it chooses to cast a ballot. Likely voter models have a role to play in election polling (with room for improvement), but will do little service to these groups without the broader context of general population trends.


Photo gallery#WhyIVote Ontario 2014 See Gallery