Research centre predicts upcoming election

Graphic by Alan Li

The Laurier Institute for the Study of Public Opinion and Policy (LISPOP), is a research initiative about the demonstration of public opinion in Canadian legislative policy process.

LISPOP is a research centre at the Wilfrid Laurier University Waterloo campus which focuses on issues pertaining to the use and demonstration of public opinion in Canada’s legislative policy process.

The organization is comprised of professors and doctorates from Laurier and a variety of other Canadian universities and institutions.

LISPOP has recently published seat projections for the upcoming provincial election. Barry Kay, associate professor of political science at Laurier and veteran in this field of study, has conducted four public opinion polls between March 11 and 14, 2018. The outcome from these polls placed the Progressive Conservative Party in the lead with 82 out of 124 seats, a majority government win.

“Prior to the last federal election, it’s been accurate over the 15 previous elections within four seats per party, per election, on average,” Kay said.

The provincial projection, made by Kay, is based on data collected from four different forms of public polls; Campaign Research, Ipsos, Leger and The Forum Poll, which collectively gathered opinion about favourable provincial party from 4,300 provincial voters.

Kay’s “regional swing model” is explained more in-depth in his paper “A Regional Swing Model for Converting Canadian Popular Vote into Parliamentary Seats 1963-2008,” which is made available through LISPOP’s website.

“It’s accuracy is less meaningfully tested by public opinion polls because we don’t know, and we will never know, what public opinion really is,” Kay said.

“The only time we really know what public opinion is, is election day itself.”

Kay has been making projections since his time as an undergraduate student in the 1960’s, when he developed his formula to begin making projections for federal elections.

He furthered his study in the 1980’s when he began forming projections for Ontario elections.

His projections are “based on the notion of swing from the previous election to an estimate of the current election,” Kay added.

Kay’s “regional swing model” is explained more in-depth in his paper “A Regional Swing Model for Converting Canadian Popular Vote into Parliamentary Seats 1963-2008,” which is made available through LISPOP’s website.

“I’m reasonably comfortable that [the polls] were accurate,”  Kay said.

“What could probably change [the projection] are mistakes by Ford – the current leader of the Progressive Conservative Party of Ontario.”

“Ford, I believe, is accident prone”.

Kay stresses that these are projections, not predictions. His method of inferencing the fluctuation of public opinion through polls does not intend to predict the future of government, but rather study the immediate representation of public opinion through polls as it concerns elections, both federally and provincially.

“Things could change between today and tomorrow, and they could certainly change over the next two-and-a-half months,” Kay said.

Leave a Reply