Why is Doug Ford’s PC Party so far ahead at this point in time?
This Campaign Research online poll was conducted between March 12 to 14, 2018 among a panel of 1,637 Ontario voters.
Beyond asking respondents which party the electorate intend to vote for, Campaign Research also asked each respondent to rate the level of importance they place on a set of policy planks. This allows for an analysis of the “stated” importance for each policy plank relative to other policy planks. The diagram below shows a top tier of policy planks that respondents told us were more important relative to the other policy planks.
At the end of the survey, respondents rated each party’s leader “as a Candidate for Premier, regardless of who they might vote for”.
Using advanced statistical analytics to evaluate correlations between voters’ leadership performance ratings on policy planks with the “overall” rating they give to each party’s leader, Campaign Research can determine which policy planks are “covertly” more important relative to each other. In other words, Campaign Research’s “derived” importance analysis determines each policy planks level of relative importance. The diagram below illustrates the “derived” importance of each policy plank.
The performance ratings of each leader on each policy plank can be assessed in terms of relative performance. The diagram below clearly shows why Ford’s Progressive Conservative Party of Ontario (“PCs”) has a significant advantage over Wynne’s Liberal Party (“OLP”) and Horwath’s New Democratic Party (“ONDP”) on key drivers.
Of the policy planks that “covertly” drive party leader ratings, Doug Ford outperforms the other leaders on the (top) four most important policy planks. Of the policy planks that are the least important, relative to each other, Kathleen Wynne outperforms the other leaders in those (bottom) four planks. Andrea Horwath has some traction on the (top) two most important policy planks and outperforms her opponents on the policy planks that are of average importance (middle) compared to others.
*See a more detailed explanation of Relative Leadership Strength at the end of this post
“The PCs have a significant lead over both the OLP and the ONDP. This is because the policy issues that matter the most to the electorate also happen to be the policy planks that Doug Ford is seen to be performing much better on. If Doug Ford and the PCs remain focused on these policy planks, the PCs could hold onto their lead. If the ONDP is able to focus the public’s attention on issues where Andrea Horwath outperforms her opponents, they will likely be able to improve their standing. Kathleen Wynne and the OLP are outperforming in a significant way on some of the policy planks, but at this point those policy planks are not seen as being as important.” – said Eli Yufest, CEO of Campaign Research Inc. Eli can be reached at firstname.lastname@example.org or (647) 931-4025 x 109
This online poll was conducted by Campaign Research as part of its monthly omnibus study between March 12 to 14 2018, The study was conducted among a random sample from an online panel of 1,637 Ontario voters whose incentives for participation were handled by the panel provider and who were selected to reflect Ontario’s age, gender and regional distributions in line with 2016 Statistics Canada census data.
A probability sample of this size would have a margin of error of plus or minus 2.4%, 19 out of 20 times. Data was weighted by age, gender and region in the Province of Ontario according to 2016 Statistics Canada census data. If you require more information, please contact us as it is available upon request.
The following screening question was asked in order to determine eligibility for participation in the study
"Are you 18 years of age or older and eligible to vote in federal elections?"
*What is Relative Leadership Strength analysis?
Absolute levels of leader performance only tell part of the story represented in a table of leader performance scores. To fully understand how leaders, compare to one another, Campaign Research goes deeper and looks at relative performance. Correspondence Analysis maps are one way to explore relative performance, but maps may suffer from difficulties in the interpretation. Relative Leadership Strength scores are based on an analysis closely related to Correspondence Analysis and are in fact derived from the same aggregate data table used in Correspondence Analysis.
In the analysis of a “leader-by-policy plank” performance summary table, Relative Leadership Strength scores highlight areas where leaders over- or under-index on key policy planks relative to the competition. As such, Relative Leadership Strength scores characterize a leader's relative strengths and weaknesses in a highly relevant and contextualized context. In this sense, Relative Leadership Strength scoring is similar in principle to double-normalization in that it accounts both for leader performance across the policy planks and ratings across leaders. Technically, Relative Leadership Strength scores are adjusted standardized residuals based on a chi-square analysis of the leader/policy plank performance summary table.
Statistically, Relative Leadership Strength scores function as a set of context-specific Z-scores that are useful for evaluating relative leader performance. Interpreted as Z-scores, the Relative Leadership Strength values can be easily stat-tested, colour-coded, or conditionally formatted to provide a quick visual assessment of relative performance across leaders and policy planks. For example, given the Z-score interpretation, a positive Relative Leadership Strength score of 2 indicates a leader-plank performance rating that is effectively 2 standard deviations above the expected level of association for that policy plank given all leaders and policy planks in the data table.
In general, Relative Leadership Strength scores tend to dovetail very nicely with the major features shown on a Correspondence Analysis map. The two analyses have close connections based on the chi-square statistic. In many cases, Relative Leadership Strength scores will highlight features that aren't obvious (or even apparent) when looking at a Correspondence Analysis map in isolation.