Menu Close

Final Federal Election Vote Projections

By Anthony Piscitelli and Porter Oleson

Election day has arrived, the polling firms have completed their polling, the projection sites have projected, and all that is left is to count the votes. On October 15, 2019, ThreeHundredThirtyEight.com summarized the election projections. Today we are updating those projections with the latest numbers. After the election, we will report on how well the projections did at predicting the actual election results.

We examine eight poll aggregators, one analysis-based projection, and two polling firm models. 338 Canada, Calculated Politics, Canadian Election Watch, CBC Poll Tracker, Lean Toss Up, LISPOP, Too Close to Call, Visualized Politics, @EarlWashburn, @repdonsman456, and @politikstcan each aggregate multiple polls from different polling firms to predict the election. Election Prediction Project and Teddy on Politics makes predictions using an analysis process, which examines data from the public alongside general trends. Teddy on Politics and @kylejhutton made similar ‘gut-based projections’ of the election. Ekos, Forum, and Mainstreet are polling firms who use their data to model the election results.

A glance at the projections shows the Liberals will likely lose significant support from the 2015 election. The Conservatives and the Bloc appear to be the most likely to gain from these losses. The NDP will likely lose some support. The Greens seem likely to hold the two seats they held at dissolution and may gain a few seats. The People’s Party of Canada is projected to have one or zero seats. The projections also predicted most independents will lose their seat. Note, at dissolution, the CCF held one seat and five seats were vacant. These are not listed in the table above.

The table above provides a snapshot of the polling results. The first row shows the highest total any projection places a party and the next row the lowest. The average of all projections is listed next. The final row shows the average of the polling aggregators. It’s worth noting that some of the aggregators provide ranges around their projections, which indicate that a large variety of outcomes are possible in an election like this one where the polling data is so close. Please check the linked website above if you are interested in seeing all the range of potential election outcomes.

The various projections models in Canada, with the exception of the Election Prediction Project,  all use some sort of variation on a regional swing model to predict the election results. Essentially this model uses the previous election results and compares it to the polling data in the current election. So, for example, in a basic swing model if in Saskatchewan a party got 10% of the vote last election and polls now have them at 20%, the projected vote total in every riding for the party would be doubled. If another party went from 30% to 20% the projected vote total for that party would drop by one third in every riding. The website ThreeHundredThirtyEight.com, which was run by Éric Grenier before he went to CBC provides a good explanation of the model he used in 2011.

The various projections models in Canada, with the exception of the Election Prediction Project,  all use some sort of variation on a regional swing model to predict the election results. Essentially this model uses the previous election results and compares it to the polling data in the current election. So, for example, in a basic swing model if in Saskatchewan a party got 10% of the vote last election and polls now have them at 20%, the projected vote total in every riding for the party would be doubled. If another party went from 30% to 20% the projected vote total for that party would drop by one third in every riding. The website ThreeHundredThirtyEight.com, which was run by Éric Grenier before he went to CBC provides a good explanation of the model he used in 2011.

Some variations in results are generally a result of variations of how the regional swing model is applied. For example, some models provide bonuses for incumbents, whereas others incorporate riding polls. How the polls are aggregated can lead to variation as well. Forum and Mainstreet make projections using their own polls only. The poll aggregators can use different methods to average the polls. Canadian Election Watch, for example, attempts to predict turnout as part of its aggregation of the polling data. Support totals are manipulated to increase Conservatives +1.5% and Liberals +0.5% and decrease the NDP -1% and Green -1%.  The table above shows that while the different methods yield relatively consistent comparisons there are subtle differences. The Liberals, for example, vary between 31% to 33% when examining the aggregators. Note, not all poll aggregators publish their aggregated totals.

If you would like to make your own projection, please feel free to post it in the comments below or post it in the Google Sheet. Share the seat total for Liberal / Conservative / NDP / Bloc / Green / PPC / Independents and what you project as the turnout for the election. In a future blog post, we plan o evaluate the predictions.

*Note projections will be updated as they are made throughout the day. Projections in Italics are not believed to be the final projections of the aggregator.

*Projections updated 6:50am added Visualized Politics, @kylejhutton, and the second 338 Canada model.

*7:20am updated with Canadian Election Watch and Visualized Politics final projections

*8:45am updated with Calculated Politics final projections and repdonsman456 projection update.

*4:30pm updated with Calculated Politics, @repdonsman456 and Ekos final projections.

Federal Election Projections: Oct. 15, 2019

Poll aggregating websites and individual polling firms use various methods to project who will win the 2019 federal election. The following table presents a summary of the projections from a number of Canadian prognosticators from October 15th, 2019 (or earlier):

LISPOP, Canadian Election Watch, Too Close to Call, 338 Canada, CBC, @politikstcan, and @repdonsman456 each use aggregating models to predict the election. While Forum and Mainstreet use their own polling data to model the election.

Liberal projections range from a high of 168 to a low of 125. (Note the 168 projection was projected from October 1st). Conservative support ranges from 127 to 150. All of these projections mean a minority government. Four of the nine projections predict the Liberals will have the most seats and five of nine projections predict the Conservatives will end up with the most seats.

The NDP projections range between a low of 24 and a high of 34 seats. The Bloc projections range between 15 and 36 seats. The Greens are projected between three and seven seats. Finally, the PPC is projected to win zero or one seat. Five of nine projections also predict an independent (presumably Jody Wilson Raybould).

It should be noted while these projections are the predictions published, many of the sites also put confidence intervals suggesting wider ranges of possible outcomes.

This is our first post summarizing election projections. If you know of a projection we missed please post it in the comments below or use the contact form to share it.

Region of Waterloo Residents Priorities 2019

By Laura Krizan

The results of this poll were based on an interactive voice response survey conducted Friday March 15thand Monday March 17th, 2019. A total of 1003 individuals completed the first question of the survey and 715 completed the entire survey. The survey was designed to aid in the development of the Region of Waterloo’s Strategic Plan for 2019-2023. The Strategic Plan helps Council and staff set priorities and achieve goals, all while keeping the community’s concerns and suggestions in mind. A significant proportion of the Strategic Plan includes hearing input from the public and listening to comments, concerns, etc. so that the Region can set appropriate priorities. The questions in the survey are also aimed to help the Region of Waterloo during the drafting of its Strategic Plan in the future.

The Strategic Plan has 5 focus areas: Thriving Economy, Sustainable Transportation, Environment and Sustainable Growth, Healthy, Safe, and Inclusive Communities, and Responsive and Engaging Government Services. The questions that were administered as a part of this survey were developed by ensuring that these focus areas were kept in mind.

The first question on the survey asked participants about the level of confidence they have in their local government. The results are shown in the chart below, indicating that a majority of citizens (56%) are confident in the regional government to some degree (including somewhat confident, confident, and very confident levels).

The survey was a way to analyze the top priorities that need to be set by the Region of Waterloo for the development of their Strategic Plan. Respondents were asked what they think the biggest priority in Waterloo Region is that the regional government should address.  The results, as shown in the chart below, have been ranked based on the number of respondents choosing a given category as their top priority. The top 3 priorities are: 1) Supporting the development of affordable housing 2) Managing growth, 3) Protecting the environment. 

A significant component of the Strategic Plan focuses on the services that are delivered by the Region of Waterloo, such as public transportation, waste collection, and so forth. In order to better understand the preferences among citizens living in the Region of Waterloo in relation to the delivery of services, the survey asked: “Regional Government must balance the cost of delivering services with taxation. Which of the following would you most prefer for property taxes in Waterloo Region?” 

Results indicated that 19% preferred increasing taxes to improve services while 14% preferred having property taxes decreased. 23% preferred keeping taxes that same and possibly reducing services. The largest proportion (44%) preferred having taxes increased with the rate of inflation and maintaining current services.

This survey was also used to analyze the best ways and platforms to receive public input in the future. Respondents were asked, “If the Region of Waterloo wanted to gather public input or engage you on major issues or decisions, what are the best ways?” It was found that the best ways to gather public input or engage on major issues/decisions in the region are: 1) Online Survey, 2) Social Media, and 3) Telephone Survey. All other options that were included in the survey are listed below.

Ultimately, the survey helped to provide the Region of Waterloo with important information that can be used during the development of the 2019-2023 Strategic Plan. A total of 9 questions were administered, yet the responses that have been analyzed above highlight the most critical results that will be taken into consideration by the Region.

Survey Details

The Interactive Voice Response (IVR) survey was conducted by Laura Krizan; Abby Schlueter; Andrea Volford, and Professor Anthony Piscitelli on March 15th and March 17th, 2019. Throughout the development of the survey, the students worked alongside Lorie Fioze, Manager of Strategic Planning and Strategic Initiatives. The questions that were formulated for the survey focused on supporting the development of the Region of Waterloo’s strategic plan. The survey was funded by the Region of Waterloo to support this initiative. 

Sampling Approach

The sample size was created by randomly selecting Waterloo Region landlines listed in a digital phone book. A sample of likely cellphone numbers was added by randomly selecting phone numbers that were originally assigned to Waterloo Region, according to the Canadian Numbering Administrator. Sampling errors exists as a result of this approach due to the mismatch created by the random dialling of phone numbers as opposed to randomly sampling actual Waterloo Region residents.

Response rate

The survey called 46,912 live lines. The response rate was 1.5%, which is based on 715 respondents who completed the entire survey. All 788 respondents who answered the first three questions were included in the results. It is worth noting that 215 (21%) respondents were not eligible to participate due to being under 18 or not living in Waterloo Region.

Weights

Results of this survey have been weighted by age, gender, and city/township according to the 2016 census. The full weights are posted along with the raw data on OpenIcpsr.org and can be found by visiting: http://doi.org/10.3886/E110225V1  

Margin of Error

Results are considered accurate +/-3.7%, 19 times out of 20. The margin of error on subsamples is higher.

Raw Data

Raw survey data is available on OpenIcpsr.org. The data can be found at: http://doi.org/10.3886/E110225V1  

Disclaimer

The survey results will exhibit sampling error as a result of the mismatch created by the random dialling of phone numbers as opposed to randomly sampling actual Waterloo Region residents. This survey was approved by the Conestoga College Research Ethics Board.