Vote Compass live blog
3rd & 7 37yd
3rd & 7 37yd
B
S
O
close
close

-





-
Live blog starts at noon ET
Hello folks. Vote Compass is live! To mark the occasion, I'll be hosting a live blog with Cliff van der Linden, one of the Vote Compass creators, starting around noon ET.You can submit your questions by clicking on the quick login function or logging in a social media platform and asking that way. You can also submit your question here. -
-
Greetings!
Let's begin. Joining me is Cliff van der Linden, the executive director of VoxPopLabs and one of the creators of Vote Compass. It just launched today and you can see it here.We're taking your questions but let's start with the basics.What is vote compass? -
Vote Compass is online application that offers citizens a personalized, informative (and, I daresay, fun) entry point into the policy platforms of the parties vying for election. Answer a series of questions about your views on Ontario politics, and the application calculates your alignment with each of the parties running for election.
-
-
The Vote Compass questions are crafted by the team of political and data scientists at Vox Pop Labs, the makers of Vote Compass, in collaboration with a team of academics from local (in this case, Ontario) universities. They reflect salient topics in the electoral discourse and try to get at key points of differentiation between the party platforms.
-
-
Each of the parties included in Vote Compass is consulted through the calibration process. So, the parties can have a say in how they're represented in Vote Compass. But whether a party leader's personal views differ from the party line—that I can't say as we haven't yet had the fortune of having a party leader reveal their results (although I'm sure some have taken Vote Compass for themselves).
-
Remember, you can ask you questions too!
Join the discussion by commenting here. Or you can respond to us on Twitter. -
-
The parties are positioned on each of the questions included in Vote Compass on the basis of their public disclosures on the issue to date—from public statements, press releases, media interviews, etc. The parties are also given the questions in advance and the opportunity to clarify their positions if they so please. As new information is released that relates to any of the subjects covered in Vote Compass, we update the calibrations accordingly.
-
-
-
Probably "does Vote Compass tell me how to vote?" It's actually not intended to direct people on how to vote, but rather to help them explore the political landscape and their place within it. It's largely about the policies being proposed by the parties—and there are other considerations that go into vote choice (party identification, attitudes towards the leaders, etc.). Vote Compass doesn't capture all of those considerations, nor is it meant to.
-
-
-
We have built Vote Compass and Vote Compass-esque for municipal elections and other political events such as referenda. The issue is that the initiative isn't that scalable as yet. It takes a lot of work in terms of research and analysis to build these applications, so we're limited in how many we can take on at a given time as we're a small team.
-
-
Notice how bunched up the Liberals, NDP and Green are on the compass, with the PCs waaay over in the bottom right (more socially and economically conservative by a huge margin). The compass quite accurately plunked me into the tiny space exactly halfway between the Libs and NDP, which somehow makes me a Green voter? About 60% of the residents of this province likely reside quite close to me in that small l/small c liberal/conservative spectrum. But this year, maybe it's our once every 25 years to hand the reins to the 40% and see what gets broken? This is why we can't have nice things. Kinda makes me miss Patrick Brown, who seemed to be promising not to break everything.
-
-
It's certainly interesting to see how the parties are plotted in Vote Compass, as it changes from election to election. The clustering of the Liberals, NDP, and Greens in one quadrant is, to my mind, a telling illustrating of how political competition operates in this province.
-
-
The best way to respond to any such claim is through transparency. Our methodology is publicly available for review (see http://voxpoplabs.com/votecompass/methodology.pdf) and we have published in peer-reviewed scientific journals on the methods we apply to calibrate Vote Compass (see https://www.tandfonline.com/doi/abs/10.1080/17457289.2016.1268144). The calibrations are thoroughly reviewed by an academic advisory committee and by the parties themselves. We've actually built quite a robust process to control for potential bias.
-
Wouldn't you agree that the selection of the questions being asked inherently reveals bias in the compass?
For example, you have 1 question on corporate taxes, which directly affects everyone in the province yet you have 3 questions on indigenous people which, although important, only directly affects a relatively small segment of the population. -
The calculus the Vote Compass uses to align both users and parties isn't weighted to the number of questions on a particular issue. We train a model using pilot data to calibrate the dimensions in the application, and users have the ability to select their own weights in the results section if they care about certain issues more than others.
-
-
For greater clarity, you can add or remove questions from the Vote Compass survey without the plots on the 2D grid changing substantially, so long as the parties are accurately coded. The purpose of using this measure is specifically to overcome the selection effects that might occur based on question selection.
-
-
The determination of the number of questions is actually informed by user experience rather than our methodology. We've found that 30 items seems to be the optimal point for retaining users. We field pilot surveys in advance of launching Vote Compass to determine which questions resonate most among Ontarians. That's how the final set is selected.
-
-
Just to push the point, even the order of questions can affect the outcome. Asking socially driven questions before fiscally driven questions can produce a different outcome. For example, most people would say they care if asked about the environment. However, if the question follows one about taxation, they may be more inclined to consider the impact of environmental policy on their own personal wealth.
-
That's a good point and one always wants to be cognizant of the potential for order bias or priming. The items in Vote Compass are actually randomly served so that we're able to detect if there is any significant effect based on ordering. We do the same thing in the pilot studies. To date we have not found any significant ordering effects among the final set of questions included in Vote Compass.
-
We're here for another 10 minutes taking your questions
You can leave a comment on the blog directly. Or you can comment here on Twitter and we'll pull it in. My guest is Cliff van der Linden, one of the creators of Vote Compass, which just launched today. Take a look!Ontario Votes 2018 - Vote Compass
Vote Compass is a tool developed by political scientists that calculates how your views compare with those of Ontario's political parties. -
The first question is really about the nature and structure of political competition in Ontario. With respect to the second question, you can actually review your responses vis-à-vis each of the parties in the results section of Vote Compass. So you can see where you are more or less progressive/conservative than each of the parties.
-
-
It's a good question, and our intention is always to include as many parties as we can in Vote Compass. But the parties have to meet certain criteria in order to meet our methodological requirements and certain practical requirements. These criteria are listed in our FAQs, but I'll re-state them here:By default, Vote Compass includes those political parties that were represented within a given jurisdiction's elected body at the conclusion of its previous election. If a political party does not satisfy this first condition, it may also be included if it meets all of the following conditions: a) it is registered under the jurisdiction's elections commission; b) it fields candidates in a majority of ridings; and c) it has a fully developed platform.
-
The fact of the matter is that most smaller parties do not have sufficiently developed platforms for us to calibrate them on all the questions included in Vote Compass. So we could not offer a viable or reliable estimation as to how they're situated in the political landscape.
-
-
It varies. People are very actively engaged in politics are often less surprised by their results than people who are less engaged. But we receive quite a bit of feedback from politicos and non-politicos alike telling us that they've been surprised by their results—but that, when digging into the reasons for their results, learned a lot about where the parties stand on key issues.
-
-
-
-
There's actually very little empirical evidence to support the idea that Vote Compass exclusively determines vote choice. In fact, the evidence is to the contrary based on data from the Canadian Election Study. So, while we hope that Vote Compass is informative, we're confident that it's not determinative.
-
-
We are just about done here. I've got a question: How do you account for changes in party leadership? I'm specifically thinking about the difference between a PC party led by Doug Ford vs. a PC party led by Patrick Brown. What did you have to change when the party went under new leadership, just a relatively short time ago?
-
It's a good question and admittedly made our lives more difficult given that we weren't certain which aspects of the People's Guarantee would make their way into Doug Ford's platform. We've had to rely on the public disclosures that Mr. Ford has made and statements by the PC Party since his leadership was announced. Again, the PC Party has had the opportunity to review all of their calibrations in advance and we have not received any indication that they disagree with how they're represented.
-
So, this is a good example of how the weighting options in Vote Compass can come in handy. If you're a single-issue voter, or you ascribe more salience to some issues than to others, you can weight your results in Vote Compass by the issues you care about the most.
-
That's about all the time we have. We'll be returning with some more lunchtime live blogs on a whole range of issues as we get closer and closer to the election.But I'm curious ... what is the strangest use of Vote Compass that you have heard? Has it been used in a bizarre place? Have you used it for a weird thing?
-
-
-
-
-
-
-
Thanks again!
And here's the link to Vote Compass, if you haven't given it a shot yet.Ontario Votes 2018 - Vote Compass
Vote Compass is a tool developed by political scientists that calculates how your views compare with those of Ontario's political parties.