An interesting thought. One idea that comes to mind is to measure the density of party lawn signs in closely run areas.
In other words, take a sample of 10 blocks and count the number of lawn signs per party. We used the upcoming Ontario elections as an example, and focused on Richmond Hill (suburb just north of Toronto).
Before the election, the lawn sign count was:
29% PC – 52% Liberal – 19% NDP
However, the outcome on the 2018 Election Day was:
52% PC – 28% Liberals – 17% NDP – 3% Greens
So what does that tell you?
Using their commanding lawn sign coverage as a predictor, the Liberals are pushing hard to keep their 2014 majority. However, it is important to choose the most accurate indicators of voter sentiment.
Lawn signs seem not a good predictor of election outcomes for many reasons. Many “ordinary” citizens (and by that we mean those who are not personally involved with a political party), often don’t want to advertise for whom they are going to vote. Others don’t really make up their mind until they get to the pooling booth. Media coverage may not be balanced or there may be a concern by the media about bias, so perhaps they are not asking the tough questions to source the true ambitions or platforms of certain parties or politicians.
Voting is a deeply personal and subjective exercise, and even with surveys by independent research organizations, considerable thought is required in formulating the questions, determining the sample size, deciding the sample components (e.g., age, region, language, etc.), and in interpreting the results.