Despite their numerous advantages, exit polls are not a common tool in the study of Canadian electoral behaviour. In this methodological note, we use data from two pilot projects to test small-scale exit polls’ accuracy when estimating party support. We mobilize exit-polling data collected in the 2018 Quebec provincial election (four voting locations) and the 2019 federal election in Quebec (two voting locations). We focus on chance error and bias error in small samples. Results obtained using parametric linear models suggest that small sample exit polls achieve relatively precise estimations. We do find, however, that right-of-the-centre parties’ vote share tends to be underestimated. These findings shed light on the strengths and shortcomings of small-scale exit polls in Canada.