Data notes to "Will no one rid us of these turnout nerds?"
Statcan wants me to make sure you know "This analysis is based on the Statistics Canada General Social Survey, Cycle 22: Social Networks, 2008. All computations, use and interpretation of these data are entirely that of Colby Cosh."
The actual sample sizes for the "Olds" and "Youngs" categories defined in the article are n=1780 for the "Olds" and n=1997 for the "Youngs". To derive national estimates of both eligible-voter totals and "did-vote" totals, the answers by respondents have to be adjusted by individual demographic weights, designed (by Statcan) such that any possible subsample reflects as closely as possible the geographic and demographic qualities of the population as a whole. Individual respondents who have demographic features underweighted in the sample are thus given extra weight for the purpose of making population estimates like mine. This makes estimating confidence bounds and calculating statistical significance a little tricky: readers who are interested can consult the documentation for GSS Cycle 22 (which, unlike the microdata themselves, is publicly viewable). But the differences between the estimated "Old" and "Young" turnouts do appear to be many, many standard deviations apart in every case.