Can someone explain why logarithmic function is used in sample sizes of election polling? For example, logarithmic function was used to estimate voting intention for different political parties during the 2019 Canadian election:
Evolution of voting intentions during the pre-campaign period of the 43rd Canadian federal election. Trendlines are local regressions, with polls weighted by proximity in time and a logarithmic function of sample size. 95% confidence ribbons represent uncertainty about the regressions, not the likelihood that actual election results would fall within the intervals.
I have general knowledge of statistics (first year level) but I don't understand how logarithmic function could be used for this. Thanks.
