Leisure and other data from the Canadian labour force survey add‐on surveys: Bias concerns

Abstract

This article provides intuitive and quantitative insights into biased estimates (e.g., means) being a problem when using Canadian Labour Force Survey (LFS) add‐on survey data. This matters because LFS add‐on surveys include key sources of leisure related data and Canada's “Data Liberation Initiative” is encouraging their use. The magnitude of bias occurring is illustrated by finding a 16% reduction in reporting having travelled the second time data on travel is obtained from a household. It is concluded that avoiding spending time reporting travelling is one reason for reporting no trips when there was travel. Evidence provided leads to the conclusion that responses to LFS addon survey questions have been significantly impacted by add‐on survey data being collected from the same household multiple times. Why some bias effects apply to most LFS add‐on surveys is examined. The impacts of bias and bias change on applied and academic research are considered.