502 – Innovative Statistical Methods for Complex Survey Data
Producing Labour Statistics by Detailed Geography and Occupation: Experiences from the New Canadian Job Vacancy and Wage Survey
Etienne Rassart
Statistics Canada
Zdenek Patak
Statistics Canada
In order to fill gaps in labour statistics data on job vacancy, the Canadian Job Vacancy and Wage Survey was launched in July 2014. The goal of the job vacancy component is to produce quarterly estimates of the number of job vacancies by economic region and occupation, while the goal of the annual wage component is to produce average hourly wage and employment estimates by economic region and occupation. Together, the two components could be used to produce vacancy rates by occupation. The components are integrated and use quarterly samples of 100,000 business locations in Canada on a rotating basis, making the survey one of the largest business surveys at Statistics Canada. The paper highlights methodological issues and solutions related to producing statistics at such detailed levels, while managing respondent burden. One challenge, common to both components, was defining a probabilistic non-response follow-up strategy prioritizing large locations. For the wage component, the main challenge was generating a sample of relevant occupations used to collect detailed wage information from each location. The solutions used sequential Poisson sampling and Pareto sampling.