A New Approach for the Development of a Public Use Microdata File for Canada's 2011 National Household Survey
1. Introduction

Statistics Canada has been producing anonymized Public Use Microdata Files (PUMFs) from the Census Program long form samples since the 1971 Census. For the 2011 National Household Survey (NHS), which replaced the long form, an individual (person level) and a hierarchical (household level) PUMF were produced using the same approach as previous censuses. Representative subsamples of 2.7% and 1% of the Canadian population, respectively, were taken from the 2011 NHS and global recoding and local suppression were used to protect data confidentiality. Additionally, in answer to requests from Integrated Public Use Microdata Series-International (IPUMS-International), and to meet the needs of users who prefer to work with unsuppressed data, Statistics Canada has agreed to also produce a third, hierarchical, PUMF that relies on perturbation instead of suppression to protect data confidentiality. This paper describes the creation of this new type of PUMF for the Agency. The next section gives background information on the project. Section 3 presents the approaches used for sample selection and disclosure control. The disclosure control strategy is centered on targeted data perturbation, whose application is covered in detail in Section 4. Concluding remarks are given in Section 5, which also provides summary results on the outcome.