Sampling and Weighting Technical Report, 2011 National Household Survey (NHS)
Revised March 24, 2015

Introduction

The 2011 National Household Survey (NHS) is a new voluntary survey distributed to about 30% of Canadian census households. It collects information previously collected by the long-form census questionnaire which was distributed to one in five households every five years. The 2011 Census questionnaire consists of the same eight questions that appeared on the 2006 Census short form questionnaire as well as two additional questions on language. Contrary to the NHS, participation to the census is mandatory.

The NHS collected social and economic information that communities need to plan services such as child care, schooling, family services, housing, roads and public transportation, and skills training for employment. The goal of the NHS is to provide data for small geographical areas and small population groups.

In order to help users assess the usefulness of NHS data for their purposes, this 2011 National Household Survey Technical Report on Sampling and Weighting details the conceptual framework and definitions used in conducting the NHS, as well as the data collection and processing procedures employed. Because those procedures are linked to the data collection and processing procedures employed by the census, the 2011 Census data collection and processing procedures are also described. The report then outlines the method of sampling and weighting used in the 2011 NHS as well as its effect on the results. Because some information is collected on a sample basis and weighted to the full population level, bias and discrepancies may be observed in the final estimates. This report identifies some of these observed differences and explains the probable causes.

Because of its potential impact on response rates, the new voluntary nature of the NHS has brought more complex sampling and weighting methods than those used for the 2006 Census.

In general, the motivation to use sampling stems from a desire to reduce respondent burden, reduce costs, obtain results sooner, or a combination of all three. A disadvantage of sampling is that the results based on a sample may not be as precise as those based on the whole population. Nonetheless, when the loss in precision (which may be quite small when the sample is large) is tolerable in terms of the intended uses of the data, sampling is often cost-effective.

The 2011 Census of Population and NHS used sampling in a variety of ways. It was used in:

• Ensuring that the quality of the enumerator's work in collecting questionnaires met certain standards.
• Controlling the quality of coding responses during processing.
• Evaluating the quality of the census and NHS data.

The primary use of sampling, however, was to select dwellings to be included in the NHS sample. This report describes this specific use of sampling and evaluates its effect on the quality of NHS data.

Chapter 1 details the collection methods used in the 2011 Census and NHS. Chapter 2 explains the data processing procedures. Chapter 3 describes how sampling is applied in the NHS. Chapter 4 gives an overview of the procedures used in applying weights to the NHS sample data in order to obtain estimates for the population. Chapter 5 provides information on NHS non-response bias indicators while Chapter 6 describes the weighting areas. Chapter 7 presents an evaluation of the weighting procedures and Chapter 8 provides an overview of the variance estimation methodology in the 2011 NHS.

• Chapter 1 Data collection
• Chapter 2 Census and National Household Survey data processing
• Chapter 3 Sampling
• Chapter 4 National Household Survey estimation
• Chapter 5 National Household Survey non-response bias indicators
• Chapter 6 Weighting areas
• Chapter 7 Evaluation of weighting procedures
• Chapter 8 Variance
• Chapter 9 Conclusion
• Appendix A Glossary
• Appendix B The history of sampling in the Canadian censuses
• Appendix C Constraints used in the weighting process
• Bibliography