9. Conclusion

The introduction of the 2011 National Household Survey (NHS) to replace the census long-form questionnaire resulted in many challenges. As a result of the non-mandatory nature of the NHS, the non-response rates were higher than for previous long-form censuses. The types and patterns of non-response were also different. A more complex approach was thus required for the sampling design, non-response follow-up, weighting and estimation procedures for the NHS.

Changes included an increase in the sample size from one in five to one in three private households. An improved strategy for non-response follow-up was implemented to ensure that certain areas and populations identified to be at risk for higher than usual non-response rates were targeted. Although there are several limitations, bias indicators due to total non-response were produced for a few characteristics. Finally, the complex sample design required for the NHS led to the development of innovative new weighting and variance estimation methodology that are quite different from previous censuses.