2011 NHS: Concerns and Solutions

Now that most of the 2011 National Household Survey (NHS) has been released, we are receiving many questions from clients who are curious about what the NHS is and how useful it is. The NHS is a voluntary survey meant to replace the Census Long Form for the 2011 Census. Contrary to popular perception, the NHS will not be provided by Statistics Canada completely free of charge.  While NHS numbers at the Census Subdivision (CSD) level of geography and higher are available at no cost, Dissemination Area (DA) tabulations are not. A Dissemination Area is the smallest standard area for which census data is aggregated, and our PCensus software requires DA level data to generate small-area demographic profiles.

Some clients are concerned about the usefulness and strength of the NHS and its content. These concerns are not unfounded- due to the new methodologies that the NHS is based on, and the fact that the survey is voluntary. The topic of greatest concern is the amount of suppression present in the NHS. Suppression is a technique used to ensure the privacy of those surveyed, and it has always been present in the Canadian Census. However, new in the NHS is suppression at the demographic variable level. This means that some variables – income for instance – can be given values of zero if too few people in the area fill out the survey. This is significant for two reasons. First, users have no way of knowing which variables have been suppressed. In some cases, it is difficult to tell if the zero-value on the survey reflects reality, or if it is a result of suppression. The second issue is the frequency of the problem. It has been estimated that over 10% of the 56,204 Dissemination Areas in Canada are affected by suppressed records. Due to the nature and frequency of suppressed records in the NHS, the accuracy of the data may be inadequate for some market analysis and site selection studies. As Such, many analysts’ confidence in the survey’s utility has been shaken.

Notwithstanding these concerns, the National Household Survey is still a useful data source if its limitations are understood and it is used correctly. At high-level geographies, it is as effective as the historical Census reports we are used to. For use in small-area geographic analysis, however, it is beneficial to complement the NHS with supporting data. If your organization requires DA level data for trade area analysis using circles, custom drawn polygons, or drive-times, it is advisable that you ‘fill in the gaps’ of the NHS with modeled data such as our 2014 current year demographic estimates and projections. We will also offer ‘Adjusted Census’ products from Environics and Pitney Bowes Software which resolve some of the Census data inconsistencies and offer Dissemination Area tabulations.