Methodological issues in analyzing small populations using CCHS cycles based on the official language minority studies.

Statistical analyses for small populations or small domains of interest can be challenging. To obtain reliable estimates, only very large surveys such as the Canadian Community Health Survey can be considered. However, despite its good geographical and temporal coverage, the analysis of small populations in smaller regions (e.g., health regions) and in regards to specific health issues remains challenging. We will look at the methodological issues in analysis of small populations in relation to sampling and non-sampling survey errors that affect the precision and accuracy of the estimates. Francophone minorities in Canada will be used to illustrate the issues throughout the paper.