Recent updates to Statistics Canada’s social data dissemination guidelines emphasize the use of confidence intervals (CIs) for reporting the accuracy of survey-based estimates. Not only are CIs better suited to the task than Coefficients of Variation for proportions, but they also serve as a valuable tool for drawing statistical conclusions from survey data, such as testing whether two proportions differ.
In this talk, we introduce CIs in the context of survey data, using analogies and accessible language. We highlight best practices to help users confidently use of CIs to report the accuracy of estimates and draw valid statistical conclusions from survey data.
Presenter: Claude Girard, Senior Methodologist, Data Analysis Resource Centre
To register for the English webinar, please fill out the following form: