Re: Review of Sampling Methodology for CAC Automobile Insurance Survey
(October 13, 2004)
Bruce Cran
Consumer’s Association of Canada
P.O. Box 18112
Delta BC V4M 2M3
Dear Mr. Cran:
Thank-you for the opportunity to review the sampling methodology for this study.
Having reviewed the quota-sampling scheme used for this study, I am satisfied on the following points:
The quota-sampling scheme used to select driver profiles is an excellent method of ensuring that the sample reflects the correct gender/age/claims history composition, and cannot be further improved, given the available data and the structure of the study.
The probability proportional to size methodology used to select postal codes is a well-established method that is well suited to the structure of this study.
The profile selection and postal code selection methodologies were performed correctly.
The study method – comparing fixed profiles across provinces – provides a fairer comparison than using actual provincial statistics. Further, the quota-sampling and postal code selection methodologies used are very appropriate for this type of study.
Any remaining bias is a result of limitations on available data (e.g. Canada-wide claims histories not being available), which cannot be further reduced without more data.
I am comfortable advising that the study proceed using this methodology.
I’ve attached my formal report and a copy of my curriculum vitae. If you have further questions, please do not hesitate to contact me.
Sincerely,
Carl James Schwarz, P.Stat.
Statistical Consulting Service
Simon Fraser University
cschwarz@stat.sfu.ca
Home / Privacy & Security Statment / Contact Us / Site Map / Search
