Revised CIS-2008 Weighting Procedure Description
The data collected for the CIS 2008 were weighted to derive national annual incidence estimates by applying a composite regionalization weight and an annualization weight. The regionalization weight was developed to estimate the number of investigations completed within the three-month data collection period by child welfare organizations across Canada. The regionalization weight includes three components: (1) a sample weight that adjusts for the disproportional selection of agencies from oversampling provinces, (2) a subsampling weight that accounts for random subsampling of investigations in agencies that investigated more than 250 cases during the three-month data collection period, and (3) an agency size correction, designed to adjust for variations in the size of agencies within a stratum. The annualization weight is used to estimate annual investigation volume based on the investigation volume during the three month data collection period of CIS-2008. The annualization weight is the ratio of all investigations conducted by a sampled agency during 2008 to investigations conducted by the sampled agency during the case selection period.
Three limitations to this estimation method should be noted. The agency size correction uses child population as a proxy for agency size; this does not account for variations in per capita investigation rates across agencies in the same strata. The annualization weight corrects for seasonal fluctuation in the volume of investigations, but it does not correct for seasonal variations in types of investigations conducted. Finally, the annualization weight includes cases that were investigated more than once in the year as a result of the case being re-opened following a first investigation completed earlier in the same year. Accordingly, the weighted annual represent the child maltreatment-related investigations, rather than investigated children.