Yes, that's a good point, but I think StatsCan uses self-reporting as well, not just police reporting, because for sure the police data is very skewed.
In terms of data on femicides, at the moment Statistics Canada's homicide survey does not capture data on femicides; it only collects some variables related to gender and risk factors for femicide, and that is one huge gap. When it does collect data on these things, it can collect whether sexual assault happened at the time of femicide, but it doesn't capture a history of sexual violence, for example, and it also does not capture underlying socio-cultural or systemic factors, only individual factors. We really want more data.
At the moment, the data collected by the Canadian Femicide Observatory for Justice and Accountability is basically collected by death review committees, which are essentially made up of volunteers or people from the sector who are doing this on the side. They're using whatever information is available to them, so it could be media reports. I think we can all agree that those will be incomplete.
They don't have access to autopsy reports, for example. We want data on whether there was sexual violence during, before or after the killing, for example, and what type of violence was present, and then disaggregated data on different populations and the relationship between the sexual violence and the population group—for example, older women, as we said earlier.