I'll start answering, and then I'll ask Mr. Swales to add a few details.
I want to specify that it wasn't a sampling method, initially. We reviewed the information regarding 19 million vehicles, and we identified issues in 500,000 cases. We selected a sample from these cases. We first analyzed the entire population, and the sample then helped us determine that the issue concerned 300,000 vehicles.
Mr. Swales may be able to provide a few more details.