I think the idea is that the apprentice's students are also taught (in part) by a "known good" teacher, which minimizes the risk to the initial students. After the three-year apprenticeship, you have three years of (complex) data for you evaluation of the "regular teacher" candidate. What I want to know is why the schools don't already have this information for their long-term teachers. The stats aren't *that* complicated to run. The data already exists. So why aren't the standardized testing companies like CTB offering this service to schools?