scientific teaching?

I am on a committee at NYU (related to NYU's Morse Academic Program) in which we were given copies of this article on scientific teaching (PDF) by Handelsman et al. The article has many sentences that include the word should, which is a little annoying, but the worst thing about it is that it advocates, repeatedly, basing our teaching methods on scientifically demonstrable successes, measured with metrics. I don't object to being scientifically accountable, and I, for one, use the research-supported techniques of active learning, participatory classroom activities, and peer instruction. However, the article goes on about metrics without giving a single clear example. Why? Because they don't want to undermine their point by bringing up controversial testing strategies. However, we can't be metric-driven if we don't have metrics! So let's all just stop talking about the philosophy of being scientific and find some metrics that we can agree on. I, for one, haven't found anything I like. After all, the goal of physics education is not to create students who perform better on isolated, decontextualized exams like the Force Concept Inventory!