Policymakers are fond of saying that we have enough research knowledge; however, one problem is that nobody is applying the knowledge we already have. This paper offers a discussion of the kinds of knowledge needed to improve students learning, what to do about this in the assessment and testing arena, and where assessment falls short. Types of knowledge are discussed, including research knowledge, which must be both usable and useful. Distinctions are made between usable and useful knowledge, and a case is made for how we might ultimately design our systems and our own actions to help us act with greater intelligence. (Author)