Without minimizing the difficulties inherent in any value added model for performance evaluation of anyone or anything in education — teachers, administrators, teacher preparation programs — the basic idea is intuitive: With regard to teachers, if one student moves from academic point A to academic point C over the course of a year, whereas another student in a similar class and with similar characteristics only moves from point A to point B, the first student's teacher should be recognized for "adding more value," even though both students end up in the same place academically.
A recent column by a fierce opponent of value added models in education asked rhetorically if we'd be happy if such models were used to evaluate doctors. (Actually, she asked if we'd be happy if doctors were measured by their outcomes without regard to the health status of the populations they serve, which is the antithesis of value-added, but why quibble?)
In fact, we suspect value added models are all around us and simply aren't noted as such. We don't know of any value added models for doctors, but the US News and World Report hospital ratings do include two types of value-add type evaluations: First, a "survival" score isn't just based on survival rates, but takes into account patients' conditions. (So two hospitals whose patients survive at the same rate (Point C) aren't rated identically if one admits patients who are sicker (Point A) than another (Point B). Second, a "patient safety" score takes into account the deaths of patients whose conditions should not have put them at significant risk. (This means that a hospital specializing in hangnail removal whose patients succumb to infection is revealed as the deathtrap it is, whereas an infectious disease hospital whose patients sometimes die from infections is revealed as a medical paragon.)
Taking starting points into account when evaluating end points makes good sense — in every field. The problem is not value added models...the problem is making them as sensitive to the nuances of the field as they can be.
Julie Greenberg
A recent column by a fierce opponent of value added models in education asked rhetorically if we'd be happy if such models were used to evaluate doctors. (Actually, she asked if we'd be happy if doctors were measured by their outcomes without regard to the health status of the populations they serve, which is the antithesis of value-added, but why quibble?)
In fact, we suspect value added models are all around us and simply aren't noted as such. We don't know of any value added models for doctors, but the US News and World Report hospital ratings do include two types of value-add type evaluations: First, a "survival" score isn't just based on survival rates, but takes into account patients' conditions. (So two hospitals whose patients survive at the same rate (Point C) aren't rated identically if one admits patients who are sicker (Point A) than another (Point B). Second, a "patient safety" score takes into account the deaths of patients whose conditions should not have put them at significant risk. (This means that a hospital specializing in hangnail removal whose patients succumb to infection is revealed as the deathtrap it is, whereas an infectious disease hospital whose patients sometimes die from infections is revealed as a medical paragon.)
Taking starting points into account when evaluating end points makes good sense — in every field. The problem is not value added models...the problem is making them as sensitive to the nuances of the field as they can be.
Julie Greenberg