Abstract
There are two aspects of this article that readers should find particularly interesting. First, Marlin and Niss provide additional data on the question of student evaluations of professors and courses. Second, they suggest the use of canonical correlation as an ideal way of facing the problems confronted in analyzing the educational production function. The authors assert that “canonical correlation would seem to answer many of the problems encountered in previous single-equation models….” Their research suggests that student evaluations can be used as surrogates for direct evaluations and that the former “do indeed measure the level of teacher input.”