The effect of school location on math and science learning is currently an important policy issue in the United States and in other countries, such as Australia. The present paper uses a 5-year series of math and science achievement data from the state of Kentucky to determine the effects of school location on learning in these subject areas. Adopting an organizational assessment approach, I show how growth models may be used to estimate achievement trends. I also demonstrate methods for discovering two important sources of invalidity in growth models: regression artifacts and spuriousness. Failure to account for these sources of invalidity may lead to erroneous policy conclusions. Two examples of growth models are provided—a linear model and a nonlinear model. The results of these analyses do not support the common contention that there is a rural achievement gap in math and science. One implication of these findings is that, if policymakers wish to enhance math learning, they will accomplish this more effectively by interventions and programs that increase the motivation and opportunity to learn among low-income students, regardless of school location. Because current U.S. education policy is focused on documenting “adequate yearly progress” in schools, growth modeling is likely to become the preferred methodology of policy researchers.
Occasional Research Paper, No. 11, 2005.