Predictive modeling in education draws on data from past courses to forecast the effectiveness of future courses. The present effort sought to identify such a model of instructional effectiveness in scientific ethics. Drawing on data from 235 courses in the responsible conduct of research, structural equation modeling techniques were used to test a predictive model of RCR course effectiveness. Fit statistics indicated the model fit the data well, with the instructional characteristics included in the model explaining approximately 85% of the variance in RCR instructional effectiveness. Implications for using the model to develop and improve future RCR courses are discussed.