Defining The Majors That Comprise "Stem": An Analytical Method For Looking Beyond The Classical Acronym

There is no universally agreed-upon definition for "STEM." Research in STEM education would be aided by developing a convention for how to categorize students' potential college majors. This study develops and tests a definition of STEM and a method for verifying the definition's validity and usefulness. A classification scheme is developed from reviewing prior educational research; examining the National Science Foundation's categorizing of fields of study; and constructing a predictive model. Demographic, attitudinal, experiential, and standardized test score variables from the National Educational Longitudinal Study of 1988 (NELS:88) dataset are used to predict which students will earn a STEM degree vs. another educational outcome. The predictive model tests the classification of majors into different four year degree outcomes. It tests the inclusion of majors beyond the classic "bench" sciences, engineering, and mathematics. Models were fitted to predict whether students had a STEM vs. other outcome, and the predicted outcomes were compared to actual outcomes. The models with more divergent outcomes had greater accuracy. Students earning a degree involving significant quantitative coursework in a field not generally considered STEM had more factors in common with students earning a Non-STEM degree. A narrow definition of STEM focusing on the bench sciences, engineering, and mathematics produces more accurate predictions. This method for classifying educational outcomes allows researchers to precisely define which majors or non-degree outcomes are included in the analysis. The method is flexible enough to accommodate a mutually exclusive and exhaustive set of outcomes.