Spatial Econometrics and Network Analysis as Means to Assess the Assumption of Independence in Higher Education Research

Abstract

The study guides institutional research professionals through the steps of identifying, diagnosing, and modeling outcome dependence of credit accumulation given students' major of ascription, a pragmatically and timely relevant endeavor when considering the current emphasis on outcome-based or performance-based funding policies. with a highly prolific colleague surely impacts my publication prospects). At the student level, relationships to one another and/or their shared participation in academic activities such as courses/seminars and/or majors affect GPA variation, credit accumulation, major persistence, and success in these indicators have a direct effect on continuation and eventual graduation. In both cases, the tenets of outcome dependence (Bivand et al., 2013;Cliff & Ord, 1969, 1972 indicate that failing to account for individuals' systemic and systematic exposure to colleagues or peers and/or their common exposure to courses and/or majors, may yield model estimates that are over or under-reporting the estimates of interest given the influence of over-or underperforming peers.The main premise of this chapter is that even though outcomes taking place in higher education settings are not independent, our traditional naïve quantitative models assume that (a) peers' performance do not influence their peers' performance and are not simultaneously affected by them in return and (b) rarely test whether this independence assumption holds true in the analytic samples. To address this potential source of bias, the chapter offers an analytic framework to address outcome dependence that relies on the spatial statistics or geostatistics framework (Cliff & Ord, 1969, 1972. The main contribution of the chapter consists of adapting this spatial framework to address dependence emerging in social settings with direct applicability to institutional and higher education research. From this perspective, the chapter aims to guide institutional research professionals through the steps of identifying, diagnosing, and modeling outcome dependence given actors' connections to one another either directly or through their ascriptions to a common exposure, such as major field of study. These steps are attained through the merger of spatial and network analysis principles.The procedures and analyses contained in this chapter are of particular interest to institutional research professionals and higher education administrators and scholars as they relate to outcome dependence of credit accumulation based on students' major of ascription.These analyses and examples were conducted using institutional research data obtained from two community colleges located in different states and regions of the country (California and Georgia). These examples are pragmatically and timely relevant, particularly when considering