Structuring First-Year Retention at a Regional Public Institution: Validating and Refining the Structure of Bowman’s SEM Analysis

Abstract

Structural equation modeling (SEM) considering how students’ non-cognitive attributes influence first-year college student persistence remain extraordinarily rare—as are studies that test and expand upon published structural models or studies that include college student food security. This study addresses each. We surveyed “Beginner” Freshmen, capturing eight non-cognitive measurements and ussing institutional data on performance and fall-to-fall persistence measures, we then tested the structure of Bowman et al.’s (Res High Educ 60:135–152, 2019) SEM model. In Model 1, we mimic the Bowman model’s financial variable by only including financial stress. We confirm that Bowman is a good structural model of student persistence, although our data were collected for another purpose, using different scales for non-cognitive elements and even one different non-cognitive measurement. We found students’ non-cognitive attributes remain importantly influential to social adjustment (r = .65), commitment to persist (r = .40), college GPA (r = .25), and fall-to-fall persistence (r = .30). In Model 2, we generated a latent financial security variable incorporating financial stress and food security. Including food security generated a direct influence from the financial security variable to high-school GPA (r = .25), not found in the Bowman model or Model 1, and a direct significant relationship from financial security to social adjustment (r = .11)—not found in Model 1. Further changes are observed in the indirect relationship from financial security to college GPA from Model 1 (r = .29) to Model 2 (r = .51). We highlight the robustness of the Bowman model and that the inclusion of food security brings increased strength to several relationships without sacrificing optimal fit.