Modeling Parallel Latent Growth Trajectories with Time-Varying Baselines: A Demonstration Examining the Intersection between Minority-Serving Institution Status and Females' Participation in STEM Majors over Time

Abstract

The current paper demonstrates an application of parallel latent growth modeling with time-varying baselines and high missingness levels using extant data on higher education institutions participating in the National Center for Women and Information Technology Extension Services (NCWIT-ES) program between 2006 and 2018. Using this methodology, we test the relationship between the growth trajectory factors for the number of women applying to STEM undergraduate departments (of participating institutions) and the growth trajectory factors for the number of women graduating from STEM undergraduate departments. Further, we also illustrate use of an intersectionality lens in framing the research by including a key institution-level predictor, minority-serving institution (MSI) status, to test whether undergraduate women’s involvement in STEM significantly differed for MSI and predominantly white institutions (PWIs). Analysis results showed that the number of women applying to STEM departments at the year their institution joined NCWIT-ES was positively predictive of the number of women graduating from those STEM departments five years later. Moreover, growth in the number of women applicants per year was positively predictive of growth in the number of women graduates five years down the line. Last, year joining NCWIT-ES was positively predictive of higher baseline and greater growth in both applications and graduations, whereas MSI status had less application growth acceleration and was negatively related to graduation baseline and growth factors. Keywords: parallel latent growth model, time-varying baseline, higher education, STEM, women, minority-serving institutions, intersectionality