Rural Crossroads: The Influence of Rurality on Education Policy Decisions and Subsequent STEM Outcomes

This dissertation considers educational policy implementation in rural Texas schools across a variety of policy types and STEM outcomes. Rural schools have unique limitations to their resources, setting them apart from their suburban or urban counterparts. Multiple initiatives can stretch these resources thin and require tough decisions that impact both the students and their communities. Using quantitative causal methods and descriptive analysis of state-wide data from the University of Houston’s Education Research Center (UH-ERC), rural implementation of the policies: Foundation High School Program (FHSP) established by state House Bill 5, bond issues for capital improvements, and Computer Science for All are examined and evaluated. The longitudinal data used in this study comes from the Texas Education Agency (TEA) and the Texas Higher Education Coordinating Board (THECB), which are used to consider both short- and long-term student outcomes post-policy implementation. The first study takes a deeper look at how the term rural is defined and how that definition can influence findings. Three differing classification systems for rurality are compared, first by their definition and student demographics and then through the lens of evaluating the implementation of FHSP. Findings suggest there are key differences in the results linked to how the term rural is defined, highlighting the importance of using a shared and clear definition when considering rural policy outcomes. The second study considers bond issues for capital improvements in rural schools and their links to STEM outcomes both in school and after graduation. While renovations and capital improvements are often much needed, they can divert administrative focus. Understanding the long-term implications of these efforts is important for informing decision-making on how time and resources are spent. This causal analysis relies on a regression discontinuity design (RDD) comparing those districts that narrowly pass a bond issue and those that narrowly do not. Findings show that while there are more immediate benefits to test scores in mathematics, there is little benefit to long-term outcomes, including teacher retention, high school course-taking, certification, or degree obtainment. The third study considers long-term outcomes for rural students who engage in computer science (CS) in high school. Using a latent class analysis (LCA) with propensity score matching (PSM), this causal analysis first identifies clusters of course-taking for CS students and then focuses on the class of students who engage in skills-based courses. The PSM is used to compare the outcomes for these students to those who did not engage in a skills-based pathway. Since 50% percent of rural Texas schools do not offer any CS courses, by running a second level of evaluation with only these students, a measure of unobserved variable bias can be determined and subtracted from the model to determine causal links between course-taking and post-secondary outcomes. Findings show those students who are likely to engage in a skills-based set of courses are six percentage points more likely to pursue a STEM degree in college.