A Statistical Examination of Factors Influencing Graduation Status of Students with Disabilities
Loading...
Authors
Roddenberry, Missy
Issue Date
2026-01-09
Type
Dissertation
Language
en_US
Keywords
Special education , Education , Education, Secondary , High school graduates , High schools , Dissertations, Academic
Alternative Title
Abstract
The current study examined student-level and school-level factors influencing on-time high school graduation among 425 students with disabilities (SWD) across the 2022, 2023, and 2024 graduation cohorts in a South Georgia school district and compared the predictive accuracy of multiple statistical and machine learning models.Results revealed that students who were not chronically absent and those spending more time in general education classrooms had substantially higher odds of graduating. In contrast, students placed in alternative schools demonstrated lower graduation odds. Unexpectedly, students who qualified for free or reduced-price lunch had higher graduation odds in this district, contradicting national patterns linking economic disadvantage to lower graduation rates. In addition, logistic regression provided the highest overall accuracy (82.4%) and interpretability, compared to neural network, random forest, and support vector machine models. For class imbalance, none of the models accurately classified non-graduates, with specificity ranging from 29% to 48%. These results highlight the importance of early detection and intervention efforts focused on attendance, inclusive placement, and support for students in alternative settings. While advanced machine learning models offer incremental gains in accuracy in some contexts, the current study highlights the practical value of logistic regression for school districts seeking clear, data-driven methods to identify students at risk and inform policy interventions that support the graduation of students with disabilities.
Description
Citation
Publisher
License
This dissertation is protected by the Copyright Laws of the United States (Public Law 94-553, revised in 1976). Consistent with fair use as defined in the Copyright Laws, brief quotations from this material are allowed with proper acknowledgement. Use of the materials for financial gain with the author's expressed written permissions is not allowed.