Person-Oriented and Variable-Oriented Statistical Approaches to Understanding At-Risk, Online High School Students: Predictions Of Academic Outcomes and The Moderating Role Of Protective Factors
Paek, Sue Hyeon
College of Education and Behavioral Sciences; School of Psychological Sciences Educational Psychology
University of Northern Colorado
Type of Resources
Place of Publication
University of Northern Colorado
Many students come to school with risk factors that negatively impact their educational attainment and overall well-being. These risk factors may include but are not limited to abuse, poverty, teen pregnancy, low academic achievement, and learning disabilities. Other types of factors have been shown to provide a protective aspect against risk, and examples include social emotional skills such as grit, flexibility, and social-awareness. There is benefit in individualizing educational experiences for at-risk adolescents based on their specific risks and available protective factors, but knowing how these factors influence and interact with one another to predict academic outcomes is the first step in identifying the most effective approaches for an individual’s education. This study sought to use a person-oriented statistical approach to determine what profiles of risk could be found in individual students in a large, at-risk, online high school population. Once profiles were identified, their relation to academic outcomes would be investigated, as well as any moderating effect of protective factors. Moreover, a variable-oriented statistical approach was used to analyze to what degree latent risk factors predicted academic outcomes and if the presence of protective factors moderated that relationship. Confirmatory factor analysis (CFA) and latent profile analysis (LPA) was used to identify latent profiles of risk in a sample of 734 at-risk, online high school students. Unfortunately, while three profiles were identified, they lacked sufficient distinction to be used effectively, so the person-oriented approach was not pursed further. Subsequently, a variable-oriented approach involving a series of hierarchical regressions was conducted to investigate if risk factors would predict academic outcomes and if social emotional protective factors would decrease the negative effects of the risk factors. It was found that for some academic outcomes, such as grade point average (GPA) and passer rate, educational risk factors, such as previous suspensions or expulsions, were often significantly correlated with decreases in academic achievement. This was above and beyond what family risk factors, such as abuse and poverty, might predict alone. Unfortunately, social emotional protective factors had little moderating influence between risk and academic outcomes. Based on the findings of this study, it would seem important for educators to work towards lessening the amount and severity of risk factors students experience in school. Additionally, it would be wise to support and nurture specific social emotional skills in our at-risk youth that will help them be successful while in school environments. Limitations include the restrictions that are associated with cross-sectional, self-reported, and archival data. Additionally, some factors were skewed left, which made prediction difficult. Finally, the number of regressions may have impacted the interpretability of some results.
Copyright is held by the author.