An Investigation of the Usefulness and Predictive Ability of the College Student Inventory at a Small, Private, Technological Institution in the Southeast
Date of Award
Doctor of Philosophy (PhD)
Thomas J. Marcinkowski
Joo Young Park
Over the past several decades, retention research has focused on influential factors related to student persistence, specifically first-year persistence. Past research has identified the importance of prior academic achievement and demographic factors and, more recently, non-cognitive, psychosocial factors. In practice, first-year programs and the use of early intervention tools such as the College Student Inventory (CSI) have increased. However, practices still seem to lag behind theory and research, as first-year retention and six-year graduation rates remain stable at 74% and 59%, respectively (Kena et al., 2015; NCHEMS, 2016). We also know that the causes and characteristics of retention tend to be institution-specific. The overarching question that guided this research was: To what extent is the CSI (Form A) valid and useful for identifying first-year students at risk of leaving a small, private, independent, technological institution? There is almost no published research on the use of and/or subsequent results of these kinds of assessments at comparable institutions. Thus, from a research perspective, one of the primary aims of this study was to add to this gap in the retention literature and provide a potential point of reference for similar institutions. From a practical perspective, the purpose of this research was to contribute to the understanding of retention on this campus. Secondary data analysis served as the primary methodology. Hierarchical regression analyses were conducted to investigate the relationship between six sets of variables and two college outcomes: academic performance and persistence. Sets A-D (i.e., academic motivation, general coping ability, receptivity to support services, and social motivation) mirrored the overarching constructs measured by the CSI, and included its 17 major scales. The remaining sets, Sets E and F, were comprised of academic and demographic factors commonly included in persistence studies. Academic performance was defined as GPA after two points: the first semester (FSGPA) and the first year (FYGPA). Persistence was determined by enrollment status at three transition points (i.e., the second, third, and fifth semesters). Following IRB approval, data were provided by the Office of Institutional Research. Multiple regression results indicated that eight variables had significant positive relationships with academic performance at the end of the first semester (FSGPA) and first year (FYGPA). These included two CSI scales (i.e., Study Habits and Family Emotional Support), five academic factors (i.e., standardized test score, high school GPA, and a declared major in aeronautics, business, and psychology and liberal arts), and one demographic factor (i.e., international student status). The final models explained 27.5% and 32.5% of the variance in FSGPA and FYGPA, respectively. Logistics regression results indicated that four variables had significant positive relationships with persistence at two or more points in time. Three were academic factors (i.e., standardized test score, high school GPA, and a declared major in aeronautics), and one was a demographic factor (i.e., gender). The amount of variance in persistence explained by the study variables increased from the second semester to the third year, with the latter explaining 9.8%. Results indicated that the study variables were able to explain considerably more variance in academic performance than persistence. The CSI-A made a relatively small contribution to the explanation of variance in academic performance and persistence, explaining only about 6% of the variance in GPA. The academic factors included in this study explained the bulk of the variance in both outcomes at every point in time. Results indicated that high school GPA was the strongest and most consistent IV across all sets. These results are consistent with prior theory and research. Recommendations for further research, as well as for practice at this institution, were offered.
Ha, Jessica Paula, "An Investigation of the Usefulness and Predictive Ability of the College Student Inventory at a Small, Private, Technological Institution in the Southeast" (2018). Theses and Dissertations. 994.