DescriptionPURPOSE AND GOALS This presentation follows up on the previous work in the field establishing connections between students’ use of libraries and student success measures such as GPA and retention (e.g. Soria, Fransen & Nackerud, 2014; Soria, Fransen & Nackerud, 2017; Nackerud, Fransen, Peterson & Mastel, 2013). Critiques of this body of work (e.g. Robertshaw & Asher, 2019) suggest that the correlations between students’ use of libraries and their GPA have low to non-existent effect sizes. This critique prompted a deeper analysis of more specific library usage types and the distinct populations that use them, since we know that students from different backgrounds experience college in disparate ways and have varying needs. Analyses of Fall 2011 first year undergraduates showed correlations between use of library resources and typical success measures (GPA, retention). Effect sizes were small when compared to predictors such as ACT score and college credits earned while in high school, but still significant. The studies led to different kinds of outreach efforts, and by Fall 2016 88 percent of undergraduates used the library compared to 77 percent in Fall 2011. Buoyed by our belief that using libraries helps students attain their goals, we sought to do more with what we have. Rate of library use is now so high overall that looking for correlations, while controlling for the many other factors that influence student success, has little value. Instead, we shifted our focus to specific populations, defined by characteristics such as college of enrollment and first generation status. For example, if data show that first generation students in the college of biological sciences are less likely to use the digital resources, and that those students in that cohort who DO use digital resources are more successful, we can devote time and effort to reaching those students. Yet, how can we efficiently analyze all of these potential characteristics? DESIGN, METHODOLOGY, OR APPROACH Using multiple demographic and institutional indicators, we took an iterative approach to building statistical models using one, two, and three predictor variables to create ANOVA models. This allowed us to estimate the correlation and see the effect sizes and eta-squared measures to have a more nuanced understanding of the importance specific correlations have. This approach follows the blueprint established in the previous round of analysis (Gyendina and Fransen, 2019). We started with creating one-way ANOVA models based on focused subsets created using three demographic and institutional variables. For example, bring together characteristics like age, college (e.g. science and engineering), and had library instruction to see if there is a meaningful correlation. After analyzing the preliminary results, we identified key areas with the highest potential for actionable steps and developed relevant three-way ANOVA models. FINDINGS We present the relevant models with moderate to high effect sizes and corresponding actionable steps we are proposing. The presentation also describes the process we used and invites the audience to consider this methodology. PRACTICAL IMPLICATIONS OR VALUE The presentation invites the audience to consider the potential use of data to communicate with college stakeholders.
|Period||Jan 21 2021|
|Event title||Library Assessment Conference 2020: null|
|Degree of Recognition||National|