Concomitant-Variable Latent-Class Beta Inflated Models to Assess Students’ Performance: An Italian Case Study
- 74 Downloads
Students’ performance is a crucial aspect for university programs effectiveness and organization. In this paper, we introduce and analyze a performance index for the first-year students of a private Italian university, namely the Libera Università Maria Ss. Assunta. We use administrative data on 532 undergraduate students enrolled in any of the eight available bachelor degrees in 2015. Our aim is to improve the general understanding of performance linking it with personal student’s characteristics and with degree-specific aspects. A beta inflated latent class approach is employed to identify clusters of performance establishing a link with all available explanatory variables. The empirical analysis unveils that a good and balanced degree organization may improve students’ performance. The student’s ability plays a crucial role in discriminating between good and bad performances, and also strongly depends on individual-specific characteristics, such as the final mark obtained at high school.
KeywordsLatent class Students’ performance Private university Beta distribution Zero-inflation
- Bini, M., Bertaccini, B., & Bacci, S. (2009). Robust diagnostics in university performance studies (pp. 139–160). Heidelberg: Physica-Verlag HD.Google Scholar
- Johnson, N. L., & Kotz, S. (1970). Continuous univariate distributions (Vol. 2). Hoboken: Wiley.Google Scholar
- McLachlan, G., & Krishnan, T. (2007). The EM algorithm and extensions (Vol. 382). Hoboken: Wiley.Google Scholar