Comparison of two statistical methods to identify groups in a student satisfaction survey
DOI:
https://doi.org/10.34115/basrv8n2-003Keywords:
comparación, análisis de clases latente, análisis de cluster, satisfacción estudiantilAbstract
This work studies the main characteristics of the construction of Student Satisfaction in the undergraduate courses of the Facultad de Ciencias Económicas y de Administración from Universidad de la República, Uruguay, by using and comparing two multivariate data analysis techniques: Cluster Analysis and Latent Classes Analysis. Data used for the implementation presented in this work originate from a survey applied on a sample of undergraduate students in 2009. This questionnaire is structured in sets of questions: the first set contains variables which allow characterize students sociodemographically. On the other hand, variables from the ECSI (European Customer Satisfaction Index) model which shall be used to characterize Student Satisfaction are presented. The 6 observed/manifest variables considered as input for the construction and characterization of Student Satisfaction are: student expectations (E) when they begin studying in the institution, the image (I) they have of the institution, the quality of education received (CSA) and of services provided (CSF), the personal needs and desires with regards to College (ND), and the perceived value (VP). On the one hand, results presented arise from considering that there is an actual hidden/latent variable referred to Student Satisfaction and that this variable is defined from the interaction of the 6 manifest variables, by 4 latent classes. On the other hand, with reference to clusters, it is proposed to group students into 3 groups based on the analysis of results arising from clustering using Ward's method. Additionally, the comparison of results obtained using both techniques is presented.
References
AGRESTI, A. Categorical data analysis. Wiley-Interscience, Hoboken, N.J., 2013
ÁLVAREZ-VAZ, R., Freira, D., Vernazza, E., and Alves, H. Can students’ satisfaction indexes be applied the same way in different countries? Int Rev Public Nonprofit Marketing, 13(101). 2016.
ALVES, H. and Raposo, M. La medición de la satisfacción en la enseñanza universitaria: El ejemplo de la Universidade da Beira Interior. Int Rev Public Nonprofit Marketing, 1(1):73–88. (2004).
BANDEEN-ROCHE, K., Miglioretti, D. L., Zeger, S. L., and Rathouz, P. J. Latent variable regression for multiple discrete outcomes. Journal of the American Statistical Association, 92(440):1375–1386.1997.
BLANCO, J., Camaño, G., Nalbarte, L., and Álvarez-Vaz, R. Introducción al Análisis Multivariado. Instituto de Estadística, FCEA. 2006.
DEMPSTER, A. P., Laird, N. M., and Rubin, D. B. Maximum likelihood from incomplete data via the EM algorithm. Journal of the Royal Statistical Society, 39:1–38. 1977.
EVERITT, B. S. An Introduction to Latent Variable Models. Springer Netherlands, Dordrecht. 1984.
HAGENAARS, J. Applied latent class analysis. Cambridge University Press, Cambridge New York. 2002
LAZARSFELD, P. The logical and mathematical foundations of latent structure analysis. ISA Stouffer (ed.), Measurement and Prediction, pp. 362–412. 1950.
LINZER, D. A. y Lewis, J. B. poLCA: An R package for polytomous variable latent class analysis. Journal of Statistical Software, 42(10):1–29. 2011.
MAECHLER, M., Rousseeuw, P., Struyf, A., Hubert, M., and Hornik, K. cluster: Cluster Analysis Basics and Extensions. R package version 2.0.4. 2016.
MCLACHLAN, G. Finite mixture models. Wiley, New York. 2000
R CORE Team. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3-900051-07-0. 2017
VERNAZZA, E. (2013). Evaluación de un instrumento de medición del nivel de satisfacción estudiantil en los cursos de formación superior de la FCCEEyA de la UDELAR a través de la aplicación de Structural Equation Modelling (SEM). Informe de Pasantía, Universidad de la República.
