Efficiency of Selecting Important Variable for Longitudinal Data
- 1 Department of Education, Kookmin University, Seoul, South Korea
- 2 Department of Education, Kookmin University, Seoul, South Korea
Abstract
Variable selection with a large number of predictors is a very challenging and important problem in educational and social domains. However, relatively little attention has been paid to issues of variable selection in longitudinal data with application to education. Using this longitudinal educational data (Test of English for International Communication, TOEIC), this study compares multiple regression, backward elimination, group least selection absolute shrinkage and selection operator (LASSO), and linear mixed models in terms of their performance in variable selection. The results from the study show that four different statistical methods contain different sets of predictors in their models. The linear mixed model (LMM) provides the smallest number of predictors (4 predictors among a total of 19 predictors). In addition, LMM is the only appropriate method for the repeated measurement and is the best method with respect to the principal of parsimony. This study also provides interpretation of the selected model by LMM in the conclusion using marginal R 2 .
- Agresti, A. (2002). Categorical data analysis (2nd ed.). Boboken, NJ: John Wiley & Sons.
- Agresti, A., & Finlay, B. (1986). Statistical method for the social sciences (2nd, ed.). San Francisco, CA: Dellen.
- Akaike, H. (1973). Information theory and an extension of the maximum likelihood principle. In B. N. Petrov, & F. Csaki (Eds.), Second international symposium on information theory (pp. 267-281). Budapest: AcademiaiKiado.
- Altman, D. G., & Andersen, P. K. (1989). Bootstrap investigation of the stability of a Coxregression model. Statistics in Medicine, 8, 771-783.
- Bernstein, I. H. (1989). Applied multivariate analysis. New York: Springer-Verlag.
- Bondell, H. D., & Reich, B. J. (2008). Simultaneous regression shrinkage, variable selectionand clustering of predictors with OSCAR. Biometrics, 64, 115-123.
- Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multipleregression/correlation analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Lawrence Erlbaum.
- Derksen, S., & Keselman, H. J. (1992). Backward, forward and stepwise automated subset selection algorithms. British Journal of Mathematical and Statistical Psychology, 45, 265-282.
- Efron, B., Hastie, T., Johnstone, I., & Tibshirani, R. (2004). Least angle regression. The Annals of Statistics, 32, 407-489.
- Fan, J., & Li, R. (2001). Variable selection vianonconcave penalized likelihood and its oracle properties. Journal of the American Statistical Association, 96, 1348-1360.
- Foster, D. P., & George, E. I. (1994). The risk inflation criterion for multiple regression. The Annals of Statistics, 22, 1947-1975.
- George, E. I., & McCulloch, R. E. (1993). Variable selection via Gibbs sampling. Journal of the American Statistical Association, 88, 881-889.
- Gilks, W. R., Wang, C. C, Yvonnet, B., & Coursaget, P. (1993). Random effects models for longitudinal data using Gibbs sampling. Biometrics, 49, 441-453.
- Hosmer, D. W., & Lemeshow, S. (2000). Applied logistic regression. New York: John Wiley& Sons.
- Laird, N., & Ware, J. H. (1982). Random effect models for longitudinal data. Biometrics, 38, 963-974.
- Mallows, C. L. (1973). Some comments on Cp. Technometrics, 15, 611-675.