M. Regúlez-Castillo, J. M. Pérez-Salamero González, C. Vidal-Melia
This paper proposes a procedure for selecting large subsamples drawn from a simple random sample to be more representative of the population under study than the original one. The mathematical approach corresponds to a problem of Convex Mixed Integer Nonlinear Programming (Convex MINLP). The procedure seeks to maximize the size of the subsample taken from a stratified random sample with proportional allocation, restricting it to a p-value high enough to achieve a good fit using Pearson’s chi-square goodness of fit test. Whenever the problem is feasible, the procedure obtains, in a short time, the global solution. By means of a simulation exercise, the procedure is applied successfully to 4,000 cases of stratified populations. Finally, using the Continuous Sample of Working Lives (CSWL) we show the utility of the procedure in a real case application.
Palabras clave / Keywords: chi-square test, Continuous Sample of Working Lives, optimization, p-value, subsampling
Programado
Pósteres I
30 de mayo de 2018 15:30
Zona EXPO
