Multilevel wage prediction with graduate survey data

L. E. Vila Lladosa, M. Caballer Tarazona

Human Capital theory predicts individual wages from years of education and work experience (Mincer, 1974). Nonetheless, substantial differences in sample wage (mean & variance) for groups of graduates suggest that clusters (hierarchy) may be relevant to predict wages with graduate survey data. Two theoretical models compete to predict graduate wages for ML analysis of the response variable. First, signaling theory predicts individual wages from HE credentials; that is, wage reflects the average labor-market value of diverse types of educational credentials (assumption: credential as screening device for workers productivity). Second, segmentation theory predicts individual wages combining HE credentials and job segments. Wage reflects the average labor-market value of diverse HE credentials within diverse job segments (assumption: credentials as criteria to assign workers to job segments).

Palabras clave: productivity, segmentation, signalling

SI-DE-1 Sesión Invitada. Analítica de Datos en Educación
6 de septiembre de 2019  11:20
I3L10. Edificio Georgina Blanes

J. V. Benlloch Dualde, L. G. Lemus-Zúñiga

I. Baeza-Sampere, L. E. Vila Lladosa

J. M. Carot Sierra, A. Conchado Peiró, E. Vazquez Barrachina, C. Castro López