RESUMEN
The nature of the relationship between ability and performance is of critical importance for admission decisions in the context of higher education and for personnel selection. Although previous research has supported the more-is-better hypothesis by documenting linearity of ability-performance relationships, such research has not been sensitive enough to detect deviations at the top ends of the score distributions. An alternative position receiving considerable attention is the good-enough hypothesis, which suggests that although higher levels of ability may result in better performance up to a threshold, above this threshold greater ability does not translate to better performance. In this study, the nature of the relationship between cognitive ability and performance was examined throughout the score range in four large-scale data sets. Monotonicity was maintained in all instances. Contrary to the good-enough hypothesis, the ability-performance relationship was commonly stronger at the top end of the score distribution than at the bottom end.
Asunto(s)
Logro , Pruebas de Aptitud/estadística & datos numéricos , Evaluación Educacional/métodos , Evaluación Educacional/estadística & datos numéricos , Empleo/estadística & datos numéricos , Selección de Personal/métodos , Aptitud/fisiología , Estudios de Cohortes , Escolaridad , Humanos , Estudios Longitudinales , Selección de Personal/estadística & datos numéricos , Análisis de Componente Principal , Reproducibilidad de los Resultados , Estudiantes/psicologíaRESUMEN
Critics of educational admissions tests assert that tests measure nothing more than socioeconomic status (SES) and that their apparent validity in predicting academic performance is an artifact of SES. The authors examined multiple large data sets containing data on admissions and related tests, SES, and grades showing that (a) SES is related to test scores (r = .42 among the population of SAT takers), (b) test scores are predictive of academic performance, and (c) statistically controlling for SES reduces the estimated test?grade correlation from r = .47 to r = .44. Thus, the vast majority of the test?academic performance relationship was independent of SES: The authors concluded that the test?grade relationship is not an artifact of common influences of SES on both test scores and grades.