Análisis Estadístico de Covarianza Múltiple no Paramétrico y su aplicación en la biomedicina

This work has a like a had the object to present alternatives of multiple covariance analysis, valid in situations where the answer variable is quantitative and the covariables are qualitative (ordinals or nominals). In this sense, it is presented a minute study from ANCOVA Proposal Multiple Non-Parametric (extension of the method of Conover and Iman) applicable to situations where the covariables are ordinal or nominal and the answer variable is quantitative, is presented. The proposed proceeding consists of transforming in ranks all the experimental observations made from what the stability and high power of test demonstrated, obeys to the fact that this proceeding diminishes the variability, solving the problems that causes the lack of homogeneity, in addition is less sensible to the nonnormality of the answer. This is demonstrated by the analytical form compare it with his accredit parametric and through the tests powers obtained for each one of the experiments. In oreder to validate