Abstract

ACOSTA AVENA, L. M  and  SALAZAR URIBE, J. C. Evaluating the Presence of Confounding in Some Members of the Exponential Family. Ciencia en Desarrollo [online]. 2015, vol.6, n.1, pp.99-112. ISSN 0121-7488.

A technique to detect confounding, which is recommended in the literature to detect confounders in the Cox models, is adapted and applied to some models of the exponential family. The aim is to assess whether this technique is useful in the logistic, Poisson and linear models. An alternative technique, based on the odds, is proposed for the case of the logistic model. The results show that the techniques have a similar behavior when confusion is present in the model; however, the proposed technique proved to be more effective in scenarios in which this phenomenon was absent.

Keywords : Confounding; Exponential Family; Cox Models; Regression.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )