Topics and controversies in biostatistics
← vista completaPublished on September 10, 2014 | http://doi.org/10.5867/medwave.2014.08.6016
Causality and prediction: differences and points of contact
Causalidad y predicción: diferencias y puntos de contacto
Abstract
This contribution presents the differences between those variables that might play a causal role in a certain process and those only valuable for predicting the outcome. Some considerations are made about the core intervention of the association and the temporal precedence and biases in both cases, the study of causality and predictive modeling. In that context, several relevant aspects related to the design of the corresponding studies are briefly reviewed and some of the mistakes that are often committed in handling both, causality and prediction, are illustrated.