摘要：This paper stands that Political Science empirical research designs should be crafted in order to produce falsifiable causal inferences. Infer as a process of using available information to know about unavailable data. Causal in the sense that the occurrence of x should changes the probability of occurrence of y. And falsifiable in order that at any moment the inference could be demonstrated wrong by a concurring research design. In addition, explanations should identify the causal mechanism that links explanans and explanandum.
关键词：Causality;Mechanics;Inference;research design.;Causalidade;Mecanismos;Inferência;Desenho de pesquisa.
