Confounding is a kind of bias which occurs in a research. Confounding is less frequent in randomized controlled trials (RCT) for evaluation of influenza vaccines. However, there are obstacles or difficulties in conducting RCT for evaluation of influenza vaccines, particularly, in the elderly people. Therefore, a retrospective or prospective cohort study has been primarily performed to evaluate effectiveness of influenza vaccine in elderly people. Confounding by indication or other confounding exist in most observational studies. Accordingly, at the stage of designing or analyzing a study, confounding should be controlled with a restriction, matching, stratified or multivariate analysis technique.