Bias

May result from flaws in the design of a study or in the analysis of results and may result in either an underestimate or an overestimate of the effect.