How to design a study that everyone will believe

Bias and confounding are major threats to study validity. Bias refers to a systematic error in a study, which can occur during participant selection, data collection, analysis, or reporting. Confounding refers to study bias caused by confounding variables, which are associated with the independent and dependent variables of interest. Randomization is one way to minimize confounding, but there are others. In this chapter, I describe types of bias, further define confounding, and describe ways to minimize the influences of bias and confounding so that your estimates of effect are as valid as possible.