Bias

Abstract

There are several reasons (e.g. sources of bias), why a study’s results may deviate from the truth. Recruitment and/or follow-up, which is non-random as to who does not provide data is a major source of bias. Other sources include observer preferences and the subject’s answers being informed, for example, by their current or prior status. It is not easy to undertake the perfect, bias-free, study. The challenge is to design the study to minimize the likelihood of bias and to undertake the necessary subanalyses, where appropriate, to obtain a sense of how important a particular bias might have been in explaining the results obtained.