Biases in Scientific Inquiry

The purpose of this paper is to review, organize, and analyze disparate kinds of biases that can impact the quality of scientific research. We offer an analysis of what constitutes a bias and then provide a novel taxonomy of biases in science based on the criteria used to individuate them in the literature, i.e., based on the different criteria researchers use to identify biases. We identify five criteria: mechanism, effect, content, stage, and feature. We then conceptualize the impact of different biases using the notion of a ‘research pipeline.’ Furthermore, we identify a type of bias that can permeate the research pipeline as a whole, called ‘systemic bias,’ which includes biases that are likely to be more pervasive and thus more difficult to counteract. We conclude by stressing the importance of classifying biases in ways that are epistemically useful for science.