Evidence of Experimental Bias in the Life Sciences: Why We Need Blind Data Recording

• DOI: 10.1371/journal.pbio.1002190
• Publication Date: 2015-07-08T17:52:18Z
• Authors: Luke Holman; Megan L. Head; Robert Lanfear; Michael D. Jennions
• Keywords: Reporting bias; Response bias; Confirmation bias

Observer bias and other "experimenter effects" occur when researchers' expectations influence study outcome. These biases are strongest researchers expect a particular result, measuring subjective variables, have an incentive to produce data that confirm predictions. To minimize bias, it is good practice work "blind," meaning experimenters unaware of the identity or treatment group their subjects while conducting research. Here, using text mining literature review, we find evidence blind protocols uncommon in life sciences nonblind studies tend report higher effect sizes more significant p-values. We discuss methods urge researchers, editors, peer reviewers keep mind.