Observer bias is a broad category of bias that occurs when the method or process of observation results in systematic differences from the truth. Practically, this often occurs because the research “sees what she wants to see”. For example, when performing a chart review, there are often multiple observations recorded in the notes. If the researcher looking at aortic dissection, they may choose the one description that states the pain is “severe” and ignore other descriptions, because that is what they expected to see. The interpretation of almost all medical tests is somewhat subjective, and therefore observer bias is almost always possible.
The primary methodological tool used to limit observer bias is blinding of outcome assessors.
Observer bias is a type of measurement bias.
Observer bias is also referred to as experimenter bias or researcher bias.
This post is part of a series of posts on bias in medical research. You can find the whole bias catalogue here.