Errors in categorising exposure, outcome, or diagnosis distort a study's results.
Misclassification bias arises when participants are incorrectly assigned to an exposure, treatment, or outcome category (e.g., a diagnostic tool mislabels injured vs. uninjured athletes). If the error rate is similar across comparison groups (non-differential), it usually — but not always — biases results toward the null; if the error differs by group (differential), it can bias results in either direction and is generally more concerning. The direction and size of the effect can be hard to predict with certainty, so this term is often flagged as a limitation rather than precisely quantified.
This guide was auto-drafted and is pending editorial review.