Fixed effects (FE) have emerged as a ubiquitous and powerful tool for eliminating unwanted variation in observational accounting studies. Unwanted variation is plentiful in accounting research because we often use rich data to test precise hypotheses derived from abstract theories. By eliminating unwanted variation, FE reduce concerns that omitted variables bias our estimates or weaken test power. FE are not costless, though, so their use should be carefully justified by theoretical and institutional considerations. FE also transform samples and variables in ways that are not immediately apparent, and in doing so affect how we should interpret regression results. This primer explains the mechanics of FE and provides practical guidance for the informed use, transparent reporting, and careful interpretation of FE models.
