How observed effects in behavioral science should (not) be reported

Empirical studies in behavioral science typically report observed effects by relying on standardized effect size measures (e.g., Cohen’s d). These measures relate the observed mean difference to the observed standard deviation. The latter thus functions as an empirical measurement unit. Since the observed standard deviation is conceptually related to the quality of measurement, the interpretation and the meta-analytically aggregation of standardized observed effects require a transparent account of the measurement scale reliability, the measurement error, the quality of the empirical setting, and how it is standardized. Even top-tier journals, however, do not demand that published studies provide such an account. This arguably places a cumulative science of human behavior farther out of reach. Standardized observed effects, therefore, should be reported alongside a transparent error account.