Explicit documentation of publication and other reporting biases requires availability of protocols, data, and results of primary analyses from conducted studies so that these can be compared against the published literature. However, these are not often available. A few empirical studies have retrieved study protocols from authors or trial data from FDA submissions [6–9]. These studies have shown that deviations in the analysis plan between protocols and published papers are common [6], and effect sizes of drug interventions are larger in the published literature compared with the corresponding data from the same trials submitted to FDA [7, 8]. It is challenging, but more common, to detect these biases using only the published literature. Several tests have been developed for this purpose; some are informal, others use statistical methods of variable rigor. These tests may be applied to single studies, multiple studies on specific questions, or wider disciplines.
