Background: The systematic error introduced by summarizing data that are not representative of the available evidence is commonly referred to as publication bias. Although it has been acknowledged that this term is inaccurate and alternatives have been proposed, to date no coherent terminological framework has been developed. Issues: The term publication bias is unfortunate in two respects: it insufficiently differentiates between biased processes and biased results, and it is inaccurate in that it describes more than just the bias induced by selective publication. Alternatives: A distinction can be made between selective processes and their respective biases. Process-selectivity can be present in the reporting of outcomes and studies, the publication of studies in paper-based and electronic media, and the inclusion of studies in databases and reviews. These selective processes can but do not necessarily lead to reporting bias, publication bias, and inclusion bias, respectively. Instead of using publication bias as a pars pro toto – a term that names a part to describe the whole – dissemination bias is more appropriate as a summary term. The distinction between selective processes and biased results described above can also be applied to statistical methods in meta-analysis, differentiating methods that assess data trends associated with a selective dissemination process from methods that attempt to correct for the bias that may have been induced by the selective dissemination. Rank correlation and regression tests belong to the former category, and selectionmodels or imputation techniques that provide bias-corrected estimates belong to the latter. Conclusions: The proposed terminological framework (Figure 1) explicitly defines selective processes in the dissemination of evidence and their potentially resulting biases. The ideas are extended to the classification of meta-analytical statistical methods that deal with selectivity and bias. The terminology can be implemented in the planning and reporting of meta-analytical studies.