Main content



Loading wiki pages...

Wiki Version:
Bayesian inference allows to assess whether a claim about an effect (e.g., effect > 0, effect > Δ) is justified based on a data estimate and a prior distribution that expresses an individual’s information, assumption, or belief about the effect before seeing the data. Recipients of a Bayesian inference, however, have different or differently assessed information and accordingly vary in their prior distributions. Thus, it remains unclear whether they should agree on deciding whether the claim is supported or not. “Reverse-Bayes analysis" and the "sufficiently sceptical prior" address this problem by providing how strongly one may believe in the non-existence of an effect to be convinced by the data to the contrary. A graphical tool called *Region of Evidence (RoE)* is presented. It shows all the normal priors for which, given a normally distributed estimate, the *posterior probability* that a claim is true exceeds 1 – α, that is, where one may decide to accept a claim. A RoE assesses sensitivity by covering any prior that one might have, including advocacy priors that favour a positive or negative effect and cases where the support for a claim might largely come from the prior. A RoE visualises all the priors that are in line with a claim of superiority or non-inferiority, assuming that the prior and an effect estimate are normally distributed. Since RoE only requires an effect estimate and its standard error, it can be easily applied to previously published results. The paper describes an open-source implementation in R and a stand-alone web-application for its use outside of R. Its utility is demonstrated with the clinical example of Esketamine treatment in major depression patients, where the RoE method illustrates how the benefit appraisals of different German stakeholders translate into formal priors that are compatible with their diverging claims. Reasons for the Bayesian approach to data analysis include ● interpretability: it yields an explicit probability (posterior probability) that a claim (i.e., a parameter is smaller or larger than a specific relevance threshold Δ) is true (given the prior distribution on the parameter, the data and the model that describes the data ● transparency of prior assumptions ● clarification of misconceptions about frequentist p-values (Wagenmakers et al., 2018). ● accounting for complexity ● stabilisation of estimates, especially in small samples with many predictors ● ability to incorporate assumptions about biases in estimates **Implementations in R and Stata** The RoE method is implemented in an open-source package called [BayesROE][1] for R statistical software. A stand alone version is accessible [here][2]. An implementation in Stata can be found [there][3] or in a syntax container in [Code Ocean][4]. Greenland, S. (2005). Multiple-bias modelling for analysis of observational data (with discussion). Journal of the Royal Statistical Society Series A (Statistics in Society) 168(2), 267-306. [] Greenland, S. (2006). Bayesian perspectives for epidemiological research: I. Foundations and basic methods. International Journal of Epidemiology 35(3), 765–775, [] Marsman, M., & Wagenmakers, E.-J. (2017). Three insights from a Bayesian interpretation of the one-sided P value. Educational and Psychological Measurement, 77(3), 529–539. [] Mayo, D. G. (2018). Statistical Inference as Severe Testing: How to Get Beyond the Statistics Wars. Cambridge: Cambridge University Press. [] O'Hara, R. B., & Sillanpää, M. J. (2009). A review of Bayesian variable selection methods: what, how and which. Bayesian Analysis 4(1), 85-117. [] Wagenmakers, E.-J., Marsman, M., Jamil, T., Ly, A., Verhagen, J., Love, J., Selker, R., Gronau, Q. F., Šmíra, M., Epskamp, S., Matzke, D., Rouder, J. N., & Morey, R. D. (2018). Bayesian inference for psychology. Part 1: Theoretical advantages and practical ramifications. Psychonomic Bulletin & Review, 25(1), 35–57. [] [1]: [2]: [3]: [4]:
OSF does not support the use of Internet Explorer. For optimal performance, please switch to another browser.
This website relies on cookies to help provide a better user experience. By clicking Accept or continuing to use the site, you agree. For more information, see our Privacy Policy and information on cookie use.

Start managing your projects on the OSF today.

Free and easy to use, the Open Science Framework supports the entire research lifecycle: planning, execution, reporting, archiving, and discovery.