An annual JCR impact factor calculation based on Bayesian credibility formulas

► We propose a Bayesian approach for impact calculation of a journal, combining both its current prestige and its recent trajectory. ► Credibility theory establishes an adequate mechanism determining how credible is the impact factor calculation. ► We model the citation-quality process by choosing a conjugated family Poisson–Gamma in order to obtain a net impact credibility formula. ► The Subject Category Listing “Information Science & Library Science” of 2009–2010 JRC Social Science Edition (ISI Web of Knowledge) is used for illustration. Our aim deals with appraising the annual impact calculation of journals belonging to the JCR, in terms of the expected citation (with or without selfcitations) by published paper in a range of k-years. A Bayesian approach to the problem, should reflect not only the current prestige of a journal, but also taking into account its recent trajectory. In this wide context, credibility theory becomes an adequate mechanism deciding whether journal's impact factor calculation to be more or less plausible. Under prior belief that journal quality is determined by its impact factor, we model the citation-quality process by choosing a conjugated family of the exponential class in order to obtain a net impact credibility formula. Proposed weighting schema produces the effect of smoothing out any sudden increases or decreases in the year-by-year impact factor.