A Comprehensive Statistical Analysis and Modification of Impact Factor of Medical Journals

Impact Factor (IF) is a bibliometric method used for evaluating the published work of scientists. In 2012, a research of Adriano B. L. Tort, Ze H. Targino, and Olavo B. Amaral affirmed that the time lag from online to print publications, which defined as publication window of journal (PWj), could unintentionally raise the value of a journal's IF. This inflation is greater for longer online-to-print time delays. This study continuously provided more evidence to point out the influence of PWj to IF calculation. Our study found that the 1-year IF has the effect on PWj with the significant finding Spearman of rank correlation 0.352 (p=0.0002). The citation distribution skewed on the right, and IF of a journal was stable in the period 2009-2010( p=0,1411). Interestingly, all indices including article influence score, eigenfactor score, total cites, self-cites, cited-half-life showed no statistical significance without the PWj after adjusted for six independent variables in regression model. We suggest that IF computation should be calculated in the median method, which is suitable for non-normal distribution of citation numbers. All types of papers should be counted in the denominator of IF formula with the median of these types of articles.