Statistical Investigation of Scientific Journals Impact Factors in Relation to Benford's Law

Abstract

Benford's law (BL) has been used as a tool for the detection of possible data manipulation in datasets. The statistics of the first digit(s) of the observed values are compared to a set of expected values given by BL, and violations signal possible data anomalies. Though this method is first used for economical purposes, it has been recently used in bibliometric statistical analyses. The method provides alert that records may have certain unusual trends. In the present work, we investigate the impact factor statistics for 24 years starting from 1997 up to 2021, for open access and subscribed journals. We study conformity to BL using the basic first digit(s) violation test and the mantissa arc test, besides using the relative size factor test. Results show that open access journals adhere to Benford's Law more closely than subscribed journals.