Modeling the probabilistic distribution of the impact factor

The study of the informetric distributions, such as distributions of citations and impact factors is one of the most relevant topics in the current informetric research. Several laws for modeling impact factor based on ranks have been proposed, including Zipf, Lavalette and the two-exponent law proposed by Mansilla et al. (2007). In this paper, the underlying probabilistic quantile function corresponding to the Mansilla's two-exponent law is obtained. This result is particularly relevant, since it allows us to know the underlying population, to learn about all its features and to use statistical inference procedures. Several probabilistic descriptive measures are obtained, including moments, Lorenz and Leimkuhler curves and Gini index. The distribution of the order statistics is derived. Least squares estimates are obtained. The different results are illustrated using the data of the impact factors in eight relevant scientific fields.

► The theoretical probabilistic distribution of the two-exponent law (Mansilla et al., 2007) is obtained.
► Several probabilistic features are derived.
► Estimation methods based on least squares are discussed.
► Applications with data taken from Journal Citation Reports Edition 2010 are considered.