A Note on Empirical Sample Distribution of Journal Impact Factors in Major Discipline Groups
34 Pages Posted: 16 Feb 2010
There are 2 versions of this paper
A Note on Empirical Sample Distribution of Journal Impact Factors in Major Discipline Groups
A Note on Empirical Sample Distribution of Journal Impact Factors in Major Discipline Groups
Date Written: February 14, 2010
Abstract
What type of statistical distribution do the Journal Impact Factors follow? In the past, researchers have hypothesized various types of statistical distributions underlying the generation mechanism of journal impact factors. These are: lognormal, normal, approximately normal, Weibull, negative exponential, combination of exponentials, Poisson, Generalized inverse Gaussian-Poisson, negative binomial, generalized Waring, etc. It is pertinent to note that the major characteristics of JIF data lay in the asymmetry and non-mesokurticity. The present study, frequently encounters Burr-XII, inverse Burr-III (Dagum), Johnson SU, and a few other distributions closely related to Burr distributions to best fit the JIF data in subject groups such as biology, chemistry, economics, engineering, physics, psychology and social sciences.
Keywords: Journal impact factor, JIF, Theoretical distribution, Burr, Dagum, Generalized extreme value, generalized gamma, Inverse Gaussian, Johnson SU, Johnson SB, Kumaraswamy, Log-logistic, lognonmal, log-Pearson, Weibull, Generalized normal, Hypersecant, Beta
JEL Classification: C, C4, A14, Z
Suggested Citation: Suggested Citation