The stratified literature produced from 1922 to 2003 on Lotka's law is analyzed by using the Poisson lognormal and the Poisson Gauss inverse generalized models. For this purpose, three types of how the literature produced is counted are used: straight, complete and fractioned ones. The data of the research are evaluated by using the chi-square test at 0.05 significance level. Both models fit very well into the distribution of literature produced, but the Poisson Gauss inverse generalized distribution produces a smaller chi-square and better anticipates the total number of authors than the Poisson Lognormal distribution.
Lotka's law; Authors' productivity; Scientometrics; Bibliometrics; Informetrics; Poisson lognormal model
