A characterization of the law of Lotka in terms of sampling

Abstract

In order to model the variable T (the age of citations received by scientific works) with data elaborated by the Institute of Scientific Information, we have used some of the instruments already developed in the survival models to this type of retrospective analyses in the presence of censored data. This analysis is used because, usually, the citations of ages greater than or equal to 10 years appear added together. For a set of journals related to the field of Applied Economics, we have explored which models fit better among those commonly used. Two different approaches to assess the goodness-of-fit for each selected model have been suggested: an analysis through graphical methods and a formal analysis to estimate the parameters of each model by the method of maximum likelihood estimation with data censored to the right.