Modeling Consumer Purchase Events: A Reply to Lawrence

Abstract

In the May 1980 issue of JMR , Lawrence replaces the gamma distribution on the purchasing rate component of the negative binomial model with the lognormal distribution. In this response the author shows that Lawrence's modeling effort is incomplete and that his data analysis is inconsistent with his implied model. In addition it is argued that the gamma distribution per se has some major advantages with respect to the lognormal distribution proposed by Lawrence.