Goodness of fit tests of the two-parameter gamma distribution against the three-parameter generalized gamma distribution.

The generalized Gamma distribution is widely used in survival analysis, engineering and economics, and it includes the Gamma distribution as a special case. In real life applications, the Generalized Gamma distribution may be an unnecessarily complicated distribution to fit when one of the simpler models may be sufficient to analyze the data. The aim of this paper is to develop tests of goodness of fit of the two parameter gamma distribution against its three parameter counterpart. Specifically, we develop Wald, C (α) , score, likelihood ratio, and gradient tests. These tests are then compared in terms of empirical size and power using a simulation study. [ABSTRACT FROM AUTHOR]