Job search duration data analysis with a Gamma-Gamma distribution model

Abstract

Job search duration is an important subject in labor economics, and studies analyzing the impact of job seeker attributes and labor market characteristics on job search duration often involve statistical modeling based on user behavior data. However, current models have limitations, with simple models providing poor fits and complex models being difficult to fit. To address these issues, a new probability model has been proposed for analyzing job search duration. This model uses a simple probability density function with two parameters to represent the probability distribution of event intervals in a unidimensional point process. It is based on a constrained gamma-gamma distribution and can represent a wider range of cases than the exponential distribution model does. The proposed model showed a better fitting compared to current models, such as the exponential and gamma distribution models. Additionally, the use of Bayesian information criterion for model selection allows for adjusting the granularity scale of job seeker segmentation based on attributes. This new model has the potential to improve the accuracy of labor economics research on job search duration.