A1542
Title: A Bayesian panel data study on socioeconomic factors influencing exercise habits in Japan
Authors: Tomoki Toyabe - Kanazawa Gakuin University (Japan) [presenting]
Makoto Nakakita - RIKEN (Japan)
Sakae Oya - Keio University (Japan)
Naoki Kubota - Keio Univeristy (Japan)
Teruo Nakatsuma - Keio University (Japan)
Abstract: Bayesian estimation of a panel logit model for binary longitudinal data using Markov chain Monte Carlo is examined. As an empirical application, panel data on exercise habits from the Japanese population is analyzed. Building on related existing approaches, the Ancillarity-Sufficiency Interweaving Strategy is incorporated into the MCMC algorithm to improve computational efficiency. An application of this approach to Japanese data is provided and its practical usefulness for the analysis of binary panel outcomes is illustrated.
