This paper presents a Gibbs sampler for the estimation of Thurstonian ranking models. This approach is useful for the analysis of ranking data with a large number of options. Approaches for assessing the goodness-of-fit of Thurstonian ranking models based on posterior predictive distributions are also discussed. Two simulation studies and two ranking studies are presented to illustrate that the Gibbs sampler is a promising solution to the numerical problems that previously plagued the estimation of Thurstonian ranking models.
|Original language||English (US)|
|Number of pages||14|
|Journal||British Journal of Mathematical and Statistical Psychology|
|State||Published - Jan 1 1999|
ASJC Scopus subject areas
- Statistics and Probability
- Arts and Humanities (miscellaneous)