The bradley-terry model for handling categorical response variables from farmer participatory trials

Abstract

This paper looks at responses from participatory on-farm farmer participatory trials that are often measured as ratings (farmers score each given treatment on a scale that is ordered but arbitrary) or rankings (where farmers arrange the treatments in order from most preferred to least preferred). Simple methods such as the preference statistic that uses the proportion of responses, Kruskal-Wallis test which is a one-way analysis of variance by ranks and the Friedman test that is a two-way analysis of variance by ranks are outlined. The Bradley-Terry model for ranks which is a logit model for paired comparisons is described and used to fit models for plot level covariates.