Abstract
Ordered response models are used to analyse surveys in which respondents are asked to measure something (e.g. their life-satisfaction) using ranks. Ordered outcomes (`y`) ranked in a way `y/in /{1, 2, ..., J/}`, where `J` is the total number of distinct categories.
The idea behind ordered probit model is that the observed variable represents ranges of an unobserved latent variable for which cutoffs are unknown. http://www.cerge-ei.cz/research/applied_sw/bivord/