Skip to contents

Ordered Logit Response Model

Source: R/model_ordered.R
ordered_logit.Rd

Ordered responses based on a latent logistic evidence variable. Binary response models are the 2-category special case.

Usage

ordered_logit()

Value

A model list with all the necessary functions to sample

Details

The scale parameter is the latent logistic scale, and thresholds follow the same ordered K - 1 parameterization as ordered_probit(), with derived ordered thresholds returned in cut_expanded.

Examples

dord <- design(
  Rlevels = c("left", "right"),
  factors = list(subjects = 1, S = c("left", "right")),
  formula = list(location ~ 0 + S, scale ~ 1, cut ~ 1),
  matchfun = function(d) d$S == d$lR,
  constants = c(scale = log(1)),
  model = ordered_logit
)
#> 
#>  Sampled Parameters: 
#> [1] "location_Sleft"  "location_Sright" "cut"            
#> 
#>  Design Matrices: 
#> $location
#>      S location_Sleft location_Sright
#>   left              1               0
#>  right              0               1
#> 
#> $scale
#>  scale
#>      1
#> 
#> $cut
#>  cut
#>    1
#>