Convert deviances to thresholds

Converts an OpenMx model in which ordinal thresholds are parameterized as deviances between successive thresholds to an equivalent model in which the parameters are thresholds.

Usage

deviances_to_thresholds(model)

Arguments

model

An MxModel containing ordinal indicators whose thresholds are specified using deviance parameterization, as created by as_ram with threshold_method = "mx_deviances".

Value

An MxModel.

See also

as_ram

Examples

# Plain model
if(isTRUE(requireNamespace("OpenMx", quietly = TRUE))){
library(OpenMx)

set.seed(1)
dat <- data.frame(
  X1 = ordered(sample(1:2, 200, replace = TRUE)),
  X2 = ordered(sample(1:2, 200, replace = TRUE)),
  X3 = ordered(sample(1:3, 200, replace = TRUE))
)

mod <- as_ram(
  "
  X1 | t1
  X2 | t1
  X3 | t1
  X3 | t2
  ",
  data = dat,
  threshold_method = "mx_deviances"
)
res <- run_mx(mod)
coef(res)
res_thresholds <- deviances_to_thresholds(res)
coef(res_thresholds)

# Multigroup model
set.seed(2)
dat1 <- data.frame(
  X1 = ordered(sample(1:2, 100, replace = TRUE)),
  X2 = ordered(sample(1:2, 100, replace = TRUE)),
  X3 = ordered(sample(1:3, 100, replace = TRUE))
)

dat2 <- data.frame(
  X1 = ordered(sample(1:2, 100, replace = TRUE)),
  X2 = ordered(sample(1:2, 100, replace = TRUE)),
  X3 = ordered(sample(1:3, 100, replace = TRUE))
)

group1 <- as_ram(
  "
  X1 | t1
  X2 | t1
  X3 | t1
  X3 | t2
  ",
  data = dat1,
  threshold_method = "mx_deviances"
)
group1 <- mxModel(group1, name = "group1")

group2 <- as_ram(
  "
  X1 | t1
  X2 | t1
  X3 | t1
  X3 | t2
  ",
  data = dat2,
  threshold_method = "mx_deviances"
)
group2 <- mxModel(group2, name = "group2")

multigroup <- mxModel(
  "multigroup",
  group1,
  group2,
  mxFitFunctionMultigroup(c("group1", "group2"))
)
res2 <- run_mx(multigroup)
coef(res2)
res2_thresholds <- deviances_to_thresholds(res2)
coef(res2_thresholds)
}
#> group1.Thresholds[1,1] group1.Thresholds[1,2] group1.Thresholds[1,3] 
#>            -0.02506889             0.20189348            -0.64334539 
#> group1.Thresholds[2,3] group2.Thresholds[1,1] group2.Thresholds[1,2] 
#>             0.55338476             0.17637416            -0.02506890 
#> group2.Thresholds[1,3] group2.Thresholds[2,3] 
#>            -0.46769879             0.49585034