sdt_rating_logmu: Rating SDT category log-probability (multinomial logit)

In bmm: Easy and Accessible Bayesian Measurement Models Using 'brms'

Rating SDT category log-probability (multinomial logit)

Description

R companion to the Stan sdt_rating_logmu function. It returns log(p_k) for rating category cat, which the sdt_rating() multinomial formula uses as the category logit (so softmax recovers the SDT category probabilities). brms evaluates the non-linear formula in R for posterior_predict() and posterior_epred(), so this function must be on the search path; it is exported for that reason and is not called directly.

Usage

sdt_rating_logmu(
  cat,
  K,
  dist,
  thresh,
  anchor,
  d,
  criterion,
  spacing,
  sdratio,
  stimulus,
  ...
)

sdt_dpsdt_logmu(
  cat,
  K,
  dist,
  thresh,
  anchor,
  d,
  criterion,
  spacing,
  sdratio,
  Ro,
  Rn,
  stimulus,
  ...
)

sdt_metad_logmu(
  cat,
  K,
  dist,
  thresh,
  anchor,
  d,
  criterion,
  spacing,
  sdratio,
  logmratio,
  stimulus,
  ...
)

Arguments

cat	Integer rating category index.
K	Integer number of rating categories.
dist	Integer noise-distribution id (see the .sdt_dists registry).
thresh	Integer threshold-parameterization id.
anchor	Integer: 1 when criterion is measured from the noise mean, 2 when it is the bias parameter, measured from the midpoint between the means.
d, criterion, spacing, sdratio	Model parameters (draws-by-observation matrices supplied by brms). d is d', or d_a when sdratio is not 0; spacing is 0 for threshold types without it.
stimulus	Stimulus covariate (0 = noise, 1 = signal).
...	Threshold delta parameters, when the threshold type uses them.
Ro, Rn	Recollection parameters of the dpsdt version on the logit (linear-predictor) scale; inv_logit(Ro)/inv_logit(Rn) are the recollection and recall-to-reject probabilities.
logmratio	Log M-ratio for the metad version; meta-d' is recovered as exp(logmratio) * d, on the same scale as d.

Value

log(p_cat), matching the shape of d.

bmm documentation built on Oct. 16, 2026, 9:08 a.m.