GDINA (version 1.4.2)

modelcomp: Item-level model comparison using Wald test

Description

This function evaluates whether the saturated G-DINA model can be replaced by reduced CDMs without significant loss in model data fit for each item. See de la Torre and Lee (2013), and Ma, Iaconangelo and de la Torre (2016) for details. This function also calculates the dissimilarity between the reduced models and the G-DINA model, which can be viewed as a measure of effect size (Ma, Iaconangelo & de la Torre, 2016).

Usage

modelcomp(GDINA.obj, item = "all", DS = FALSE, SE.type = 2,
  models = c("DINA", "DINO", "ACDM", "LLM", "RRUM"), varcov = NULL)
# S3 method for modelcomp
extract(object, what = c("wald", "wald.p", "df", "DS",
  "models"), digits = 4, ...)
# S3 method for modelcomp
summary(object, ...)

Arguments

GDINA.obj

An estimated model object of class GDINA

item

a vector of items to specify which items the Wald test is applied to

DS

whether dissimilarity index should be calculated? FALSE is the default.

SE.type

the type of standard error estimates.

models

a vector specifying which reduced CDMs are possible reduced CDMs for each item. The default is "DINA","DINO","ACDM","LLM",and "RRUM".

varcov

Optional; user specified variance-covariance matrix. If supplied, it must be a list of length \(J\), giving the variance covariance matrix of item success probability for each item. The default is NULL, in which case, the estimated variance-covariance matrix from the GDINA function is used.

object

object of class modelcomp for various S3 methods

what

argument for S3 method extract indicating what to extract; It can be "wald" for wald statistics, "wald.p" for associated p-values, "df" for degrees of freedom, and "DS" for dissimilarity between G-DINA and other CDMs.

digits

How many decimal places in each number? The default is 4.

...

additional arguments

Value

an object of class modelcomp. Elements that can be extracted using extract method include

wald

wald statistics

wald.p

p-values associated with the wald statistics

DS

dissimilarity between G-DINA and other CDMs

Methods (by generic)

• extract: extract various elements from modelcomp objects
• summary: print summary information

References

de la Torre, J., & Lee, Y. S. (2013). Evaluating the wald test for item-level comparison of saturated and reduced models in cognitive diagnosis. Journal of Educational Measurement, 50, 355-373. Ma, W., Iaconangelo, C., & de la Torre, J. (2016). Model similarity, model selection and attribute classification. Applied Psychological Measurement, 40, 200-217.

See Also

GDINA, autoGDINA

Examples

## Not run: ------------------------------------
# dat <- sim10GDINA$simdat
# Q <- sim10GDINA$simQ
# # --- GDINA model ---#
# mod1 <- GDINA(dat = dat, Q = Q, model = "GDINA")
# mod1
# wmod1 <- modelcomp(mod1)
# wmod1
# # wald statistics
# extract(wmod1,"wald")
# #p values
# extract(wmod1,"wald.p")
# wmod1r <- modelcomp(mod1,DS=TRUE)
# #dissimilarity index
# extract(wmod1r,"DS")
# wmod2 <- modelcomp(mod1,models = c("DINA","DINO"))
# wmod2
# wmod3 <- modelcomp(mod1,item=c(8,9,10),DS=FALSE)
# wmod3
## ---------------------------------------------