- Fixed tests and vignette that failed due to changes in package mirt version 1.32.1.
augment()
now consistently prefixes new columns with.fitted.
or.se.fit.
. Furthermore, its argumentse.fit
was renamed tose_fit
(as recommended by package modeltests).- Uncoupled README from the vignette; README is now much shorter.
- R (>= 3.6.0) is now required in alignment with the dependency mirt.
- Condition handling was refactored (e.g.,
myTryCatch()
was replaced withsimsalapar::tryCatch.W.E()
). sym_diff()
was replaced withsets::set_symdiff()
, andsort2()
was removed.
- Minor, internal changes in the documentation
- Modified package Description field for CRAN
- Modified vignette,
install.packages
was commented out. - Credit to Jackson for the
jackson
data and Dontas forsort2()
is now given via appropriate attribution rather than usage of\author
.
- Modified package Title field for CRAN
- In the model syntax, the section Subtree is deprecated and was merged with section Constraints.
- Long variable names are no longer automatically replaced. They are now accepted for mirt and TAM, but no longer for Mplus.
logit
is now the default link function everywhere.mirt
is now the default engine infit()
.
- More complex models using constraints (that were only available for Mplus) are now implemented for mirt and TAM.
- The new function
irtree_create_template()
helps to create a model string. - Data set
jackson
(Big Five questionnaire) was added. - Many internal changes to increase robustness.
- Controlling features of the engine in
fit()
is now done via a singlecontrol
argument (rather than...
), and helper functionscontrol_<engine>()
make that painless. - The documentation was improved.
- The output of
tidy()
was enhanced (added columnsparameter
andcomponent
), and the argumentpar_type
is now required for mirt.
- Add wrapper functions for the TAM package
- Add the multidimensional partial credit model (not an IR-tree model); works only in TAM
- First stable version