Cat objects are reverted to the bounds of
[-5,5]. Likewise, the
Cat class defaults are now
Newton Raphson bug fixed. Bug was causing estimates when no questions have been answered to go to extremes rather than prior mean.
simulateThetas() has a new argument (defaulted to
FALSE for backwards compatibility) that when TRUE returns a list of
dataframes containing adaptive answer profiles for each Cat objected
involved in the simulation
Cat objects now use integration bounds of
[-4,4]. Likewise, the
Cat class defaults are now [-4,4].
This narrows the bounds of integration from [-5,5] to avoid
computational issues that arise at “extreme” values of the latent
z slot still defaults to .9,
but now in calculating delta for certain integration routines, the
package executes qnorm(z).
oracle() function adds option for parallel computing
probability() for categorical data now throws errors
to account for extreme values of latent trait that may cause
simulateRespondents() bug fix when respondent’s
answer in raw data is
NA, now transform to -1 to indicate a
added dataset of Need to Evaluate raw response profiles
readQualtrics() now has respondent ID as rownames
instead of a column.
Added example data for the
oracle() allow for simulation exercises to evaluate model
quality and performace.
plot.Cat() allows for visual
representation of item parameters.
readQualtrics() aid the user
in creating an adaptive battery in Qualtrics using
ids was added to the
object representing each question item’s unique identifier.
selectItem() returns a third item,
next_item_name which represents the unique identifier of
the item that should be asked next.
lookAhead() now returns the next best item given the
question is skipped.
simulateThetas()allow for estimation of ability parameter for dataframe of response sets.
gpcmCat()by adding the
ltmpackage to Imports.
methodsimports as it caused errors in testing with r-oldrel.