Includes marginal maximum likelihood estimation and joint maximum
likelihood estimation for unidimensional and multidimensional
item response models. The package functionality covers the
Rasch model, 2PL model, 3PL model, generalized partial credit model,
multi-faceted Rasch model, nominal item response model,
structured latent class model, mixture distribution IRT models,
and located latent class models. Latent regression models and
plausible value imputation are also supported. For details see
Adams, Wilson and Wang, 1997

If you use `TAM`

and have suggestions for improvement or have found bugs, please email me at [email protected].

The official version of `TAM`

is hosted on CRAN and may be found here.
The CRAN version can be installed from within R using:

`utils::install.packages("TAM")`

The version hosted here is the development version of `TAM`

.
The GitHub version can be installed using `devtools`

as:

`devtools::install_github("alexanderrobitzsch/TAM")`

**** CHANGES IN TAM PACKAGE VERSIONS ****

++++++++++++++++++++++++++++++++++++++++++++++++++++++++

Test Analysis Modules (TAM) A. Robitzsch, T. Kiefer, & M. Wu

```
___ ___
___ / /\ /__/\
/ /\ / /::\ | |::\
/ /:/ / /:/\:\ | |:|:\
```

/ /:/ / /:/~/::\ **|**|:|:\
/ /::\ /**/:/ /:/:\ /**/::::| :

/**/:/:\ \ :/:/**/ \ :~~_*/
_*/ :\ \ ::/ \ :\

\ :\ \ :\ \ :\

_*/ \ :\ \ :\
_*/ __/

Alexander Robitzsch [email protected]

For reporting a bug, please always provide a reproducible R script and (if necessary) a corresponding dataset.

For TAM tutorials see http://www.edmeasurementsurveys.com/TAM/Tutorials/

+++++++++++++++++++++++++++++++++++++++++++++++++++++++

FIXED * fixed numerical instabilities in latent regression
models caused by variables with very different
standard deviations. Replaced MASS::ginv() by the
internal function tam_ginv_scaled() which is now
exported (thanks to Simon Grund).
NOTE * renamed weighted_curtosis() into weighted_kurtosis()

NOTE * fixed a bug in plot.tam() when response data with
completely missing observations for some persons is
used (thanks to Kow Kuroda)

DATA * included/modified datasets: --- EXAMP * included/modified examples: weighted_Stats (1,2), tam.mml (26)

NOTE * fixed a bug in output of standardized coefficients for multidimensional IRT models ADDED * included function tam.np() for nonparametric estimation of unidimensional item response models NOTE * added S3 method IRT.residuals() for objects of class 'tam.jml' (suggested by Matthew Courtney) NOTE * tamaan() cannot process item names with underscore "_" included. This is now included as an error message for the user (reported by Boris Girnat). NOTE * renamed tam.jml.WLE() into tam_jml_wle() NOTE * now (again) export tam.mml.wle() (used in ?tam.linking) (thanks to Boris Girnat)

DATA * included/modified datasets: --- EXAMP * included/modified examples: tam.np (1), tam.linking (3,4)

ADDED * added "quick" standard error computation method for fitted TAM models of class 'tam.latreg' in tam.se() using helper function tam_latreg_se_quick() (thanks to Yuan-Ling Linda Liaw) FIXED * fixed a bug in tam.wle() which was introduced due to changes in TAM 2.11 (thanks to Anna Scharl) NOTE * added utility functions tam_max_abs() and tam_max_abs_list()

DATA * included/modified datasets: --- EXAMP * included/modified examples: ---

FIXED * fixed a bug in tam.pv.mcmc() when almost singular predictor matrices are used (thanks to Florian Keller) NOTE * included argument 'calc_ic' in tam.pv.mcmc() NOTE * included argument 'groups_by_item' in plot.tam() function (thanks to Laura Helbling) NOTE * changed behavior of 'verbose' argument in tam.mml(), tam.mml.2pl(), ... functions when it is called in other functions (thanks to Hansjoerg Plieninger) NOTE * moved some parts of the M-step computations in tam.mml.2pl() and calculation of probabilities to Rcpp

DATA * included/modified datasets: --- EXAMP * included/modified examples: tamaan (3)

FIXED * fixed problems of numerical overflow in tam.pv.mcmc() (thanks to Domenico Angelone and Florian Keller) NOTE * handled non-vector input for 'theta' in tam.wle() with tam.jml() values (thanks to Andres Alberto Burga Leon) NOTE * changed default in plot.tam() to 'fix.devices=TRUE' (communicated by David Torres Irribarra) NOTE * included Example 25 in ?tam.mml which illustrated different GPCM parametrizations in the TAM, mirt and ltm package NOTE * exported tam.jml.fit() function ADDED * added tam.personfit() function NOTE * updated documentation ?IRT.itemfit. Use now the label RMSD statistic instead of RMSEA fit statistic. NOTE * included ECV statistic in tam.fa() function

DATA * included/modified datasets: --- EXAMP * included/modified examples: tam.mml (25), tam.jml (1), tam.personfit (1)

NOTE * columns with NAs are now omitted from tam.ctt() (thanks to Hiroyuki Yamada) NOTE * compute pseudoinverse matrix of covariates in latent regression model to allow estimation in case of multicollinearity NOTE * avoided error in tam.pv() when the prior only has zeroes in a row NOTE * included class-specific item averages in summary output of tamaan() function for methods LCA, LOCLCA and OLCA NOTE * compute correct number of parameters with non-identified located latent class models (LOCLCA) in tamaan(). The estimation is conducted without constraints, but one non-identified parameter will be subtracted. ADDED * included standardized regression coefficients and explained variance in output for functions tam.mml(), tam.mml.2pl(), tam.mml.mfr() and tam.latreg() with value 'latreg_stand' (suggested by Jana Gross Ophoff) NOTE * renamed tam.mml.se() into tam_mml_se_quick() NOTE * included argument 'item_pars' in tam.se() which indicates whether standard errors should be computed for item parameters (suggested by Karoline Sachse) NOTE * included argument 'stringsAsFactors' in tampv2datalist()

DATA * included/modified datasets: --- EXAMP * included/modified examples: ---

FEATURES:

- include individual class probabilities as an output value 'ind_classprobs' in tamaan() (thanks to Rudolf Debelak)

BUG FIXES:

- fixed a bug in tam.mml.3pl() for models with stochastic nodes (thanks to Bodo Eickhorst and Hansjoerg Plieninger)
- fixed a bug with 'verbose' input of 'T' or 'F' (thanks to Hansjoerg Plieninger)
- fixed a bug in tamaan() with latent classes estimation which was recently introduced due to other bug fixes
- fixed a problem causing a warning message in multiple group models (thanks to Hon-Jang Yang)

NOTES:

- added an example (tam.mml(), Example 20, Model 2) for conversion of design matrices of multidimensional partial credit models from tam.mml() into tam.mml.3pl() (created by Hansjoerg Plieninger)
- allowed stochastic integration and Quasi Monte Carlo integration for multiple group models in tam.mml.3pl() (suggested by Hansjoerg Plieninger)

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

- tam.mml (20.2)

FEATURES:

- included option for providing initial theta values in tam.pv.mcmc()

BUG FIXES:

- fixed a bug in tam.pv.mcmc() which was recently introduced with an updated version of the CDM package
- fixed a bug in tam.pv.mcmc() which caused a crash when only one plausible values is requested
- fixed a bug in output of variance matrices in tam.pv.mcmc()
- included residual correlation matrices in output of tam.pv.mcmc()
- fixed a bug in tam.latreg() caused by subfunction tam_mml_create_nodes() (thanks to Tom Benton)
- fixed a bug in tamaan() (and tam.mml.3pl()) caused by tam_mml_create_nodes() (thanks to Markus Buehner)
- fixed a bug in tam.mml.wle2() when submitting a new dataset (thanks to Plamen Mirazchiyski)

NOTES:

- included residual variances and correlations in summary output of tam.pv.mcmc()
- added argument 'subtract_max' in tam_irf_3pl()
- calculation of item response probabilities for tam.pv.mcmc() now takes place in Rcpp, some speed improvement
- exported utility functions add.lead(), tam_bayesian_bootstrap(), tam_cov_wt() and tam_cor_wt()
- included two options for Bayesian bootstrap (Rubin, 1981) in tam.pv.mcmc(): sampling of case of probabilities and sampling of cases (which corresponds to sampling of integers)
- exported utility functions tam_round_data_frame(), tam_round_data_frame_print(), tam_osink(), tam_csink() and tam_ginv()
- included argument 'file' in summary.msq.itemfit()
- included argument 'select' for weighted statistics (e.g., in weighted_mean(), weighted_sd(), ...)
- included argument 'verbose' in tam.pv()
- included argument 'as_mids' in tam.pv() which allows to convert the set of plausible values into an object of class mids for analysis in the mice package.

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

- tam.pv (1)

FEATURES:

BUG FIXES:

NOTES:

- set 'dontrun' to example 1 of 'tam.latreg' because the CRAN check failed

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

FEATURES:

BUG FIXES:

- argument 'progress' was disabled due to changes in last TAM version and has been fixed (thanks to Sebastian Weirich)

NOTES:

- replaced MASS:mvrnorm by mvtnorm::rmvnorm in functions in the TAM package. Removed MASS package from 'Imports'. Moved GPArotation, lattice, lavaan, plyr, psych and WrightMap packages from 'Imports' to 'Suggests'.
- included utility function 'require_namespace_msg'.
- included d values in output of 'tam.linking'
- exported utility function 'tam_dmvnorm'

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

FEATURES:

- included multidimensional plausible value imputation with multivariate normal approximation of the posterior distribution

BUG FIXES:

- fixed a recently introduced bug in 'tam.wle' since TAM 2.2 (thanks to Mark Dulherty). The function 'tam.mml.wle' is not exported anymore.

NOTES:

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

FEATURES:

- included S3 method 'IRT.frequencies' for computing univariate and bivariate observed and expected frequencies (thanks to a suggestion by Rudolf Debelak)

BUG FIXES:

- fixed a bug in 'tam.linking' (thanks to Sandra Preusser)

NOTES:

- set argument 'beta.fixed' to FALSE as the default in 'tam.latreg'
- added 'print' method for 'tam.linking'

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

- IRT.frequencies (1), tam.linking (3)

FEATURES:

- included function 'tam.pv.mcmc' which allows drawing plausible values based on fully Bayesian estimation. A recommendation for using 'tam.pv.mcmc' in case of stochastic nodes is now displayed in 'tam.pv'. TAM now depends on the coda package for analyzing the MCMC output.
- included linking methods Stocking-Lord and Haebara in the functions 'tam.linking' and 'tam_linking_2studies'

BUG FIXES:

- fixed a recently introduced bug in estimating models with stochastic nodes in TAM functions (thanks to Boris Girnat, Steffen Wild and Reg Allen)

NOTES:

- exported utility functions 'tam_interval_index' and 'tam_rowCumsums'
- disabled argument 'normal.approx' for multidimensional models in 'tam.pv' because it is not properly working
- allowed nonnumeric person identifiers for matching in 'tampv2datlist' function (thanks to Stefan Behrendt)
- restructured 'tam.mml.wle' and 'tam.mml.wle2' functions. Now, WLEs can also be estimated without fitting the model in TAM (see Example 1 in ?tam.wle).
- included a summary method for 'tam.wle'. In addition, more descriptive output is computed as values in 'attr(object, ...)' in 'tam.wle'.
- exported functions 'weighted_mean', 'tam_irf_3pl', 'tam_NA_pattern', 'tam_01_pattern', 'tam_normalize_vector', 'tam_remove_missings', 'tam_AXsi_compute', 'tam_AXsi_fit'

DATASETS INCLUDED OR MODIFIED:

- data.ex16, data.ex17

EXAMPLES INCLUDED OR MODIFIED:

- tam.pv (1,3), tam.wle (1), tam_irf_3pl (1), tam_NA_pattern (1,2), tam.linking (1,2)

FEATURES:

BUG FIXES:

NOTES:

- included argument 'verbose' in TAM functions which is frequently used in many other R functions
- exported utility functions 'tam_args_CALL_search', 'tam_matrix2', 'tam_outer', 'tam_normalize_matrix_rows', 'tam_trim_increment', 'tam_difference_quotient', 'tam_assign_list_elements', 'tam_aggregate'
- fixed bug in 'IRT.informationCurve' when only one item is used (thanks to Asghar Minaei)
- changed the default of the argument 'est.variance' to TRUE in 'tam.mml' and 'tam.mml.mfr'
- included optional control argument 'trim_increment' in argument list 'control' in 'tam_foo' functions which allows the user to choose between two methods 'half' and 'cut' for reducing the size of increments
- internal restructure of the main functions

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

FEATURES:

- extended multiple group models to multidimensional IRT models in 'tam.mml.3pl' (see Example 1.9)

BUG FIXES:

- fixed a bug in 'tam.mml.3pl' with option skillspace='normal' and estimated means (i.e., means not fixed to zero)
- fixed a bug in handling multiple groups in 'tam.mml.3pl' (thanks to Laura Zieger)

NOTES:

- changed calculation of item response probabilities to avoid numerical overflow
- removed dependencies to the packages msm and tensor
- included example for extracting posterior probabilities in Rasch mixture models (requested by Ling Liaw)
- changed order of items in output of value 'item' in 'tam.mml.mfr'
- exported utility functions 'tam_packageinfo', 'tam_print_call' and 'tam_rsessinfo' for usage in other packages
- added 'logLik' method for objects of class 'tam.jml'
- moved function 'tam.jml2' into 'tam.jml' (now available as the default with argument 'version=2')

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

- data.geiser (2), tam.mml.3pl (1.9), tam.jml (1,2,3)

FEATURES:

BUG FIXES:

- removed adjustment for WLE estimation of extreme scores in 'adj' in 'tam.wle'. Now, WLE estimation in TAM should be equivalent to ConQuest software, sirt::IRT.mle and mirt::fscores. The adjustment 'adj' is only applicable for MLE estimation in 'tam.wle'. The WLE estimation of the recent TAM version is equivalent to 'adj=0' of previous TAM versions.

NOTES:

- restructured Rcpp functions

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

FEATURES:

BUG FIXES:

- changed handling of item names in 'tam.mml.mfr' back to TAM 1.995-0 version (problem communicated by Barth Riley)
- fixed a bug in 'tam.mml.2pl' for irtmodel='GPCM.design' with provided design matrices for item slopes (thanks to Matthias Trendtel)

NOTES:

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

FEATURES:

BUG FIXES:

- fixed incorrect handling of item renaming in 'tam.mml.mfr' (thanks to Barth Riley)
- fixed a bug in calculation of p values in 'tam.fit' (thanks to David Torres Irribarra)
- fixed a bug in 'IRT.simulate' with used argument 'nobs' (thanks to Mark Dulhunty)
- the fixing of the random seed within the 'tam.fit' function is removed (thanks to Sebastian Weirich)
- fixed crucial bug in 3PL estimation in 'tam.mml.3pl' function (thanks to Petra Kleiber)

NOTES:

- stop estimation in 'tam.jml' (and 'tam.jml2') if there are persons with only missing item responses (communicated by Ricardo Primi)
- included standard errors for constrained parameters in 'tam.mml.mfr' similar to the software ConQuest software (requested by Florian Keller und Domenico Angelone)

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

FEATURES:

BUG FIXES:

- fixed a bug in 'tam.fit' when proving a fit matrix F with missing labels (thanks to Maxim Morin)

NOTES:

- included argument 'version' in 'msq.itemfit'. Another Rcpp implementation (version=2) of the fit calculation has been added which avoids the storage of memory intensive objects (previous function; version=1). Additionally, a pure R function of calculating fit can be requested using version=3 (thanks to Frank Musekamp).

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

FEATURES:

BUG FIXES:

- fixed a bug in 'tam.wle' which occured if some persons were eliminated from the analysis
- fixed a recently induced bug in 'tam.wle' estimation of multidimensional models (thanks to Sascha Bernholt)

NOTES:

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

FEATURES:

- included function 'IRTLikelihood.ctt' which computes the individual likelihood under classical test theory and normally distributed measurement errors

BUG FIXES:

- fixed a bug in plausible value imputation 'tam.pv' in case of 'normal.approx=FALSE' (thanks to Matthias Trendtel)
- fixed a bug in plotting 'png' files in 'plot.tam' (thanks to Rainer Alexandrowicz)
- fixed a bug in calculating standard errors with only one estimated xsi parameter
- fixed a crash in 'tam.jml2'
- fixed a bug in 'tam.wle' when using the argument 'score.resp'

NOTES:

- included Example 24 for specifying sum constraints in partial credit models (requested by Chen-Wei Liu)
- output of 'tam.wle' is now an object of class 'tam.wle'. A print method for this class is defined.
- WLE reliability in 'tam.wle' is now calculated by taken sample weights 'pweights' into account
- included Example 11 of 'tam.mml.3pl' with mixed item response format (thanks to a correspondence with Andreas Seifert)
- included utility functions for weighted statistics 'weighted_Stats'
- included utility functions 'WLErel' and 'EAPrel' for computation of WLE/EAP reliability
- added 'print' methods for tam classes ('tam.mml', 'tamaan', ...)
- added S3 methods 'Scale' and 'DescribeBy'. The application of these methods to object classes will be included in future TAM versions.
- slight change in summary output of 'IRT.modelfit'

DATASETS INCLUDED OR MODIFIED:

EXAMPLES INCLUDED OR MODIFIED:

- tam.mml (24), WLErel (1), tam.mml.3pl (11), weighted_Stats (1,2,3), IRTLikelihood.ctt (1), tam.pv (2)

Features:

Bug fixes:

- fixed a crucial bug in estimation of 'tam.mml.2pl', irtmodel='GPCM.design'

Notes:

- included a warning message in 'IRT.truescore' when not all values in the theta grid can be computed
- cleaned namespace file and defined imported functions in a more explicit way
- included argument 'Q' in 'IRT.truescore' function which allows for computation of weighted true scores (thanks to David Sichau)
- fixed a bug in 'tam.mml.2pl' with irtmodel='GPCM.design' in case of blank categories

Datasets included or modified:

Examples included or modified:

- IRT.truescore (2)

Features:

Bug fixes:

- fixed a bug in postprocessing of parameter names in 'tam.mml.mfr' with the option "constraint='items'"
- avoided NA entries in posterior calculation in 'tam.mml' and 'tam.mml.2pl'

Notes:

- corrected output of likelihood when minimum value was not obtained in last iteration
- changed behavior of parameters 'max.increment' and 'increment.factor' in 'tam.mml.mfr' which should result in more stable deviance behavior
- included attributes 'pid' and 'pweights' in 'IRT.likelihood' and 'IRT.posterior'
- included optional argument 'gamma.init' for 'tam.mml.2pl' and irtmodel='GPCM.design'
- included argument 'item.elim' in 'tam.mml', 'tam.mml.2pl' and 'tam.mml.3pl' which prevents the omission of items with a maximum score of zero (suggested by Anja Mikolajetz)
- changed output of 'pweights' to original provided person weights, not the transformed person weights

Datasets included or modified:

Examples included or modified:

Features:

Bug fixes:

- fixed a bug in creating design matrices with more than 9 levels of a factor in a facet for facet models with items with varying number of categories
- fixed a recently introduced bug in 'tam.pv' which occured due to a change in the list of arguments in auxiliary function 'stud_prior.v2' (thanks to Florian Gigl)

Notes:

Datasets included or modified:

Examples included or modified:

Features:

Bug fixes:

- fixed a recently introduced bug in TAM 1.12-0 when creating the item parameter table

Notes:

- extended value 'item' to output of category 0 if design matrices put values on zero categories

Datasets included or modified:

Examples included or modified:

- tam.mml (23)

Features:

Bug fixes:

- fixed a recently included bug (in TAM 1.11-0) in 'tam.mml.3pl' estimation when guessing parameters are estimated (thanks to Rolf Strietholt and Laura Zieger)

Notes:

Datasets included or modified:

Examples included or modified:

Features:

Bug fixes:

- fixed a recently included bug in TAM 1.10-0 in calculation of deviance in 'tam.mml.2pl', 'tam.mml.3pl' and 'tam.mml.3pl' (thanks to Christopher Gess)
- fixed a recently included bug in TAM 1.9-0 in estimation of models with estimated means of latent trait (thanks to Frank Musekamp)

Notes:

- included guessing parameters and its standard errors in output of 'tam.mml.3pl' ('guess' and 'se.guess')

Datasets included or modified:

Examples included or modified:

Features:

- added simulation function 'IRT.simulate'
- added function 'IRT.truescore' which converts theta scores into true scores

Bug fixes:

- fixed a bug in calculation of expected counts

Notes:

- included argument 'dev_crit' in 'control' in tam functions which allows a decision whether change in absolute deviance difference (dev_crit="absolute") or in relative deviance differences (dev_crit="relative") should be assessed.
- extended 'IRT.itemfit' S3 method to objects of classes generated in TAM (suggested by Karoline Sachse)
- included item-wise calculation of thresholds according to Ali, Chang and Anderson (2015) suggestion in 'IRT.threshold'

Datasets included or modified:

Examples included or modified:

- IRT.itemfit.tam (1,2), IRT.threshold (3), IRT.simulate (1,2,3,4), IRT.truescore (1)

Features:

- included accelerators of the EM algorithm according to Ramsay (1975) and Yu (2012). In the recent TAM package version, no acceleration is used as an default.

Bug fixes:

- Fixed recently broken 'tam.mml.3pl' function when the design matrix E is used.

Notes:

- small speedup for large samples of students in unidimensional models

Datasets included or modified:

Examples included or modified:

- tam.mml (21,22)

Features:

- added S3 method 'IRT.informationCurves' for computing item and test information curves (suggested by Ray Opao)

Bug fixes:

- fixed a crucial bug in 'tam.mml.3pl' for mixed dichotomous and polytomous data occur and guessing parameters are estimated (thanks to Matthias Trendtel)

Notes:

- added the argument 'Q.fixed' in 'tam.mml.3pl' for usage of fixed item loadings
- small speedup in M-step for datasets with large number of items

Datasets included or modified:

Examples included or modified:

- tam.mml (15c), tam.mml.3pl (10), IRT.informationCurves (1,2)

Features:

- included S3 method 'IRT.residuals' (and 'residuals') for objects of classes 'tam.mml', 'tam.mml.2pl' and 'tam.mml.mfr'
- added function 'msq.itemfitWLE' (see ?msq.itemfit) which computes outfit and infit statistics based on weighted likelihood estimates (WLEs)
- added function 'tam.Q3' for calculation of Q3 statistics based on WLEs

Bug fixes:

- fixed a bug in 'tam.mml.3pl' (which led to an error) in case of multiple groups
- fixed a bug in 'tam.mml.mfr' when no person identifiers were provided (thanks to by Michal Modzelewski)
- fixed a bug in 'msq.itemfit' with differing number of categories per item
- fixed a bug in calculation of expected counts for multifaceted models
- fixed a bug in 'tam.mml.mfr' with non-unique combinations of person and facet identifiers

Notes:

- speeded 'tam.mml.3pl' for models with many item parameters
- added argument 'change' in 'plotDevianceTAM' for plotting iteration progress in deviance values
- added arguments 'progress' and 'output.prob' in 'tam.wle'
- added summary of fit statistics in output of 'msq.itemfit' and 'summary.msq.itemfit'
- included sample size of each item pair in 'tam.modelfit' output value 'stat.itempair'
- renamed datset 'sim.rasch' into 'data.sim.rasch'. The same procedure was conducted for all other 'sim.xxx' datasets.

Datasets included or modified:

Examples included or modified:

- IRT.residuals (1), msq.itemfit (1,2,3), tam.modelfit (4)

Features:

- included function 'IRT.drawPV' for drawing plausible values based on individual posterior (output of 'IRT.posterior')

Bug fixes:

- fixed a bug for input of a B loading matrix with no item labels
- fixed a bug in 'plot.tam' with missing data in item responses (bug fix provided by Michal Modzelewski)
- corrected an included bug in recent TAM versions in AIC calculation (thanks to Simon Grund)
- fixed a bug in variance fixing in 'tam.fa' which occured in a recent TAM version due to changes in defaults in 'tam.mml.2pl'
- fixed a bug when constraining of item parameters was ignored in 'tamaan' (thanks to Stefan Behrendt)

Notes:

- speeded data/design matrix preparation phase in 'tam.mml.mfr'
- included internal function 'tamcat' for analyzing computation time
- changed numerical behavior in 'tam.mml.mfr' in order to avoid a zero sum of all individual posterior probabilities (thanks to Roman Freunberger)
- included argument 'xsi.setnull' in 'tam.mml.mfr' for convenient fixing of xsi parameter groups. A warning is now given if the combination of person and facet identifiers is not unique.
- allowed also explicit use of facets without effects in 'tam.mml.mfr'
on 'formulaA', e.g. 'item
*step + 0*rater' sets all rater effects to zero - included warning in 'tam.wle' if the WLE is (possibly) not estimable (suggested by Rainer Alexandrowicz)

Datasets included or modified:

Examples included or modified:

- tam.mml (19, 20, 10a), IRT.drawPV (1), tamaan (1.9)

Features:

- included S3 methods 'IRT.expectedCounts', 'IRT.modelfit', 'predict' and 'IRT.data' for fitted models in TAM
- included item fit statistic on item level 'msq.itemfit' as an alternative to the simulation based 'tam.fit' function
- included linear approximation for confirmatory factor analysis

'IRT.linearCFA'

Bug fixes:

- fixed a bug in plausible value imputation in multidimensional models and normal approximation (reported by Dan Su)
- changed resp in output of 'tam.mml', 'tam.mml.2pl' and 'tam.mml.3pl' to output the original used data with missings (thanks to Matthias Trendtel)
- extended tam.modelfit to handle multi matrix design appropriately
- changed defaults in 'est.variance' if 'variance.fixed' is provided in 'tam.mml.2pl'. The previous default led to bugs in the tamaan function. (reported by Christoph Duchhardt)

Notes:

- added output values 'xsi.fixed.estimated', 'B.fixed.estimated' for default item parameter fixings in tam.mml, tam.mml.2pl, tam.mml.mfr and tam.mml.3pl (suggested by Matthias Trendtel)
- extended tan.modelfit function for more flexible input to allow for assessing model fit for objects which were not fitted in TAM ('tam.modelfit.IRT', 'tam.modelfit.args')
- added 'IRT.likelihood' and 'IRT.posterior' S3 method for object of class 'tam.latreg'
- extended 'plotctt' and 'plot.tam' the using graphics package in addition to lattice (suggested by Michal Modzelewski)
- included function for computing Thurstonian thresholds ('IRT.threshold')

Datasets included or modified:

Examples included or modified:

- tam.mml (1,5t,6), IRT.expectedCounts.tam (1), tam.modelfit (1), predict (1), msq.itemfit (1), IRT.linearCFA (1,2), IRT.threshold (1,2)

Features:

- included function 'cfa.extract.itempars' for extracting item parameters from a fitted confirmatory factor analysis with the cfa function in lavaan package
- included the function 'IRTLikelihood.cfa' for extracting the individual likelihood from fitted confirmatory factor analyses
- included an S3 method 'IRT.WrightMap' which is a wrapper to the WrightMap package

Bug fixes:

- fixed a bug for item response dataset with only one column
- fixed a bug in tamaan function when applied to TYPE="TRAIT" and models without guessing parameters
- fixed a bug in plot.ctt function for input of item response data with only one columns (thanks to Michal Modzelewski)

Notes:

- added item types 'PCM(kk)' and 'GPCM(kk)' for smoothing item intercepts in tamaan() function
- included a summary method for 'tam.fit'
- included 'MEASERR' convenience method for single indicators with measurement errors in 'lavaanify.IRT'
- changed some internal handling in 'lavaanify.IRT' due to announced changes in lavaan 0.5-18 in 'lavaanify'
- changed equality constraints in the section "MODEL CONSTRAINT" in 'tamaanify' to the usage of "==" instead of "=" (as in lavaan)

Datasets included or modified:

- data.ctest2

Examples included or modified:

- cfa.extract.itempars (1), IRTLikelihood.cfa (1), tamaan (2), tam.fit (1,2), data.geiser (2), tam.latreg (2,3), IRT.WrightMap (1,2,3,4), data.cqc (1,2,3), lavaanify.IRT (6)

Features:

- included option for sum constraints on item intercepts (constraint="items") in 'tam.mml'. For polytomous data, this option automatically specifies irtmodel="PCM2".
- included truncated normal prior distributions for gammaslope parameters in 'tam.mml.3pl'
- included 'tamaan' and 'tamaanify' functions for more more convenient model specifications in TAM
- included 'doparse' function as a shortcut for writing loops in model syntax
- included item-category specific curves in 'plot.tam' using

the argument type="items" (suggested by Cheng-Hsuan Li and Michal Modzelewski) - plot function for type="items" now also works for multidimensional item response models with between item dimensionality
- included normal approximation and sampling of regression coefficients for plausible value imputation ('tam.pv') for multidimensional item response models. The theta grid can also be fixed (argument theta.model=TRUE) and is taken from the fitted model.
- included the function 'tampv2datalist' for converting plausible values and a set of covariates into a list of multiply imputed datasets
- included the function 'tam.latreg' which performs a latent regression model in which only the individual likelihood is needed as an input

Bug fixes:

- corrected algorithmic stopping criteria in 'tam.mml.3pl'
- fixed a bug in 'tam.mml.2pl' when design matrix B.fixed for fixed loadings contains only one row
- another bug was fixed by Michal Modzelewski in data preparation in the multi-facet function 'tam.mml.mfr'
- fixed a bug in estimation routines if a dataset resp is provided without column names (thanks to Alec Nyce)

Notes:

- included parametrization of mean item intercept (irtmodel="PCM2") in output of tam.mml and tam.mml.2pl (value item$xsi.item)
- included nonlinear terms in lavaan syntax in 'lavaanify.IRT'
- included option for initial delta parameters in 'tam.mml.3pl'
- added AIC3 information criterion in summary
- included datasets data.geiser and data.janssen for latent class analysis and mixture Rasch analysis (thanks to Christian Geiser and Anne Janssen)
- included Example 3 in 'tam.wle' comparing WLE person parameter estimates from the TAM and the PP package
- included Example 3 in 'tam.threshold' for creating Wright maps with the WrightMap package

Datasets included or modified:

- data.geiser, data.janssen, data.janssen2

Examples included or modified:

- tam.modelfit (3), tam.mml (1,4,5,6,13,18), lavaanify.IRT (3,4), tam.mml.3pl (1.6b,4,7,8,9), tamaanify (1,2), tamaan (1,2), plot.tam (1,2), tam.wle (3), tam.threshold (3), tam.pv (2,3), tam.latreg (1)

Features:

- included fit of constrained facet parameters for output of 'tam.mml.mfr'
- included extended lavaan syntax 'lavaanify.IRT'. This function was previously included in sirt package as 'lavaanify.sirt'.
- added argument 'gammaslope.center' in 'tam.mml.3pl' for centering item parameters within iterations in estimation
- included an argument of fixing person parameters in 'tam.jml' (see tam.jml, Example 1)
- added S3 methods for fitted models in TAM: 'logLik', 'IRT.likelihood', 'IRT.posterior', 'IRT.irfprob', 'IRT.factor.scores'. These S3 methods were defined in the recent version of the CDM (>= 4.0) package.

Bug fixes:

- improved model identification in 'tam.mml.mfr' in case of differing number of categories per item and fixed a bug in creating the A matrix (thanks to Michal Modzelewski).
- fixed a miscount of number of estimated parameters in multiple group models (thanks to Michal Modzelewski).
- fixed incorrect handling of irtmodel="PCM2" in 'tam.mml' for multidimensional models (thanks to Chen-Wei Liu)
- fixed a bug in 'tam.jml2' when some item parameters were fixed with the xsi.fixed argument
- fixed a bug in fit calculation for tam.jml objects (thanks to Hillary Ferral)

Notes:

- included Rcpp based version of item fit calculation to reduce computational time
- changed calculation of tam.fit in case of missing item responses which substantially changes fit statistics especially for faceted models.
- changed handling of simulating abilities in 'tam.fit'
- included p values in output for fit statistics in 'tam.fit'
- included estimation of Thurstonian thresholds for multidimensional models
- added an example of a 2PL mixture distribution IRT model with two latent classes (tam.mml.3pl, Simulated Example 8)
- changed summary layout in 'tam.mml.3pl'
- made item parameter estimation in 'tam.jml' more computationally efficient
- added anova method for 'tam.mml.3pl'
- added value of the individual likelihood in 'tam.mml', 'tam.mml.2pl' and 'tam.mml.3pl'

Datasets included or modified:

- rating dataset data.exJ03 (provided by Matt Barney)

Examples included or modified:

- tam.fit (4), tam.threshold (2), tam.mml.3pl (7,8), lavaanify.IRT (1,2), tam.jml (1,4), IRT.likelihood.tam (1), IRT.factor.scores (1), IRT.irfprob (1)

Features:

- added user defined variance function in tam.mml and tam.mml.2pl
- added user defined constraints on slope parameters in 'tam.mml.3pl'

Bug fixes:

- fixed a bug in deviance computation in 'tam.mml.3pl' when stochastic nodes are used
- (hopefully) fixed computational problems in multidimensional models with nearly singular covariance matrix (thanks to a discussion with Johannes Hartig)

Notes:

- added a plot function 'plot.ctt' accompanying the classical test theory based statistic
- fixed a bug in creation of design matrices in faceted models with differing categories per item (thanks to Michal Modzelewski). Note that the treatment of faceted models is still not compatible with other software (like ConQuest). A substantial bug fix is planned in future TAM relaeses.

Datasets included or modified:

- data.ctest1 (thanks to HGP)

Examples included or modified:

- tam.fit (4), tam.mml (17), tam.ctt (1)

Bug fixes:

- modified incorrectly calculated thresholds with 'tam.threshold' for multidimensional models (reported by Janine Buchholz and Joerg-Henrik Heine)

Notes:

- included SRMSR fit statistic in tam.modelfit
- included argument "Nsimul" in "tam.fit" which improves calculation of fit statistics

Features:

- included model fit statistics (Q3, SRMR, MADRESIDCOV): 'tam.modelfit'
- included 'tam.mml.3pl' for 3PL estimation, general design matrices for item slopes and options for discrete latent variables (estimation of latent class models, located latent class model, skew distributions, mixed IRT models, cognitive diagnosis models, ... is now possible)
- extended tam.pv for objects of class 'tam.mml.3pl'
- speeded data preparation of faceted models a bit
- tam.se now computes standard errors for estimated slope parameters, B (for irtmodel in c("2PL", "GPCM", "2PL.groups")); highly experimental for irtmodel="GPCM.design").
- plot.tam now prints a message about the location of the exported graphics.
- plot.tam has a new argument 'items' for selecting the items to be displayed. RD-file adjusted.
- plot.tam has new arguments 'export.type="png"' and "export.args=list()" for selecting the file type for the graphics export. The supported file types are those supported by ?dev.new. The argument 'export.args=list()' is for specifying a list of arguments for the export device method.
- tam.threshold has a new argument prob.lvl to set the probability level of the thresholds.

Bug fixes:

- fixed a bug in pid output in object 'person' for multidimensional models (thanks to Sebastian Weirich)
- fixed a bug in variance estimation for multiple group 2PL models
- included 'designMatrices.mfr2' for a provisional bug fix for multifaceted models with differing number of categories per item (see Example 16 in tam.mml; thanks to Christoph Duchhardt)
- fixed a bug in 'tam.mml.mfr' for differing number of categories per item -> now a xsi vector with parameter fixings to values of 99 is used
- plot.tam is debugged regarding the case of missing WLEs for some person (thanks to Florian Buchwald).
- plot.tam now doesn't throw a warning when the 'Plots' directory already exists.
- fixed a bug in plot.tam for cross-platform accordance (thanks to Uli Keller).
- fixed a bug in tam.fit with missing item scores
- fixed a bug in tam.jml so that ../Tutorials/6Population.htm now runs without error.

Notes:

- tam.fit is experimental for objects obtained from tam.mml.mfr.
- changed convergence behavior in tam.wle (thanks to Peter Harych)
- changed item sorting in output object 'item'
- tam.mml(/.2pl/.mfr) methods now return Q, tam.mml.2pl additionaly returns B.fixed, basispar, E, and est.slopegroups objects. RD-file adjusted.
- tam.jml(/.jml2) do not return fit statistics anymore. Item and person fit can be obtained by a call to tam.fit (/.jml.fit).
- in case of multidimensional analyses, we recommend an update of the R-package mvtnorm to version 0.9-99992 or higher.
- extended example 13 in tam.mml.Rd to the case of polytomous longitudinal data (requested by and thanks to Florian Buchwald)
- extended example 6 in tam.mml.Rd to the specification of sum constraints on the item parameters in a PCM model for differing number of categories.
- Some minor changes in RD-files (eg., irtmodel for argument 'E' in tam.mml.2pl is "GPCM.design", theta value in tam.jml is "MLE in final step").

- fixed a Bug in the optional specification of a B matrix in tam.mml.mfr
- new faster tam.mml.wle2; included Rcpp helper functions for tam.mml.wle2 for fast computation; tam.mml.wle2 new reference in tam.wle for objects of class tam.mml, rd-file adapted
- tam.mml.mfr now accepts polytomous data with more than ten categories per item
- specification of response matrix in 'long format' fixed, see example section (thanks to Takuya Yanagida)
- some minor changes in usability (eg., pid can be specified as vector, one-column matrix or data.frame)

- updated DESCRIPTION and NAMESPACE due to changes in Rcpp
- change in package maintenance: Thomas Kiefer is now the package maintainer
- fixed a bug in item labeling in tam.se (thanks to Sebastian Weirich)
- fixed a bug in labeling of reduced reponse matrices in tam.mml.mfr
- fixed some examples in .Rd files

- added 'AXsi_' as value in tam.mml, tam.mml.2pl and tam.mml.mfr
- cleaned labels in 'tam.mml.mfr' for faceted models
- included the argument 'score.resp' in tam.wle for scoring a new dataset different than the one which is used for item calibration in tam (requested by Jan Boehnke)
- added examples for applying a sum constraint of item parameters in 'tam.mml' (request by Kelvin Gregory)
- included argument 'adj' in tam.jml/tam.jml2 for modifying the adjustment in case of perfect scores (thanks to Courtney Castle)
- included dataset 'data.timssAusTwn.scored'
- included summary for tam.jml
- use plausible values for calculation in item fit (in tam.mml.fit)
- included an example for creation of user-defined fit statistics

- removed blanks in item category labels in tam.ctt2 and tam.ctt3
- allowed non-integer group labels
- fixed a bug in data preparation for facet models which appeared since a change in TAM 0.7 (thanks to Jonathan Williams)

- changed rating scale parametrization for unequal number categories per item
- included a warning in tam.ctt2 if memory allocation is reached
- included deviance plot 'plotDevianceTAM' (thanks to Sebastian Weirich)
- included tam.ctt3

- include crossprod in some functions
- fixed a bug for using latent regression models in facet model function 'tam.mml.mfr' (thanks to Jonathan Williams)
- included another facet example (Example 14)
- fixed a bug in tam.mml.mfr if all facets possess the same number of levels
- fixed a bug in multidimensional generalized partial credit model estimation and speeded 2PL estimation a bit
- included bifactor model and exploratory factor analysis -> function 'tam.fa' with arguments irtmodel='efa','bifactor'
- now printing progress in tam.ctt2 correctly
- fixed a bug in data processing in tam.mml.mfr for factor or string input of facets
- included more examples in tam.fit, anova.tam.mml, tam.pv, tam.jml, tam.se, tam.wle
- avoided inversion of singular matrices in multidimensional WLE estimation tam.wle (thanks to Iasonas Lamprianou)
- removed bugs for item response datasets with only one item in tam.mml and tam.mml.2pl
- fixed a bug in tam.ctt2: multiple groups were not displayed
- included an option for deleting redundant generalized items which sometimes improves computation speed
- cleaned the TAM manual (thanks to Michaela Mayr)
- changed output structure of tam.fa and included rotation examples in tam.fa.Rd
- included additional convergence stabilizing approach which can be activated using the argument 'fac.oldxsi' in control -> See Example 15 for such a non-converging model where changing algorithmic defaults seems necessary (problem addressed thanks to Sebastian Weirich)
- substituted internal 'theta.sq' function by Rcpp based function 'theta.sq2'
- started with including examples similar to the ones in the ConQuest 2.0 manual (see data.cqc.Rd)

- fixed a bug when in latent regression when providing a formula in argument 'formulaY'
- fixed a bug in output of tam.se (thanks to Sebastian Weirich)
- included a faster alternative to tam.ctt -> tam.ctt2 is based on a Rcpp function
- allowed freeing beta constraints by the user by including the possible argument beta.fixed=FALSE
- removed a bug in data processing for faceted designs (thanks to Michaela Mayr)
- fixed a bug with varying string lengths of facet levels
- fixed a bug in tam.ctt2 for numeric input matrices
- fixed incorrect handling of constraints in regression coefficients which was just included in a recent 0.6 version
- fixed an output bug in tam.wle in multidimensional models (thanks to Jan Henning-Kahmann)
- accelerated estimation of multdimensional models (different approach of calculating the prior distribution)
- fixed a bug when submitting xsi.inits and beta.inits to tam
- fixed a bug in facet models when a facet contains only one element (twice thanks to Michaela Mayr)
- included irtmodel='PCM2' (partial credit model with ConQuest parametrization) and irtmodel='RSM' (rating scale model)
- added a longitudinal data example in 'tam.mml' manual

- included data.ex11 example dataset
- removed a bug in WLE calculation for multifaceted designs
- avoided warning in cov2cor for tam.mml.2pl function (thanks to Linda Graefe)
- included argument increment.factor for obtaining convergence behavior in bad suited data constellations
- changed default of np.adj in tam.pv to 8
- included two helper functions written in Rcpp in plausible value imputation function tam.pv for better performance
- included S3 plot method
- included data.numeracy and data.timssAusTwn datasets

- included quasi-Monte Carlo integration as an alternative to stochastic integration
- multifaceted model completely implemented
- print warning if there are items in resp with only zeroes (thanks to Ronny Scherer)
- fixed a bug in tam.wle, tam.pv and tam.se
- fixed a bug in creating design matrices for faceted designs
- changed initial xsi values for indetermined parameters (thanks to Sebastian Weirich)
- included function for calculating Thurstonian thresholds
- included Rcpp function in M step for calculating item parameters
- fixed a bug for partial credit models
- included parametrization of the dispersion model for ordered data
- included a function tam.ctt for calculating some statistics based on classical test theory
- fixed a bug in tam.mml.2pl with irtmodel="GPCM" for models with differering numbers of item categories

- cleaned Rd manual files a bit
- included standard errors in tam.mml and tam.mml.2pl
- included summary.tam and summary.tam.mml
- included anova.tam for model comparisons
- fixed a bug in multiple group estimation during tam.mml and summary application
- fixed a bug in WLE estimation in the multdimensional case

- removed bugs in tam.pv for sampling plausible values with the argument normal.approx=FALSE (thanks to Sebastian Weirich)
- included seed for stochastic integration in the tam control argument
- fixed a bug in stochastic integration (thanks to Steffen Brandt)
- fixed a bug in the calculation of reliability and person parameter estimates when applying stochastic integration (thanks to Steffen Brandt and Caroline Just)
- use estimated parameters with minimal deviance in the output (thanks to Ulrich Schroeders and Steffen Brandt)

- 2PL calibration outsourced
- designMatrices and tam.mml.mfr method for multifacet rasch model calibration included. Both highly experimental.
- Examples for the calibration included, see help file of sim.mfr