craschR: Item and person estimates and fit statistics.
In amyarneson/crasch: Construct Mapping with Rasch

Item and person estimates and fit statistics.

craschR(scores, itemInfo = NULL, consInfo = NULL, varsInfo = NULL,
  estPackage = "TAM", retainOrig = FALSE, missingAs0 = FALSE,
  longFormat = FALSE, persMethod = "EAP", consecutive = FALSE,
  writeout = TRUE, fileSuffix = NULL)

`scores`	A data frame or path to CSV file in crasch format with the scoring information. The scores are be coded as integers in terms of the construct levels. Missing responses are coded as `NA`.
`itemInfo`	A data frame or path to CSV file in crasch format with information on each item. There must be `I` rows and as many columns as specified in the crasch format documentation. No factor variables are allowed.
`consInfo`	A data frame or path to CSV file containing labeling information about the construct(s) in crasch format. Contains 1 row per construct.
`varsInfo`	A data frame containing demographic variable information for respondents. One respondent per row, IDs must match those of `scores` data frame. Numeric, character, and factor variables are allowed.
`estPackage`	A string that identifies which IRT package will be used to conduct estimation of item (and possibly person) parameters. See details.
`retainOrig`	A logical indicating whether one element of the output object should include the original result object from `mirt` or `TAM`.
`missingAs0`	A logical indicating whether `NA`s should be recoded to the lowest level (available on a given item) of the construct. If `FALSE`, missing data will remain as missing data.
`longFormat`	A logical indicating whether the scores object is in long format. If `TRUE`, reshaping to wide format will occur internally.
`persMethod`	Choose from c("EAP", "MLE", "WLE")
`consecutive`	A logical indicating whether to perform a 'consecutive' analysis in which an independent analysis is conducted for each dimension. If `TRUE`, the correlation between any two dimensions will be constrained to be 0. If `FALSE`, multidimensional analysis that allows non-zero correlations among dimensions will be conducted.
`writeout`	A logical indicated whether the estimate objects should be written to your working directory as CSVs. See Return below for the names of each file produced. Each will also include any `fileSuffix` that you specify.
`fileSuffix`	A character string that will be affixed to the end of each file name (if `writeout = TRUE`). Use this if you are conducting multiple analyses in the same working directory and do not wish for your existing files to be overwritten.

A list with the following entries

itemPars

A matrix with the item parameter estimates. If estPackage = "mirt", this matrix will be... estPackage = "TAM", this matrix will be in the ConQuest parameterization with an overall item ‘delta’ and step ‘tau’s.

(itemGrid.csv)

itemSEs

A matrix with the standard error estimates corresponding to the estimates given in itemPars.

(itemGrid.csv)

itemThres

A matrix giving the Thurstonian threshold estimates.

(itemGrid.csv)

itemFit

A data frame providing the fit statistics for items and steps.

(itemFit.csv)

persPars

A data frame containing estimates for person locations for each dimension.

(persGrid.csv)

persSEs

A data frame with standard errors for person estimates.

(persGrid.csv)

persFit

If non-consecutive (including unidimensional) analysis is performed, this will be a data frame with person fit statistics. If ‘consecutive’ multidimensional analysis is performed (independent analysis of each dimension), this will be a list with fit statistics for each dimension. True multidimensional fit statistics are not available if TAM is used. See details for how these are produced.

(persFit.csv)

popDist

A list containing a vector of estimated population means for each dimension (often these are constrained to be 0) and a variance- covariance matrix. Variances appear on the diagonal, covariances elsewhere. ‘Consecutive’ analyses fix all covariances to 0.

(popDist.csv)

sepRel

A vector containing the separation reliabilities for each dimension.

(classicalStats.csv)

estSummary

A list containing information about the estimation settings and model fit. Use the names() function to see what is provided here.

(estSummary.csv)

classicalStats

A list containing information about each dimension. Notably, Cronbach's Alpha and person separation reliabilities are provided. Statistics are calculated by recoding all missings to 0 as well as by only including complete cases.

(classicalStats.csv)

empties

that were deemed possible, but no one scored into them.

`estResults`	If `retainOrig = TRUE`, this will contain the FULL output provided by the estimation package. See the documentation for that package for full details. Otherwise, this will be NULL.
`scoresOrig`
`scoresRecoded`	Wide format score data, recoded to nonnegative successive integers.
`itemInfo`
`consInfo`
`varsInfo`