| Title: | CUSUM Person Fit Statistics |
| Version: | 1.0.0.0 |
| Description: | Person fit statistics based on Quality Control measures are provided for questionnaires and tests given a specified IRT model. Statistics based on Cumulative Sum (CUSUM) charts are provided. Options are given for banks with polytomous and dichotomous data. |
| Depends: | R (≥ 3.3.3) |
| License: | GPL-2 |
| Encoding: | UTF-8 |
| LazyData: | true |
| RoxygenNote: | 6.0.1 |
| Imports: | ltm, irtoys, stats, graphics |
| NeedsCompilation: | no |
| Packaged: | 2017-08-09 04:07:00 UTC; mhong |
| Author: | Maxwell Hong [aut, cre], Shao Can [ctb] |
| Maintainer: | Maxwell Hong <maxwell.hong@gmail.com> |
| Repository: | CRAN |
| Date/Publication: | 2017-08-09 04:18:09 UTC |
Generates CUSUM values for Rasch, 2PL and 3PL IRT model based on the Van Krimpen-Stoop & Meijer, (2002).
Description
Generates CUSUM values for Rasch, 2PL and 3PL IRT model based on the Van Krimpen-Stoop & Meijer, (2002).
Usage
cusum(dat, ipar = NULL, abi = NULL, IRTmodel = "2PL")
Arguments
dat |
a nxp matrix with n participants and p items. Responses are in 0 1 format. |
ipar |
a pxk matrix with given item parameters p items and k item parameters. ipar[,1] discrimination; ipar[,2] item difficulty; ipar[,3] guessing-parameter. |
abi |
a vector n ability. If not provided, estimated using Expected a Posteriori method. |
IRTmodel |
specify the IRT model ("1PL", "2PL", "3PL"). Default is "2PL" |
Value
Returns matrix with with lower and upper cusum statistics for dat.
References
Van Krimpen-Stoop, E. M., & Meijer, R. R. (2002). Detection of person misfit in computerized adaptive tests with polytomous items. Applied Psychological Measurement, 26(2), 164-180.
Examples
data(ex2PL)
cusum(dat = ex2PL)
Generates critical values for CUSUM statisitcs.
Description
cusum.cutoff Generates a bootstrap sample for cut-off scores.
Usage
cusum.cutoff(cusum.obj, upp = 0.975, low = 0.025, Breps = 1000)
Arguments
cusum.obj |
an object returned from cusum or cusum.poly |
upp |
user specified upper tail cut off. Default is .975 |
low |
user specified lower tail cut off. Default is .025 |
Breps |
number of bootstrap samples |
Value
Returns a matrix of lower and upper cut off values and corresponding standard deviations based on bootstrap sample.
Flags aberrant participants based on CUSUM statistics.
Description
Flags aberrant participants based on CUSUM statistics.
Usage
cusum.flag(cusum.obj, cutoff.obj, cut = NULL)
Arguments
cusum.obj |
an object returned from cusum or cusum.poly |
cutoff.obj |
an object returned from cusum.cutoff |
cut |
a vector for user specified cut offs (e.g c(1,1)). The first value is the upper limit. The second value is the lower limit. |
Value
Returns a true or false matrix whether a person is aberrantly responding.
Generates CUSUM plot for specified IDs.
Description
Generates CUSUM plot for specified IDs.
Usage
cusum.plot(cu.object, ID)
Arguments
cu.object |
an object returned from cusum or cusum.poly |
ID |
a numeric ID. |
Value
Returns a plot for specified cusum person chart.
Generates CUSUM values for polytomous IRT model based on Van Krimpen-Stoop & Meijer, (2002).
Description
Generates CUSUM values for polytomous IRT model based on Van Krimpen-Stoop & Meijer, (2002).
Usage
cusum.poly(dat, NCat, ipar = NULL, abi = NULL, IRTmodel = "GRM")
Arguments
dat |
a nxp matrix with n participants and p items. Responses are in 0 as the lowest scores format. |
NCat |
number of categories for each item. |
ipar |
a pxk matrix with given item parameters p items and k item parameters. Item difficulty under the "GRM" or item steps under "PCM" or "GPCM" are in the first columns. The last column is the discrimination parameter. |
abi |
a vector n ability |
IRTmodel |
specify the IRT model ("GRM","PCM","GPCM"). Default is "GRM". |
Value
Returns matrix with with lower and upper cusum statistics for dat.
References
Van Krimpen-Stoop, E. M., & Meijer, R. R. (2002). Detection of person misfit in computerized adaptive tests with polytomous items. Applied Psychological Measurement, 26(2), 164-180.
Examples
data(exGRM)
cusum.poly(dat = exGRM, NCat = 6)
Example data set based on a simulated 2PL model.
Description
Example data set based on a simulated 2PL model.
Usage
data(ex2PL)
Format
A data frame with 200 rows and 10 variables.
Source
Simulated data.
Example data set based on a simulated GRM model.
Description
Example data set based on a simulated GRM model.
Usage
data(exGRM)
Format
A data frame with 200 rows and 10 variables.
Source
Simulated data.
Example data set based on a simulated GRM model.
Description
Example data set based on a simulated GRM model.
Usage
gh
Format
Gaussian-Hermite Quadature points
Source
ltm