MIRTutils-package: MIRT utility functions
In woshikaqia/MIRTutils: MIRT Utility Package
MIRTutils-package R Documentation
MIRT utility functions
Description
A set of functions to calculate some commonly used values when doing IRT analysis
especially with new item types such as item bundles (testlets).
IRT model currently supported are 1-3 PL, GPCM, Rasch Testlet Model, and a mix of these models.
Values that can be computed include the probability of correct response, item information,
item score residuals, person scores, person fit index, etc.
It also provides a function for multidimensional IRT data simulation.
Details
In the context of the current package, test items can be either standalone items (SA) or cluster items (Cluster).
Both SA and cluster items consist of assertions, where an assertion is the same as a traditional item that was
typically modeled by a unidimensional IRT model. An SA item typically includes 1 to a handful of assertions,
whereas a cluster item typically includes 6 or more assertions. An SA item is more like a small set
of traditional items assuming local dependence among assertions, whereas a cluster is essentially a testlet.
The IRT model used is a special model similar to a Rasch Testlet model but allows for SA items to load only on the
overall dimension. When there are only standalone items, the model reduces to a regular unidimensional IRT model, and
all the values computed in this package fits to the traditional unidimensional IRT paradigm.
When there are only cluster items, the model reduces to a Rasch testlet model.
Unlike the unidimensional model where theta is usually estimated
by the univariate MLE, EAP or MAP, and the multidimensional model where theta is usually estimated by the multivarite
MLE or EAP, most of the values computed by this package involving testlets is based on the marginal maximum
likelihood estimation (MMLE) of theta. MMLE is a hybrid of MLE and EAP, where the nuisance dimension of a testlet (cluster)
is first integrated out, and the resulting marginal likelihood is then maximized to find the theta estimate.
Author(s)
Zhongtian Lin lzt713@gmail.com
See Also
Useful links:
woshikaqia/MIRTutils documentation built on Aug. 21, 2024, 4:30 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
| MIRTutils-package | R Documentation |
MIRT utility functions
Description
A set of functions to calculate some commonly used values when doing IRT analysis especially with new item types such as item bundles (testlets). IRT model currently supported are 1-3 PL, GPCM, Rasch Testlet Model, and a mix of these models. Values that can be computed include the probability of correct response, item information, item score residuals, person scores, person fit index, etc. It also provides a function for multidimensional IRT data simulation.
Details
In the context of the current package, test items can be either standalone items (SA) or cluster items (Cluster). Both SA and cluster items consist of assertions, where an assertion is the same as a traditional item that was typically modeled by a unidimensional IRT model. An SA item typically includes 1 to a handful of assertions, whereas a cluster item typically includes 6 or more assertions. An SA item is more like a small set of traditional items assuming local dependence among assertions, whereas a cluster is essentially a testlet.
The IRT model used is a special model similar to a Rasch Testlet model but allows for SA items to load only on the overall dimension. When there are only standalone items, the model reduces to a regular unidimensional IRT model, and all the values computed in this package fits to the traditional unidimensional IRT paradigm. When there are only cluster items, the model reduces to a Rasch testlet model. Unlike the unidimensional model where theta is usually estimated by the univariate MLE, EAP or MAP, and the multidimensional model where theta is usually estimated by the multivarite MLE or EAP, most of the values computed by this package involving testlets is based on the marginal maximum likelihood estimation (MMLE) of theta. MMLE is a hybrid of MLE and EAP, where the nuisance dimension of a testlet (cluster) is first integrated out, and the resulting marginal likelihood is then maximized to find the theta estimate.
Author(s)
Zhongtian Lin lzt713@gmail.com
See Also
Useful links:
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.