pseudoitems: Pseudo-Items

In ItemResponseTrees: IR-Tree Modeling in 'mirt', 'Mplus', or 'TAM'

Description

IR-tree models can be fit on the basis of so-called pseudo-items. To this end, the original, polytomous items are recoded into binary pseudo-items. Whether a pseudo-item is coded as 1, 0, or NA depends on the model equations (e.g., Böckenholt, 2012; Plieninger, 2020).

The ItemResponseTrees package internally works with pseudo-items as well. However, the user has to specify the model equations rather than the pseudo-items in the irtree_model syntax. Internally, the original responses are recoded on the basis of the model supplied by the user by the function irtree_recode(). This function may also be used directly if desired.

As an alternative to specifying the model equations themselves, users may also use the function irtree_create_template() with a mapping matrix (that specifies the structure of the pseudo-items) to generate the model equations automatically.

References

Böckenholt, U. (2012). Modeling multiple response processes in judgment and choice. Psychological Methods, 17(4), 665–678. https://doi.org/10.1037/a0028111

Plieninger, H. (2020). Developing and applying IR-tree models: Guidelines, caveats, and an extension to multiple groups. Organizational Research Methods. Advance online publication. https://doi.org/10.1177/1094428120911096

Examples

# Mapping matrix for data with three response categories:
(mm <- cbind(cat = 0:2,
             p1 = c(0,  1, 1),
             p2 = c(NA, 0, 1)))
#>      cat p1 p2
#> [1,]   0  0 NA
#> [2,]   1  1  0
#> [3,]   2  1  1

irtree_create_template(data.frame(x1 = 0:2, x2 = 0:2),
                       mapping_matrix = mm)
#>
#> m1 <- "
#> Equations:
#> 0 = (1-p1)
#> 1 = p1*(1-p2)
#> 2 = p1*p2
#>
#> IRT:
#> p1 BY x1@1, x2@1;
#> p2 BY x1@1, x2@1;
#>
#> Class:
#> Tree
#> "
#>

ItemResponseTrees documentation built on July 2, 2020, 2:25 a.m.