This function checks the assumptions of IRT. For unidimensionality the DETECT, RATIO, and ASSI statistics are calculated using sirt::conf.detect(). For local dependence, the standardized LD statistic is calculated using mirt::residuals-method() and the correlations are created using ltm::rcor.test(). Relative fit calculates BIC, AIC, and several others using code from mirt::anova-method(). Absolute fit is calculated only for models run with the mirt engine, using mirt::M2(). See the documentation on those functions for more information on calculation.

irt_assume(wizirt_fit, verbose = FALSE)

Arguments

wizirt_fit

An object from fit_wizirt

verbose

Logical. Should the names of the other packages used in this function be printed? Default is FALSE.

Value

A list with tibbles of data in them.

Examples

data("responses") my_model <- wizirt(data = responses[, -1]) assumptions <- irt_assume(my_model) print(assumptions, type = "rel")
#> # A tibble: 8 x 2 #> stat values #> <chr> <dbl> #> 1 log_lik -1026. #> 2 N 75 #> 3 n_pars 25 #> 4 AIC 2101. #> 5 AICc 2128. #> 6 BIC 2159. #> 7 SABIC 2080. #> 8 HQ 2124.
print(assumptions, type = "abs")
#> M2 df p RMSEA RMSEA_5 RMSEA_95 SRMSR TLI #> stats 303.2375 300 0.4368765 0.01207609 0 0.04630653 0.1190909 0.9524023 #> CFI #> stats 0.9524023
print(assumptions, type = "unid")
#> # A tibble: 5 x 3 #> Statistics Value Conclusion #> <chr> <dbl> <chr> #> 1 DETECT -0.62 Essential Unidimensionality #> 2 ASSI -0.290 Essential Unidimensionality #> 3 RATIO -0.36 Essential Unidimensionality #> 4 MADCOV100 1.69 NA #> 5 MCOV100 -0.62 NA
print(assumptions, type = "ld")
#> # A tibble: 300 x 5 #> item_1 item_2 LD_std pvals ccov #> <chr> <chr> <dbl> <dbl> <dbl> #> 1 A W 0.405 0.000312 0.0241 #> 2 J P 0.399 0.000398 0.0578 #> 3 K W 0.363 0.00139 0.0286 #> 4 A S 0.331 0.00371 0.0512 #> 5 A K 0.315 0.00592 0.0255 #> 6 O T 0.309 0.00695 0.0338 #> 7 D I 0.303 0.00816 0.0382 #> 8 H W 0.295 0.0101 0.0128 #> 9 M Q 0.285 0.0132 0.0456 #> 10 H P 0.282 0.0142 0.0243 #> # ... with 290 more rows