This function takes a model of the democracy scores and extracts the discrimination parameters, score cutpoints, and standard errors for all the variables involved, putting these into a tidy data frame.
cutpoints(model, type = "score")
| model | A |
|---|---|
| type | A string specifying the cutpoint type. Can be (an abbreviation of) "score" (for score cutpoints) or "discrimination" (for discrimination parameters). Default is "score." |
A tibble with either score cutpoints for each variable used to
construct the latent scores in terms of the latent variable (the default), or
discrimination parameters for each variable used to construct the index.
For the score cutpoints (type = 'score'), the columns
estimate, pct975, and pct025 report the IRT
parametrization of the model estimates, a normalized measure in the same
scale as the latent variable.
# Replicate the official UDS 2011 release and calculate its cutpoints replication_2011_model <- democracy_model(dplyr::matches("pmm"), verbose = FALSE) cutpoints(replication_2011_model)#> # A tibble: 82 x 6 #> variable estimate pct025 pct975 se num_obs #> <chr> <dbl> <dbl> <dbl> <dbl> <int> #> 1 pmm_arat -1.43 -1.44 -1.42 0.00473 3873 #> 2 pmm_arat -1.02 -1.01 -1.02 0.00136 3873 #> 3 pmm_arat -0.427 -0.402 -0.448 0.0111 3873 #> 4 pmm_arat -0.0419 -0.000439 -0.0792 0.0190 3873 #> 5 pmm_arat 0.421 0.492 0.357 0.0326 3873 #> 6 pmm_arat 1.42 1.58 1.28 0.0719 3873 #> 7 pmm_blm -0.00248 0.0907 -0.0445 0.0214 275 #> 8 pmm_blm 0.474 1.04 0.220 0.130 275 #> 9 pmm_bollen -1.53 -1.55 -1.50 0.0109 510 #> 10 pmm_bollen -1.08 -1.08 -1.07 0.00183 510 #> # ... with 72 more rows