Skip to main content

Table 1 Comparisons of unidimensional and 4-dimensional Rasch models and PCMs

From: Developing a shortened version of the dementia knowledge assessment scale (DKAS-TC) with a sample in Taiwan: an item response theory approach

Models

Deviance

Parameters

-2LL

AIC

BIC

RMSEA

Unidimensional Rasch model

11,023.44

26

− 5511.55

11,075.44

11,179.52

0.070

4-dimensional Rasch model

10,889.01

25

−5446.33

10,939.02

11,103.23

0.058

Unidimensional PCM

18,452.79

51

− 9226.40

18,554.79

18,759.62

0.071

4-dimensional PCM

18,324.44

60

− 9160.93

18,444.44

1862.84

0.072

  1. Abbreviations: PCM partial credit model, −2LL 2 loglikelihood, AIC Akaike information criterion, BIC Bayesian information criterion, RMSEA root mean square error of approximation