Table 1.
Characteristic of the study participants.
Table 2.
Conditional likelihood ratio test of overall model fit and global DIF.
Table 3.
Conversion table for score adjustment.
Fig 1.
GLLRMs of the SSS for sample 1 (left) and sample 2 (right). Note. The Markov graph nodes represent the item numbers, the exogenous variables, and the latent trait. Disconnected nodes indicate that variables are conditionally independent and partial γ inform the magnitude of the LD and DIF. When LD/DIF was unsubstantial (γ<0.1), the edges were omitted (see S1 Fig).
Fig 2.
Item characteristic curves of the GLLRM for sample 1 (left) and sample 2 (right). Note. The x-axis indicates the latent trait (“Social support”) and the y-axis indicates the item score. The points represent the average observed item responses in each of the seven class intervals. Since Item 2 had DIF by employment status, the Item Characteristic Curves (ICCs) for participants employed (black) and not employed (red) are displayed.
Fig 3.
Test Information function for sample 1 (left) and sample 2 (right). Note. The x-axis indicates the latent trait (“Social support”) and the y-axis indicates the information. The red line with shaded area represents the Test Information Function and the blue line represents the Standard Error of Measurement (SEM).