Figure 1.
The test information curve of the HADS based on the bifactor analysis for the general distress factor.
X-axis represents severity of the general factor (theta), which had been standardized (0 being average, 1 being a standard deviation). The Y-axis represents the test information value. Test information is a kind of reliability criterion in IRT models, the bigger the test information value, the less measurement error, and better reliability. In contrast to models built using CTT, in IRT models, there is a test information value corresponding to every severity point, representing the reliability at that level of severity. We get the test information curve by connecting all these values.
Table 1.
Main parameters of bifactor analysis.
Figure 2.
Structure of the hierarchical model of the HADS built using bifactor analysis.
In the original scale, 14 items load on 2 subscales (anxiety and depression) respectively. In which, item 1, 3, 5, 7, 9, 11, and 13 belong to anxiety subscale. And item 2, 4, 6, 8, 10, 12, and 14 belong to depression subscale. In the bifactor analysis, all the items have loadings on both the general distress factor and one of the subscale specific factors.
Table 2.
Distributions of anxiety and depression disorders in different diseases.
Table 3.
Correlations between items and HADS scales.
Table 4.
Model fit comparison between unidimensional, bidimensional, tridimensional, and hierarchical models.