Assessment of Score Dependability of the Wisconsin Schizotypy Scales Using Generalizability Analysis

Abstract

To investigate the reliability of the Wisconsin Schizotypy Scales, this study applied generalizability analysis with two college student samples who completed the scales at two time points. The results indicated that the Revised Social Anhedonia Scale had acceptable levels of score dependability, but that the score dependability for the other scales (the Physical Anhedonia Scale, the Perceptual Aberration Scale, and the Magical Ideation Scale) was below an acceptable level of .80. The patterns of variance components suggested that the scales' items need improvement. Researchers can use the included tables to choose the number of items and occasions needed to get dependable score interpretations.

Brief assessment of schizotypy: Developing short forms of the Wisconsin Schizotypy Scales

“…They consisted of the following items from the original scales: Magical Ideation Scale (items 1,2,5,6,9,12,13,21,22,23,24,25,26,27,29), Perceptual Aberration Scale (items 8, 10, 11, 13, 16, 19, 21, 23, 24, 25, 26, 27, 29, 30, 31), Revised Social Anhedonia Scale (items 1,3,4,7,10,13,14,17,19,21,26,30,31,35,37), and Physical Anhedonia Scale (items 3, 10, 15, 19, 24, 29, 35, 36, 39, 42, 45, 46, 47, 54, 60). Unlike the participants in our prior studies ( [Winterstein et al, 2010] and [Winterstein et al, in press]), the participants in this sample completed only the 60 items from the four short forms. To evaluate the validity of the short forms' scores, we measured many additional constructs.…”

“…In addition to their wide use in cross-sectional studies, the Wisconsin Schizotypy Scales have been used successfully in several longitudinal studies of the development of schizophrenia-spectrum disorders (e.g., [Chapman et al, 1994], [Gooding et al, 2005], [Gooding et al, 2007] and [Kwapil, 1998]). Because of the popularity of the Wisconsin Schizotypy Scales, recent work has examined their psychometric properties employing current measurement models ( [Winterstein et al, in press] and [Winterstein et al, 2010]). The scales were originally developed using classical test theory.…”

Psychometric Properties of the Wisconsin Schizotypy Scales in an Undergraduate Sample: Classical Test Theory, Item Response Theory, and Differential Item Functioning

“…The psychometric properties of the WSS within the CTT framework have been described in the original source articles and in subsequent reviews (e.g., Kwapil et al 2002Kwapil et al , 2008. In brief, research has found good evidence for internal consistency, such as internal-consistency coefficients ranging from .79 to .91 (Winterstein et al 2010). In a study of temporal stability, test-retest reliability coefficients (across a period of 2 to 3 months) ranged from .63 to .81, and subsequent generalizability analyses showed that time explained minimal variance (less than 1%) in WSS scores (Winterstein et al 2010).…”

“…In brief, research has found good evidence for internal consistency, such as internal-consistency coefficients ranging from .79 to .91 (Winterstein et al 2010). In a study of temporal stability, test-retest reliability coefficients (across a period of 2 to 3 months) ranged from .63 to .81, and subsequent generalizability analyses showed that time explained minimal variance (less than 1%) in WSS scores (Winterstein et al 2010). Confirmatory factor analyses have provided evidence for the scales' structural validity, particularly the division of the scales into broader positive and negative symptom dimensions (Kwapil et al 2008).…”

“…Exploring DIF in the WSS items is important because past work has found several group differences-such as differences between men and women and differences between racial and ethnic groups (Chmielewski et al 1995;Kwapil et al 2002)-and it is critical to know if these represent true trait differences. Recent generalizability analyses of the scales (Winterstein et al 2010) hint at the possibility of DIF in many of the WSS items. Those analyses found that significant variance in the WSS scores was associated with item-by-person interactions.…”