The Thought Disorder Index: A reliability study.

Abstract

This article reports interrater reliability among 4 teams of raters who independently evaluated thought disorder in 20 Rorschach protocols using the Thought Disorder Index (TDI). Intraclass correlation coefficients were calculated to assess the degree of association among the 4 teams for total thought disorder scores, severity levels, and qualitative thought disorder factors. Highly acceptable agreement was obtained. Spearman rank order correlation coefficients for these same variables were significant for all possible pairings of teams. A repeated-measures analysis of variance indicated that the absolute amount of thought disorder tagged by each team differed even though the teams' relative rankings of thought disorder among subjects was very similar. Such scoring differences reflect individual differences in threshold for detecting deviant thinking.

Exploring a multifactorial, clinical model of thought disorder : application of a dimensional, transdiagnostic approach.

“…TDI total score has been shown to have good inter-rater reliability, with correlations ranging from .80 to .93 (Coleman et al, 1993;Johnston & Holzman, 1979;Solovay et al, 1987) and an intra-class correlation (ICC) of .74 (Coleman et al, 1993 (Solovay, Shenton, & Holzman, 1987, p. 18). …”

Putting the person first : an examination of thought disorder and personality heterogeneity in schizophrenia.

“…Coleman et al (1993) found strong inter-rater reliability for TDI total scores ranging from r s = .80 to .90. Furthermore, Coleman et al (1993) showed intraclass correlations of .77, .72, and .77 for levels 0.25, 0.5, and 0.75, respectively (there were not enough responses at the 1.0 level to calculate reliability), and intraclass correlations of .58, .76, and .86 for idiosyncratic verbalizations, combinatory thinking, and irrelevant intrusions, respectively (these were the only categories for which there were enough instances to calculate reliability). Finally, it should be noted that the TDI has been shown to be unrelated to race and socioeconomic status (Haimo & Holzman, 1979;Johnston & Holzman, 1979) and use of the Rorschach in assessing thought disorder has received support even from strong critics of the Rorschach (Lillienfeld, Wood, & Garb, 2000).…”

“…Similarly, Solovay, Shenton, and Holzman (1987) examined a sample of individuals with schizophrenia, bipolar disorder and healthy controls, and showed strong inter-rater reliability using the Spearman-Brown formula for two independent raters for TDI total score (r = .89), severity level (r = .79), and categories (r = .81). Coleman et al (1993) used four independent teams of raters to examine thought disorder severity in a sample of individuals with schizophrenia, schizoaffective disorder, or bipolar disorder, as well as first-degree relatives of these patients. Coleman et al (1993) found strong inter-rater reliability for TDI total scores ranging from r s = .80 to .90.…”

Understanding thought disorder in schizophrenia-spectrum disorders : exploring the relation and implications of affect.

“…Similarly, Solovay, Shenton, and Holzman (1987) examined a sample of individuals with schizophrenia, bipolar disorder and healthy controls, and showed strong inter-rater reliability using the Spearman-Brown formula for two independent raters for TDI total score (r = .89), severity level (r = .79), and categories (r = .81). Coleman et al (1993) used four independent teams of raters to examine thought disorder severity in a sample of individuals with schizophrenia, schizoaffective disorder, or bipolar disorder, as well as first-degree relatives of these patients. Coleman et al (1993) found strong inter-rater reliability for TDI total scores ranging from r s = .80 to .90.…”

“…Coleman et al (1993) used four independent teams of raters to examine thought disorder severity in a sample of individuals with schizophrenia, schizoaffective disorder, or bipolar disorder, as well as first-degree relatives of these patients. Coleman et al (1993) found strong inter-rater reliability for TDI total scores ranging from r s = .80 to .90. Furthermore, Coleman et al (1993) showed intraclass correlations of .77, .72, and .77 for levels 0.25, 0.5, and 0.75, respectively (there were not enough responses at the 1.0 level to calculate reliability), and intraclass correlations of .58, .76, and .86 for idiosyncratic verbalizations, combinatory thinking, and irrelevant intrusions, respectively (these were the only categories for which there were enough instances to calculate reliability).…”