Table 1.
Reliability for Traditional Global Bias Scores from Study 1
| Test and Data | Overall Anxiety Disorders Sample (n =24) | Highest Subgroup (Group Name) |
|---|---|---|
| Test-Retest Stability | ||
| 100 Trials (Price Method) | ICC = −.12 (CI: −.51, .29) | ICC = .13, (CI = −.38, .56) (Anxiety disorder, no current mood disorder) |
| 200 Trials (Price Method) | ICC = .08 (CI: −.35, .47) | ICC = .18, (CI = −.32, .59) (Anxiety, no current mood disorder) |
| First Half (Price Method) | ICC = −.13 (CI: −.51, .29) | ICC = −.04, (CI = −.60, .50) (Panic disorder) |
| Second Half (Price Method) | ICC = −.39 (CI: −.70, .03) | ICC = −.38, (CI = −.76, .18) (Generalized social anxiety disorder) |
| All Trials (Price Method) | ICC = −.36 (CI: −.70, .06) | ICC = .13, (CI = −.46, .62) (Panic disorder) |
| All Trials (Bias for Each Word) | ICC = −.30 (CI: −.60, .10) | ICC = .08, (CI = −.43, .56) (Panic disorder) |
| All Trials (Most Negative Words) | ICC = .06 (CI: −.30, .43) | ICC = .47, (CI = −.003, .79) (Panic disorder) |
| Internal Consistency | ||
| Visit 1 (Bias for Each Word) | α = −.32 (CI: −1.20, .33) | α = −.25 (CI: −1.39, .52) (Generalized social anxiety disorder) |
| Visit 2 (Bias for Each Word) | α = −.48 (CI: −1.46, .25) | α = −.37 (CI: −1.62, .48) (Generalized social anxiety disorder) |
| Split Half Reliability | ||
| Visit 1 (Bias for Each Word) | r = −.38 (CI: −.52, .25) | r = −.16 (CI: −2.10, .57) (Anxiety, no current mood disorder) |
| Visit 2 (Bias for Each Word) | r = −.06 (CI: −1.46, .54) | r = .32 (CI: −.79, .75) (Anxiety, no current mood disorder) |
| Visit 1 (Price Method) | r = .29 (CI: −.64, .69) | r = .42 (CI: −.21, .65) (Anxiety disorder, no current mood disorder) |
| Visit 2 (Price Method) | r = −.58 (CI: −2.66, .32) | r = −.36 (CI: −3.24, .56) |
Note. In many cases, negative correlations mean that the assumptions of the reliability test were violated, resulting in numbers within confidence intervals that are outside the theoretical parameters. The confidence intervals are presented as computed to provide the clearest picture of the data. ICC = Intraclass Correlation Coefficient (two-way random for absolute agreement, single measure); CI = 95% confidence interval. Subgroups tested for reliability included all participants, participants with an anxiety disorder with no current mood disorder, participants with panic disorder with or without agoraphobia and agoraphobia without panic disorder, and participants with generalized social anxiety disorder. Each method of computing the bias score is described in full in the supplemental material. Further information is available in Study 1 of the supplemental material.