Internal Consistency. The CAARS 2 Self-Report and Observer scale scores both demonstrate excellent internal consistency. Median omega coefficient in the Normative Sample: Self-Report = .94 and Observer = .95.

Standard error of measurement (SEM). The SEM values are low for CAARS 2 scales and across raters (median SEM in the Normative Samples: Self-Report = 2.52 and Observer = 2.27). The low values indicate very little error in the estimated true scores.

Test Information. Precision is high for the target ranges of the test information functions (TIF), indicating sufficient measurement precision and low conditional standard error of measurement, providing additional evidence for the reliability of the CAARS 2 scales.

Test-Retest Reliability. The CAARS 2 scales display excellent test-retest reliability (median r: Self-Report = .92 and Observer = .84). The stable nature of the scores provides assurance that changes observed in scores over time are due to a true change in symptoms rather than to imprecise measurement.

Inter-Rater Reliability. Moderate to strong relationships were found between same rater types, while weaker relationships were found across different rater types. These results demonstrate how the CAARS 2 can effectively capture different points of view when completed by people in different roles. These results serve to highlight the importance of examining information from multiple sources when conducting a comprehensive assessment.

• Same Rater Type. When two observers of the same rater type (e.g., two friends) rate the same individual, there is evidence of moderate to strong relationships across the CAARS 2 scale scores (median r = .54).
  
  
• Different Rater Types. There are moderate associations between scores when comparing self-reported ratings to observer ratings (median r = .44), and when comparing two different rater types (e.g., one friend and one relative; median r = .45).