CAARS 2 Manual Chapter 10: Summary |
In addition to qualitative aspects of fairness that were addressed throughout development, quantitative evidence
clearly demonstrates that the CAARS 2 meets test standards for psychometric evidence of fairness (AERA, APA, & NCME,
2014). Measurement invariance, differential test functioning, and mean differences for the CAARS 2 were examined for
key demographic groups (i.e., gender, race/ethnicity, country of residence, and EL) for which sufficient data were
available. The CAARS 2 Content Scales, Associated Clinical Concern Items, and Impairment & Functional Outcome Items
were investigated with respect to the available demographic groups, using samples matched by gender, education
level, language(s) spoken, clinical status, race/ethnicity, and/or age. Consistent with predictions that the CAARS 2
would be unbiased (i.e., not favoring or disfavoring any particular demographic group), the overwhelming majority of
analyses conducted reflected similar results across groups (i.e., differences were not statistically significant),
and very small effect sizes were observed overall. In other words, there were no meaningful group differences in
CAARS 2 scores across gender, race/ethnicity, country of residence, and education level, meaning that these
demographic variables do not significantly influence CAARS 2 scores. There was strong quantitative evidence to
support the lack of measurement bias and minimal impact of demographic characteristics on the CAARS 2, indicating
that CAARS 2 results are fair, equitable, and generalizable for individuals from diverse backgrounds.