The German version of the Conners Adult ADHD Rating Scales (CAARS) has proven to show very high model fit in confirmative factor analyses with the established factors inattention/memory problems, hyperactivity/restlessness, impulsivity/emotional lability, and problems with self-concept in both large healthy control and ADHD patient samples. This study now presents data on the psychometric properties of the German CAARS-self-report (CAARS-S) and observer-report (CAARS-O) questionnaires.

CAARS-S/O and questions on sociodemographic variables were filled out by 466 patients with ADHD, 847 healthy control subjects that already participated in two prior studies, and a total of 896 observer data sets were available. Cronbach's-alpha was calculated to obtain internal reliability coefficients. Pearson correlations were performed to assess test-retest reliability, and concurrent, criterion, and discriminant validity. Receiver Operating Characteristics (ROC-analyses) were used to establish sensitivity and specificity for all subscales.

Coefficient alphas ranged from .74 to .95, and test-retest reliability from .85 to .92 for the CAARS-S, and from .65 to .85 for the CAARS-O. All CAARS subscales, except problems with self-concept correlated significantly with the Barrett Impulsiveness Scale (BIS), but not with the Wender Utah Rating Scale (WURS). Criterion validity was established with ADHD subtype and diagnosis based on DSM-IV criteria. Sensitivity and specificity were high for all four subscales.

The reported results confirm our previous study and show that the German CAARS-S/O do indeed represent a reliable and cross-culturally valid measure of current ADHD symptoms in adults.

Attention-deficit/hyperactivity disorder (ADHD) often persists into adulthood. Instruments for diagnosing ADHD in childhood are well validated and reliable, but diagnosis of ADHD in adults remains problematic. Attempts have been made to develop criteria specific for adult ADHD, resulting in the development of self-report and observer-rated questionnaires. To date, the Conners Adult ADHD Rating Scales (CAARS) are the international standard for questionnaire assessment of ADHD. The current study evaluates a German version of the CAARS self-report (CAARS-S).

Eight hundred and fifty healthy German control subjects were recruited to fill out the CAARS-S and to answer questions on sociodemographic variables. Explorative and confirmative factor analyses were conducted to obtain the factor structure for the German model and to replicate the factor structure of the original American model. Analyses on gender, age, and education level were calculated for normative data.

The explorative factor analysis of the German sample results in a six-factor solution that explained 52% of the variance. A confirmative analysis that was based on the 42 items of the original American model showed a high model-fit. Analyses of normative data showed significant influences of age, gender, and education level on the emerging subscales.

Even though the explorative factor analysis yields a solution different from the American original, the confirmative factor analysis results in such a high model-fit that use of the American version is justified with respect to international multicenter studies, for which this instrument will be highly valuable.