The main purpose of the article was to examine the reliability, validity, and cross-informant agreement of the German SCARED in a clinical sample.
The instrument yielded good psychometric properties in self- and parent report in a clinical sample. First of all, the administration of the SCARED was feasible in a sample of patients and their parents beginning psychotherapy treatment. Furthermore, internal consistencies were high for the total scores and satisfying for the scales. These results are consistent with the findings for the English version in a clinical sample and the German version in a non-clinical sample with Alphas for the total scores around α = .90 [18, 25].
Convergent validity was supported by high correlations with internalizing CBCL/YSR symptom scores in the expected direction. Discriminant validity was supported by correlative independence from externalizing scores (CBCL/YSR). These findings mirror the results of Essau and her colleagues in the community sample with moderate to high correlations with internalizing and total scores of the YSR . However, discriminant validity seemed to be lower in the community sample with moderate correlations of the SCARED scales with externalizing YSR scores .
Cross-informant agreement was in a moderate range and comparable to agreement scores for the English SCARED. The study by Wren and colleagues (2004) reports parent child agreement to average at r = .55 for the total score, while our study yielded correlations of r = .50 for father-patient agreement and r = .51 for the mother-patient agreement. Previous studies have generally found low to moderate parent-child agreement [10, 12, 18, 30, 31]. Compared to these studies, cross-informant agreement of the German SCARED can be considered being in the upper range in this sample. Previous studies did not differentiate between father and mother perspective. In our sample, the cross-informant agreement differs considerably by proxy and symptom group. While father and child agreement was higher for generalized and separation anxiety as well as social and school phobia, father's ratings of panic and somatic symptoms seemed to assess not panic but rather something similar to school phobia symptoms.
Although cross-informant agreement was acceptable in terms of correlational patterns, we also found strong differences for symptom severity and number of symptoms reported. Children and adolescents themselves tended to report more symptoms with higher degrees of intensity than their fathers and mothers. These findings are again consistent with research on the English version and other studies on informant agreement for anxiety ratings [11, 15].
Interestingly, excess symptom reporting by children/adolescents did not occur for school and social phobia. This result replicates findings by DiBartolo and her colleagues  indicating that school and social phobia were underreported compared to reports of other anxiety symptoms which were related to the child's concern regarding positive self-presentation. It is possible that for self-representation concerns, these symptoms were played down. Therefore relying on child information alone for these subscales could lead to less valid assessments. Parents underreporting of somatic/panic, and generalized anxiety symptoms on the one hand, and that children refuse (and/or deny) symptoms of school and social phobia on the other hand, stress the importance of obtaining different perspectives in the diagnostic process of children and adolescents.
There are a number of potential limitations to this study. First, the sample size was relatively small. For this reason subgroups of the sample could not be contrasted with each other (e. g. different answering patterns for age and gender). Due to the sample size, optimal cut-off points for parents could not be determined. However, power analyses with GPower yielded satisfying power (>80%) for the current calculations. Secondly, it would be of interest, whether the SCARED is suitable to differentiate between anxiety and depressive symptoms. However, the ability of the instrument to discriminate between these disorders could not be tested adequately, as comorbidity with affective disorders was too high. As typical for clinical samples [2, 32], about 20% of the children/adolescents with an anxiety disorder were also diagnosed with a comorbid affective disorder. Furthermore, future investigations could analyse a clinical group of only adolescents with anxiety disorders and then compare the subgroups of anxiety with the SCARED sub-scales, to test the ability of the SCARED to differentiate between anxiety disorders. Finally, we had no information on parents' diagnostic status, which has been found to influence their reporting of their children's problems .
Our study has a number of advantages as well. One advantage was the inclusion of father and mother perspective. Thus, the feasibility of the German parent-report version could be analyzed for both parents. A reliable and valid parent-report is necessary especially in younger children to complement self-reports with parent ratings for valid information. Furthermore, as the sample consisted of children and adolescents beginning outpatient psychotherapy treatment, this study was suitable to test the feasibility of the administration of the German SCARED in a clinical sample. The families seemed to be quite representative of the German population in terms of parents' education. 30% of the parents held a university or technical degree compared to about 24% in the general population 
A clinical implication derived from our data is that integrating both parent and child/adolescent perspectives should become the standard procedure in screening for anxiety disorders. Discrepancies in the degree of symptom reports between parents and children call for further research on individual cut-off-scores for the different perspectives.
In summary, good psychometric properties - comparable to the established English SCARED version - suggest the successful translation of the SCARED into the German language. Overall, the findings stress that the SCARED is a feasible, reliable, and valid screening instrument for parents and children/adolescents, and thus support the application of the German SCARED in clinical and research settings.