Informant agreement. For the HC group, child-reported anxiety symptoms were higher than parent-report (Figures 1–2, green lines). This suggests that children free of clinically significant levels of anxiety report more symptoms than their parents. Among anxious participants a different pattern of results emerged such that parental reports of anxiety were largely equal to or higher than children’s reports. Author Manuscript Within the treatment-seeking sample, correlations between parent and child report were relatively high when collapsed across all groups (i.e., HC and anxious youth). However, as with prior reports on child-parent agreement (Xxxxxxxx et al., 1999; Xxxxxxxx et al., 1997; Cosi et al., 2010; Xxxxx et al., 1999), when considered separately, correlations among HCs and anxious youth were relatively modest. Few prior studies have had a sufficient sample size to assess child-parent agreement in youth with specific anxiety diagnoses. We were able to assess such agreement separately for GAD, SAD, and GAD+SAD comorbid samples, identifying distinct patterns of results. Specifically, while child-parent correlations for GAD Author Manuscript Xxxxxxxxx et al. Page 13 and GAD+SAD patients were modest, strikingly little agreement was observed among patients with SAD. Author Manuscript Among the community sample, correlations between parent and child report were modest when collapsed across all groups (i.e., HC and anxious youth). This may have been driven by the low level of agreement among the HC’s, as modest correlations were observed across all patients, and within each diagnostic group. Indeed, previous studies of non-diagnosed community samples have also found that children report higher scores than their parents (Cosi et al., 2010; Xxxxx et al., 1999; Xx et al., 2008; Xxxx et al., 2007; Xxxx, Bridge, & Xxxxxxxx, 2004), further suggesting a difference in how children and parents generally report on the child’s symptoms of anxiety. While the data does not necessarily support the proposition that children are responding more accurately than their parents (as indicated by the ROC analyses), it is possible that children are, in certain cases, more knowledgeable about their own feelings and symptoms of anxiety as opposed to the symptoms externally observable to their parents. Therefore, discrepancies in child- and parent-report could be related to communication between the child and his/her parent(s). Author Manuscript Given informant discrepancy, generating a composite score...
Informant agreement. Bivariate Xxxxxxx correlations assessed the association between child- and parent report scores on the Total, GA, and SA subscales across each study’s full sample, and within each diagnostic group (HC, GAD, SAD, GAD+SAD). Non-parametric Xxxxxxxx Rho analyses largely confirmed findings (see supplementary materials for further details). Group differences in the magnitude and direction (i.e. whether the child-report was higher than parent-report or vice versa) of child-parent discrepancies on Total, GA, and SA subscales were assessed with a Group (HC, GAD, SAD, GAD+SAD) X Informant (child, parent) repeated measures ANOVA. Significant interactions were interrogated with Bonferroni-corrected pairwise comparisons. Because six comparisons were conducted for each scale, the p-value was adjusted for multiple comparisons for each scale (critical p-value = 0.05/6 = .008). Xxxxxxx-Xxxxxx tests on parent- child difference scores, and follow-up analyses with Xxxxxxxx Signed Rank tests, confirmed unequal variance did not bias findings.
