Parent and youth report of youth anxiety: evidence for measurement invariance

• Published in: Journal of child psychology and psychiatry Vol. 55; no. 3; pp. 284 - 291
• Main Authors: Dirks, Melanie A., Weersing, V. Robin, Warnick, Erin, Gonzalez, Araceli, Alton, Megan, Dauser, Christine, Scahill, Lawrence, Woolston, Joseph
• Format: Journal Article
• Language: English
• Published: Oxford Blackwell Publishing Ltd 01.03.2014; Blackwell
• Subjects: Adolescent; African Americans; African Americans - psychology; Anxiety; Anxiety Disorders - diagnosis; Anxiety Disorders - psychology; Behavior disorders; Biological and medical sciences; Black American people; Black people; Child discipline; Child psychology; Diagnostic Self Evaluation; Emotional Disturbances; Female; Humans; informant disagreement; Informants; Male; measurement; Medical sciences; Parents; Parents - psychology; Psychology. Psychoanalysis. Psychiatry; Psychometrics. Diagnostic aid systems; Psychopathology. Psychiatry; Racial differences; Self Report - standards; Symptoms; Techniques and methods; Thresholds; United States; Young people; United States; United States > US; Human; informant disagreement; Self evaluation; Evaluation scale; Anxiety disorder; Psychometrics; Invariance; Test reliability; measurement; Parent rating; Adolescent; Anxiety; Measurement method
• ISSN: 0021-9630; 1469-7610
• DOI: 10.1111/jcpp.12159

Background We characterized parent‐youth disagreement in their report on the Screen for Child Anxiety Related Emotional Disorders (SCARED) and examined the equivalence of this measure across parent and youth report. Methods A clinically referred sample of 408 parent‐youth dyads (M age youth = 14.33, SD = 1.89; 53.7% male; 50.0% Non‐Hispanic White (NHW), 14.0% Hispanic, 29.7% African‐American) completed the SCARED. We examined (a) differences between parents and youth in the total number of symptoms reported (difference scores) and in their ratings of specific symptoms (q correlations), (b) demographic factors associated with these indices, and (c) equivalence of the pattern and magnitude of factor loadings (i.e., configural and metric invariance), as well as item thresholds and residual variances, across informants. Results The mean difference score was −2.13 (SD = 14.44), with youth reporting higher levels of symptoms, and the mean q correlation was .32 (SD = .24). Difference scores were greater for African‐American dyads than NHW pairs. We found complete configural, metric, and residual invariance, and partial threshold invariance. Differences in thresholds did not appear to reflect systematic differences between parent and youth report. Findings were comparable when analyses were conducted separately for NHW and ethnic minority families. Conclusion Findings provide further evidence for the importance of considering youth report when evaluating anxiety in African‐American families. The SCARED was invariant across informant reports, suggesting that it is appropriate to compare mean scores for these raters and that variability in parent and youth report is not attributable to their rating different constructs or using different thresholds to determine when symptoms are present.