Date of Award


Document Type



Edith Cowan University

Degree Name

Master of Psychology


School of Psychology


Faculty of Community Services, Education and Social Sciences

First Supervisor

Greg Dear


It is widely accepted in the literature that internalised childhood psychopathology including anx1cty is best identified by child self-report, followed by observations by parents and then to a lesser extent, by teachers. Although anxiety scales are a useful way to quantify anxiety symptomatology and progress through treatment, changes to classification systems (e.g., DSM-111-R to DSM-IV) make their ability to screen for the presence of psychiatric problems difficult. This research looked at whether the Anxiety Cluster (AC) scale of the Piers-Harris Children's Self-Concept Scale (PHCSCS) can be used to identify anxiety disorders in children. Two separate studies were conducted in the present research. Study 1 sought to compare AC scores from 160 ten year old Western Australian school children to normative information provided in the PHCSCS manual. Study 2 involved administering the Anxiety Disorders Interview Schedule for DSM IV: Child Version (AD!S-C) to a sample of 76 children drawn from Study I to compare AC scores to anxiety diagnosis. The 76 participants comprised all II children with AC scores between 0 and 5 (indicating anxiety), all 47 children with AC scores between 6 and 9 (considered a-typical) and a sample of 18 out of 100 children with AC scores between 10 and 14 (indicating no anxiety). Sampling of the 10 to 14 group was necessary due to resource limitations for this study. Results for Study I found that AC scores for this cohort were approximately I point higher (indicating less anxiety) than reported in the manual. This result was unlikely to be due to inadequate sampling, rather, that the normative information for the PHCSCS does not provide for age and sex differences. Results for Study 2 were that I 0 out of II subjects (91%) with AC scores between 0 and 5 met sufficient criteria for one or more anxiety disorder, whilst 17 out of 18 children (94%) with AC scores between I 0 and 14 did not. Although it would increase the number false positives, raising the clinical cut-off to an AC score of 7 is recommended as it would reduce the number of false negatives. It would appear therefore, that the AC scale has good utility for use as a first stage screening tool. As such, an application could be the selection of children for intervention groups who are likely to have an anxiety disorder (AC score 0 to 7) and as a reasonable means of precluding those children who are unlikely to have an anxiety disorder (AC score l 0 to 14). A second stage screening involving a structured diagnostic interview would be necessary for the remaining children (i.e., AC score 8 and 9). Limitations of this study are discussed in terms of inter-rater reliability of diagnostic interviews and the sampling of approximately l in 5 children with AC scores of 10- 14 due to resource constraints.