I had difficulties in answering important questions about the support of some of the conclusions that White et al (Reference White, Leach and Sims1999) draw in their work on validation of the Hospital Anxiety and Depression Scale for use with adolescents.

1. (a) It is not clear whether the non-clinical sample used in the study can be considered independent since the method of selection is not mentioned.

2. (b) Of the 248 children (110 girls) who were tested first, 180 were re-tested. However, the girl/boy ratio in the latter group is not indicated. Moreover, the outpatient sample comprised 48 subjects (27 girls) and the deliberate self-harm (DSH) sample had 38 subjects (30 girls). Considering the disproportionate group sizes and gender distribution it is surprising to find that the variances in the different groups are homogeneous. However, this information cannot be deduced from the published data. As a result, it is very difficult to assess fundamental requirements for the F-test.

3. (c) Girls are 44% of the non-clinical sample, 56% of the out-patient sample and, more importantly, 79% of the DSH group. The authors conclude that there is a statistically significant gender difference, with girls scoring higher than boys in both depression and anxiety scales. Assuming that the F-test's requirements are met, then it may not seem surprising to find an overall significant difference detected by the F-test because of the characteristics of the DSH group.

4. (d) As the authors do not report multiple comparisons between the groups, it is not possible to know whether the differences remain when the DSH group is excluded.

5. (e) The analysis does not include techniques to control for gender, which appears to be a very important confounder.

6. (f) The authors assessed test-retest reliability with Pearson's correlation coefficient. As this technique does not take into account errors of measurement, it does not measure agreement and its results are not meaningful.