Analyses

Effect sizes were calculated as Cohen’s d, corrected for the association between pre- and post-transition scores (Morris and DeShon, 2002). Paired sample t-tests were used to investigate changes in group means. In the parent- and self-report data, these had sufficient (>80%) power to detect medium effects (Cohen’s d = 0.57) and above. In the teacher-report data, they were powered for large (Cohen’s d = 0.66) effects. Associations between continuous variables were tested using Pearson’s correlations.

We used the pre-established, standardised cut points (described above, in the ‘Measures’ section) to identify individuals scoring in the clinical range on measures of psychopathology (SDQ, BYI-II) and adaptive function (VABS-II). Changes over time in the proportion of children scoring in the clinical range were investigated using McNemar’s test. Also, to get a sense of whether rates of psychopathology and adaptive function impairment were elevated in this sample, two-tailed chi-squared or Fisher’s exact tests were used (depending on whether frequency counts were below five) to analyse whether more participants scored in the clinical range than would be expected in the general population.