For criterion validity, positive high correlations between the long-form and each short-form questionnaire were expected. For discriminant construct validity, the hypothesis that the scores are highest in the oro-facial, lower in the orthodontic and lowest in the paediatric dentistry group was tested. It was also hypothesized that within each of the three groups scores would be highest for those with the most severe clinical condition. For correlational construct validity, positive correlations between the scores and children's global ratings of oral health and well-being were tested. Since the former is a measure of health and the latter a measure of health-related quality of life, it was predicted that the correlation coefficient would be higher for the rating of well-being than for the rating of oral health.
Relative validity (RV) estimates were computed as the ratios of F statistics for the short-form questionnaires and the original CPQ11–14. They indicate in proportional terms how much more or less precise a short-form questionnaire is in relation to the original CPQ11–14 [16 (link),17 (link)].
Internal consistency reliability was determined determined using Cronbach's alpha. Alphas were also calculated with each item deleted. Corrected item total correlations were also compared. Test-retest reliability was assessed using the intraclass correlation coefficient (ICC). This was calculated using a one-way analysis of variance random effects parallel model [18 (link),19 (link)].