Treatment integrity is an underexposed issue in the phonological awareness intervention research. The current study assessed the integrity of treatment of the families (N = 32) participating in the experimental condition of a home-based pre-reading intervention study. The participating kindergartners were all genetically at risk for developing dyslexia. Two aspects of treatment integrity, the number of lessons completed (quantity) and the quality of the administration of the programme (observed in a videotaped session), were investigated. The level of treatment integrity turned out to be 66% when completion of all lessons was taken as quantitative criterion, and about 74% when quality of the parent-child interaction was assessed. The two measures could predict the pre-reading skills at the end of kindergarten. Together they accounted for 43% of the variance in this dependent variable. Together with pre-reading scores at the pre-test the total predicted variance was 87%. The number of lessons completed still contributed 12% to the prediction after controlling for pre-test scores. The results indicated that treatment integrity indeed appears to be an important aspect of treatment outcome and should therefore be included in intervention studies. (Contains 2 figures and 4 tables.)