Two individuals coded parenting behavior. These individuals held Bachelor's level degrees and had extensive experience in the study of child development and in observational coding of parent–child assessments. The coding of maternal and paternal behavior was alternated between the two coders so that the coder who coded one parent did not code the other parent. The coders were trained on the Clark scales and the Iowa Family Interaction Rating Scales by the second author until they achieved at least 80% reliability. The coders were unaware of group membership and all other data. A minimum of 15% of the observations were selected at random for inter-rater reliability checks. For parent–infant interactions at 12 and 24 months, inter-rater reliability was calculated for 17% of the sample (n=38) and was high, with intraclass correlation coefficients ranging from .80 to .92. Inter-rater reliability was calculated for 15% of the sample with complete observational data at 36 months (n=24)2. Intraclass correlation coefficient was .97 for mothers' warmth/sensitivity and .96 for fathers' warmth/sensitivity.
