Psychometric properties of the five-minute speech sample

Abstract

Background: Parental schemas are implicit beliefs parents have about their child that then shape their behaviors and impact child well-being and adjustment. This research evaluates the psychometric properties of the Five-Minute Speech Sample coded with the Family Affective Attitude Rating Scale of the (FMSS-FAARS) to assess parental schemas among caregivers involved in the child welfare system (CWS). Objective: Evaluate the psychometric properties of the FMSS-FAARS with a large, diverse sample of caregivers involved with the CWS. Participants: The sample consists of 2240 caregivers from the third cohort of the National Survey of Child and Adolescent Wellbeing. Methods: Caregivers completed the FMSS, which was then coded using the FAARS to assess parental schemas. Analyses included inter-rater reliability (IRR), internal consistency, and convergent validity. To assess convergent validity, the Home Observation for Measurement of the Environment - Short Form (HOME-SF), was examined for correlations with FMSS-FAARS. Results: The IRR of the FMSS-FAARS scales (Negative Relational Schema [NRS] and Positive Relational Schema [PRS]) ranged from nearly perfect (weighted kappa = 0.98) to substantial (weighted kappa = 0.63). Although the NRS exhibited acceptable internal consistency (S2 = 0.91), that of the PRS was poor (S2 = 0.58) and varied by demographic characteristics. Results of convergent validity analyses indicated variation of correlations between the HOME-SF and the FMSS-FAARS by age and subgroups. Conclusions: The FMSS-FAARS demonstrated inconsistent psychometric properties when administered to caregivers involved in the CWS, warranting additional empirical investigation. Novel approaches to coding the FMSS may be necessary to represent parental schemas effectively in a multi-dimensional and inclusive manner.