Psychometric analysis of the emotional availability scales for video-recorded interactions between parents and their preschool-aged children

TL;DR

This study evaluates the Emotional Availability Scales for measuring parent-child emotional connections in preschoolers, finding issues with the scales' reliability and structure.

Contribution

The study provides a psychometric evaluation of the EAS for preschoolers, revealing structural issues and suggesting the need for alternative measures.

Findings

Psychometric analyses did not validate the EAS's postulated structure and measurement model.

Post hoc exploratory factor analysis suggested a non-interpretable two-factor structure.

Ceiling effects on item and scale levels are not unique to this study.

Abstract

In the context of parent-child interaction, the Emotional Availability Scales have been developed to capture a dyad's emotional connection in an observational setting by four parental and two child-related scales. This study aims to test the psychometric foundation of the EAS, including basic descriptive preconditions on the item level and structural validity on the scale level, for a preschool-aged sample; as such, it complements analyses by Aran for a sample of infants. The sample of parents and their preschool-aged children is a mixed clinically referred and non-clinical sample from a midsize city in Germany. Interactions were observed in a free-play setting and rated with the EAS by two blind and certificated raters. Several model tests indicate violations for the structural model as well as all six measurement models. An additional post hoc exploratory factor analysis with parallel analysis suggests a non-interpretable two-factor structure. Psychometric analyses did not validate the EAS's postulated structure and measurement model. A post hoc literature review showed that ceiling effects on the item and scale levels are not unique to our study. However, traditionally important concepts or terms of parent-child relationships covered by the EAS are not rejected by our study and can be assessed with alternative measures, but these also need psychometric evaluation in the future.

Figures