Sequential coding of maternal sensitivity: application of nonlinear dynamic analyses and reliability

Abstract

In infant development, maternal sensitivity is a central construct in the study of the processes involved in the caregiver-infant relationship and child attachment. The relevance of these dyadic processes for infant and later child development requires the advancing in the measurement of maternal sensitivity. The measurement of this construct, that reveals its complexity through different ‘windows’ using different approaches and operational definitions, has brought researchers’ attention into the methodological arena. Mother–infant interaction has been mainly quantified through rating scales and sequential strategies using observational methodology. The present study is part of a research program whose purpose is to measure maternal sensitivity dimensions using sequential coding in real time, the free-play paradigm and the early mother–child interaction coding system (CITMI-R: by its Spanish initials) instrument. This study was designed to further the analysis of the quality of the data obtained by this coding system. For this purpose, six trained observers coded fifty-seven mother–infant free-play episodes. The reliability of their coding was estimated: firstly, examining the inter-coder agreement with the Alignment Kappa, thus, incorporating the temporal dimension of the measurement. Secondly, analyzing the inter-coder agreement on parameters that consider the sequential and temporal organization by applying variables derived from the State Space Grid methodology and the nonlinear dynamic system approach: Contingency, through dyadic events, visits to them, predictability and duration of event, visit and dyadic behavioral states. The results indicated acceptable inter-observer reliability values based on the standards for use on parameters of interest in the study of sensitivity.