Towards the development of a coding scheme for empathic effort and dyadic empathic alignment

The measurement of empathic behaviour in a social interaction poses a complex challenge. Building on the recent argument that greater empirical focus should be placed on the inherently interpersonal nature of empathy (Main, Walle, Kho & Halpern, 2017), we present a coding approach for the examination of the dynamics of live, unfolding empathy in a social interaction. Empathic behaviour is operationalised in terms of empathic effort, which is the observable effort displayed by a person to align both cognitively and affectively with their partner. Instances where interactants are displaying equal levels of empathic effort are labelled as periods of dyadic empathic alignment. An extensive series of rating sessions were conducted to create the scheme. Lab assistants and online raters coded thin-slices of videotaped interactions for empathising behaviour, non-verbal cues and conversational tone. Several behaviours were consistently associated with high levels of perceived empathy. These were integrated into a coding scheme which raters then used to reliably code a new, and more diverse, set of dyadic interactions. Designed to be flexible, the scheme can be used in multiple ways to suit a research team's resources and time constraints. Coders rating longer durations of empathic behaviour can rate their general impressions of empathic effort per minute, for example. Alternatively, the scheme can be used to obtain continuous empathy scores which are well-suited to regression analyses. By using the outlined scoring system, in conjunction with the online slider scale functionality, coders can provide precise continuous ratings for each of the specified categories, allowing for the computation of an overall score of empathic effort. Finally, researchers can perform temporal coding using annotation software, whereby the exact onset and offset times of low, moderate and high empathic effort, in addition to other micro-level non-verbal behaviours of interest, are labelled as they occur.