From Construct Definitions to Narrative Response Models: The Case of Self-Efficacy

Psychology faces crises of theory and measurement, where poorly specified constructs hinder the validity of measurement instruments. This article turns prior pleas to transparently explicate comprehensive construct definitions and accompanying measurement instructions into practice, using self-efficacy as an exemplary case. Drawing primarily on Bandura’s conceptualization, we explicate the key elements distinguishing self-efficacy from related constructs and examine implications for measurement. More importantly, this is the first time to introduce Narrative Response Models (NRMs) as a tool to describe and study how hypothesized processes that need to occur allow inferences about a psychological construct from the registered responses to an item used to assess that construct (in this case self-efficacy). Without this, there is no reason to assume the responses can tell anything about the construct: the item is invalid. Because NRMs as a tool were developed to justify confidence in the validity of an item, they explicate many underlying assumptions that warrant further investigation. In the case of self-efficacy, for example, the retrieval of gist traces and the simulation of what it takes to engage in the behaviour of interest. As such, NRMs provide scaffolding for the systematic study of a measurement instrument’s validity.