Using Implicit Measures to Assess User Experience in Children: A Case Study on the Application of the Implicit Association Test (IAT)

Abstract

The number of multimodal agents aimed at children with or without neurodevelopmental disorders (NDD) has increased tremendously during the last decade. As this expands, so does research into methods, tools, and metrics that can reliably assess their impact. Traditionally, the majority of UX Research tools have been produced for an adult audience, with fewer tools developed for a younger population. Furthermore, most of these tools use a "direct"method, in which detailed questions are asked directly to the individuals. However, when assessing youngsters, and mainly when direct inquiries are posed, the literature identifies several challenges and pitfalls not usually faced when testing adults. If overlooked, they might lead to biased judgments. This paper proposes a novel approach to UX Evaluation using implicit metrics, which offers the obvious advantage of avoiding direct questions. We investigated the application of the Implicit Association Test (IAT) - one of the most acknowledged tests in psychology to reveal unconscious attitudes, automatic preferences, and hidden biases - to determine whether 60 school-aged children enjoyed a multimodal interface dedicated to language assessment. The results, although preliminary, disclose discrepancies between what children state directly and what the test detects. With our work, we want to offer two contributions. The first, technical, describes both the logic as well as the tool we used to develop the IAT. The second, methodological, offers preliminary but exciting evidence to support the usefulness of implicit measures, and the IAT, in this field.