Egon Brunswik coined and defined the concepts of ecological validity and representative design, which are both essential to achieve external validity. However, research in HCI has inconsistently and incorrectly used Brunswik’s concept of ecological validity, which prevents the field from developing cumulative science and from generalizing the findings of user experience (UX) evaluations. In this paper, I present ECOVAL, a framework I built on Brunswik’s ideas. On the one hand, ECOVAL helps HCI researchers describe and assess the ecological validity of cues in UX evaluations. On the other hand, ECOVAL guidelines—formulated as a step-by-step procedure—help HCI researchers achieve representative design and, therefore, increase external validity. An industrial case study demonstrates the relevance of ECOVAL for achieving representative design while conducting formative UX testing. In discussing the case study, I describe how ECOVAL can help HCI researchers assess and increase the validity of UX experiments and generalize UX findings. I also illustrate the trade-offs between internal and external validities and UX resources that inevitably arise when one conducts UX experiments. From the results, I sketch avenues for future research and discuss the related challenges that future work should address.
ECOVAL: Ecological Validity of Cues and Representative Design in User Experience Evaluations.
AIS Transactions on Human-Computer Interaction, 9(2), 149-172.
Retrieved from https://aisel.aisnet.org/thci/vol9/iss2/4