Talking About Thinking Aloud: Perspectives from Interactive Think- Aloud Practitioners

It is widely reported in the literature that intervening during usability testing sessions affects user behavior and compromises the validity of the test. However, this contrasts with the ongoing popularity of Interactive Think-Aloud (ITA) amongst practitioners. We report an in-depth qualitative study that explored this tension between theory and practice through nine interviews with ITA practitioners. Our findings add nuance to many established ideas about ITA but also reveal novel practices and attitudes. For example, ITA is sometimes used to slow down users as they navigate through a system, to manage external pressures such as recruitment difficulties, and to reframe a session as a kind of interview or participatory study. We also found that participants (ITA practitioners) experienced unexpected difficulties with ITA, including the risk that it results in overly reflective think-aloud and creates challenges in team working. Participants understood that ITA causes reactivity, and they reported taking steps to reduce it. However, overall, they did not see the traditional positivist objective of valid problem discovery as a realistic or high-priority goal for usability testing. They believed that ITA data can be useful and valid even if user behavior is not wholly realistic. Based on this, we argue against the narrow problem-counting approach often employed in the comparative usability evaluation studies that have sometimes seemed to discredit ITA. We also make the case for broadening how we think about the validity of usability testing data, and we argue that forms of ITA may be appropriate in some situations.