Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.574446
Title: 'Please keep talking': an investigation of the concurrent think-aloud method in usability testing
Author: Zhao, Tingting
Awarding Body: University of Sunderland
Current Institution: University of Sunderland
Date of Award: 2012
Availability of Full Text:
Access from EThOS:
Abstract:
This thesis explores the reported disparity between the concurrent think-aloud theory (Ericsson and Simon, 1984; 1993) and its application in usability practice. A tension exists between the classic approach and usability practitioners' desire to gather relevant data. This research is the first to specifically examine the impact of a number of methodological changes in the concurrent think-aloud method, on both the utility and validity of usability test data. Three empirical studies were conducted. Study One examined the impact of adding evaluator interventions to the classic think-aloud to make it an interactive think-aloud. The results showed that there were no differences between the types of utterances produced in the classic and interactive think- aloud styles, although the latter produced more utterances relevant to usability analysis, and more verbalised usability problems. However, the interactive think-aloud also produced a large number of low impact problems. The usefulness of interventions was low, and the types of interventions issued could affect the nature of the verbalisations elicited, their relevance and the quality of problems. Study Two examined the impact of using an explicit instruction that requests specific contents in the think-aloud. The results showed that the explicit instruction neither improved the utility of data nor facilitated the usability analysis process, compared to the classic instruction. However, it served to increase participants' mental workload and led users to be more critical about the system. The third study considered the use of dual elicitation by gathering both concurrent and retrospective think-aloud data. The results showed that by adding a minimal amount of time, this approach increased the explanatory power of the data, and facilitated problem discovery and understanding. These improvements were achieved without compromising the test's validity. The empirical work reveals that methodological changes that are divergent from their theoretical motivations may not work as effectively as practitioners anticipated. The risks of introducing invalid data are not worth the limited benefits obtained. The research suggests that researchers and practitioners should adhere to the classic think-aloud, but reinforce it by collecting additional retrospective data. This dual elicitation technique appears to be a more rigorous and useful usability evaluation tool.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.574446  DOI: Not available
Share: