INSPECTION-BASED EVALUATIONS
Gilbert Cockton, Alan Woolrych, Darryn Lavery · 2007
Conclusion Validity . . . . . . . . . . . . . . . . . . . . . . . . . 1180 Improving UIM Assessment and Comparison . . . . . . 1180 Analyst Training and Preparation . . . . . . . . . . . . . . 1180 Reliability of Prediction Merging . . . . . . . . . . . . . . . 1180 Ability of User Testing to Expose Actual Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1181 Reliability of Known Problem Extraction . . . . . . . . 1181 Reliability of Matching Predicted to Actual Problems and Structured Report Formats . . . . . . . 1181 Controlling Confounding Variables . . . . . . . . . . . . 1181 Outstanding Methodological Challenges . . . . . . . . 1182 Understanding Inspection Methods . . . . . . . . . . . . . 1182 The DARe Model of UIM Effectiveness . . . . . . . . . . . . 1182 The DARe Model and the Scoping of UIMs . . . . . . . . 1182 The DARe Model and the Development Lifecycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1184 Improving Inspection Methods . . . . . . . . . . . . . . . . . 1184 Method-Independent Improvements . . . . . . . . . . . . . 1185 Method Specific Improvements . . . . . . . . . . . . . . . . . 1185 Improving Analysts . . . . . . . . . . . . . . . . . . . . . . . . . . . 1186 Future Trends and Best Practice . . . . . . . . . . . . . . . . 1187 Prospects for UIMs . . . . . . . . . . . . . . . . . . . . . . . . . . . 1187 Best Practice for UIMs . . . . . . . . . . . . . . . . . . . . . . . . 1187 Note . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1187 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1188 Usability inspection methods (UIMs) remain an important discount method for usability evaluation. They can be applied to any designed artifact during development: a paper prototype, a storyboard, a working prototype (e.g., in Macromedia Flash™ or in Microsoft PowerPoint™), tested production software, or an installed public release. They are analytical evaluation methods, which involve no typical end users, unlike empirical methods such as user testing. UIMs only require availability of a designed artifact and trained analysts. Thus, evaluation is possible with low resources (hence discount methods). Although risks arise from low resources, well-informed practices disproportionately improve analyst performance, improving cost-benefit ratios. This chapter introduces UIMs, covering six and one further method, and provides approaches to assessing existing, emerging, and future UIMs and their effective uses.