Triangulation within and across HCI disciplines
Wendy E. Mackay · 1999
How should we evaluate interactive software? Gray and Salzman (1998) examine five influential studies that compare methods of evaluating usability. In a carefully-argued critique, they identify numerous threats to validity and point out serious flaws in both the designs and claims made in each study. I think there will be various reactions to this article. Some people will react with alarm, concerned that if frequently-cited articles in respected publications are so full of errors, perhaps the field of human-computer interaction is in serious trouble. Others may attack the article, mis-interpreting it as the blind application of laboratory-style experiments to field studies. My own reaction is to commend Gray and Salzman for daring to critique some of the most influential work in the field, showing not only how to identify certain kinds of problems, but also how to address them. Their study helps to advance the field by raising the level of critical analysis in a graceful and constructive manner. This is how research disciplines make progress: through self-analysis and the creation of a solid research context. Field studies vs. laboratory experiments Like many others in HCI, I started out with academic degrees in experimental