Some Whys and Hows of Experiments in Human-Computer Interaction

摘要

Experiments help to understand human-computer interaction and to characterize the value of user interfaces. Yet, few intermediate guidelines exist on how to design, run, and report experiments. The present monograph presents such guidelines. We briefly argue why experiments are invaluable for advancing human-computer interaction beyond technical innovation. We then identify heuristics of doing good experiments, including how to build on existing work in devising hypotheses and selecting measures; how to craft challenging comparisons, rather than biased win-lose setups; how to design experiments so as to rule out alternative explanations; how to provide evidence for conclusions; and how to narrate findings. These heuristics are exemplified by excellent experiments in human-computer interaction.