Human-Computer Interaction

Chapter 6

Hypothesis Testing

At a cocktail party, it is advisable to avoid discussions on religion or politics. The reason, of course, is that people’s opinions on these thorny topics are often both strong and divergent. Let’s add statistics to the list. Few topics raise more controversy than the application of statistical techniques to the data and hypotheses in experimental research. Fortunately, the reader is not dragged into the debate here.¹ Instead, the approach in this chapter is to describe and present a few useful statistical tools and techniques as they are typically used in HCI research. In the pages ahead I will work hard to minimize theory and controversy while maximizing practical value.

Statistical procedures for hypothesis testing ...