Appendix: A Crash Course in Fundamental
Statistical Concepts
INTRODUCTION
Throughout the book we’ve attempted to provide as much statistical background as possible without letting it get too overwhelming. In this appendix we review some fundamental statistical concepts and provide pointers to chapters where the concepts are covered in greater detail. If you’ve never had an introductory statistics class or don’t remember basic concepts such as measuring central tendency and variability, then you can use this appendix for a quick review.

TYPES OF DATA
The first step in using statistics to make better decisions is to obtain measurements. There are two major types of measurements: quantitative and categorical. Task time, number of usability problems, and rating-scale data are quantitative. Things like gender, operating system, and usability problem type are categorical variables.
Quantitative data fall on a spectrum from continuous to discrete-binary, as shown in Figure A.1.
Note that the extreme discrete end of this spectrum includes binary categorical measurements such as pass/fail and yes/no. The more discrete the data, the larger the required sample size for the same level of precision as a continuous measure. Also, you’ll usually use different statistical tests for continuous versus discrete data (see Chapters 3–6).
Discrete data have finite values, or buckets. You can count them. Continuous data technically have an infinite number of steps, which form a continuum. The number of usability problems would be discrete—there are a finite and countable number of observed usability problems. Time to complete a task is continuous since it could take any value from 0 to infinity, for example,
178.8977687 seconds.
You can tell the difference between discrete and continuous data because discrete data can usually be preceded by the phrase “number of