The primary motive behind sample size estimation, as illustrated in Figure 6.1, is economics.

the sample of users you measure represents the population

Things like gender, operating system, and usability problem type are categorical variables.

Confidence intervals are used just like margins of errors. In fact, a confidence interval is twice the margin of error.

To find the geometric mean, transform the raw times to log times (using the Excel function =LN()), find the arithmetic mean of these log times, then convert this mean of the logs back into the original scale (using the Excel function =EXP()).

For a comprehensive discussion of usability testing, see the chapter “Usability Testing” in the Handbook of Human Factors and Ergonomics (Lewis, 2012

Usability testing remains a central way of determining whether users are accomplishing their goals.

we also need a measure of the spread of the data around the average. The most common way to do this is using a metric called the standard deviation.

The larger our sample size, the smaller we would expect the standard error to be and the less we’d expect our sample mean to differ from the population mean.