13.1  If there is something there, make sure you find it

13.2  The noise–effect–number triangle

13.2.1  General strategies

13.3  Subjects

13.3.1  More subjects or trials (increase number)

13.3.2  within-subjects/within-groups studies (reduce noise)

13.3.3  Matched users (reduce noise)

13.3.4  Targeted user group (increase effect)

13.4  tasks

13.4.1  Distractor tasks (increase effect)

13.4.2  Targeted tasks (increase effect)

13.4.3  Demonic interventions! (increase effect)

13.4.4  Restricted tasks (reduce noise)

A–B test, alternative hypothesis, balance orders, Bayesian statistics, between-subjects design, cognitive load, control conditions, controlled environments, controlled experiment, controlled tasks, coping mechanisms, demonic intervention, distractor task, ecological validity, effect size, elderly users, error bars, expert slip, false baseline, false negative, false positive, Fitts' Law, five users (the myth), formative evaluation, HCI studies, Human–Computer Interaction, hypothesis testing, image recognition, in-the-wild, increase effect, increase number, individual differences, interference effects, iterative development, learning effects, matching subjects, mathematics, mean (μ), menu navigation, more subjects, Müller–Lyer illusion, narrow user group, noise–effect–number triangle, non-significant result, nonlinearity, null hypothesis, order effects, pair subjects, parameters, power-law distribution, random noise, randomise orders, real world, reduce noise, restricted device, restricted system, restricted tasks, sample, sensitivity, size of the noise, size of the study, social network data, sonic enhancement, standard deviation (s.d., σ), standard error (s.e.), statistical analysis, statistical power, statistical significance, steering task, subjects, summative evaluation, targeted tasks, targeted user group, task completion time, tasks, the job of statistics, uncontrolled factors, uncontrolled variation, user experience, user experience studies, user studies, variability, WEIRD people, within-groups studies, within-subjects experiment, Wizard of Oz prototyping