How users interact with dark patterns: researching effects on time spent and gaze patterns with eye tracking

Abstract

Dark patterns (also known as deceptive design) are techniques that can be applied in digital user interfaces to steer the behaviour of the user into a certain direction that is beneficial for the owner of the website or app. A lot of research has been done over the last few years on this topic, but not much is yet known about how users look at instances of dark patterns nor how much time they cost or save users. Hence this research investigates user gazing behaviour using eye tracking equipment and simultaneously measures the time users spend looking at dark patterns and how this influences their total time spent on a task. The research consists of an experiment (N=13) in which people completed two tasks in a randomised order. Both tasks had an experimental version with dark patterns and a control version without dark patterns. Each participant saw one task in the control version and one task in the experimental version. Divided over the two tasks seven instances of dark patterns were implemented: sneak into basket, trick question, false hierarchy, preselection, nagging, low-stock/high-demand messages, and confirmshaming. All participants also completed a demographics questionnaire and took part in a retrospective think-aloud session. The results of the experiment showed that participants spent significantly more time on the checkout page in the experimental version of the first task, in which sneak into basket, trick question and false hierarchy were applied, compared to the control condition without dark patterns. For all of these three patterns it was also found – using Area of Interest analysis – that the participants spent more time looking at the deceptive version of the element, compared to the control version. For the other four patterns no such results were found. For the preselection pattern the opposite was found: participants spent on average less time looking at the options when one was already selected. This is probably caused by the fact that they were not required to perform an action in this area if they were satisfied with the selected option. In addition to time based metrics also the gaze patterns (the way the gaze of the participants moved over the screen) were analysed. For some types of dark patterns typical gaze patterns could be distinguished. For the trick question for example regressions (gaze jumping back to a previous word while reading) were seen often. This research contributes to the further understanding of how dark patterns are perceived by users. It can help designers to weigh the effects they want to achieve with the application of dark patterns against the adverse effects, such as extra time that is needed for the users. On a more theoretical side it offers extra insight in how users interact with and look at dark patterns on websites. This can also help in determining the severity of types of dark patterns, which in turn can help deciding what legislation is needed. Future research can consist of researching more types of patterns with a larger and a more diverse sample in order to be able to draw stronger conclusions.