Putting the fun into usability

We’ve all felt like punching a computer or laughed at something on a website. We might have felt close to others we’ve connected with online, or gone surfing around our hobbies for pleasure. Console games have spiked our hearts with terror when zombies attack, and relief as the level boss is defeated.

Computers aren’t about work, work, work any more. Technology engages us emotionally, yet few usability studies take this into account. At a time when fun is fundamental to the perception of an interface, how can you measure it?

The most common approach is to ask users how they feel about an activity, using a questionnaire after the test or by scoring feeling ratings in an interview or focus group. This method depends on users being able to recall and articulate their feelings, though. One idea to increase the accuracy of recall is to show the user a video replay of the test on the assumption it will immerse them in activity again, and bring their feelings bubbling back. Any interruption in the activity to prompt users to reveal their feelings during it would likely change how they feel and negatively affect the experience the user has.

Another common approach is to observe participants during a think-aloud commentary to assess their spoken and body language. Cues of emotion, such as facial and verbal expressions and gestures, can be coded. The analysis needs to be conducted with great care to avoid bias because it involves so much interpretation.

Researchers are starting to use biometric tools to measure physiological changes. One example is using sensors on face muscles to detect the minute changes in electrical activity that accompany facial expressions. Biometric tools facilitate continuous data collection and are relatively non-intrusive, but the complex data output may not be well understood. Measures such as heart rate and galvanic skin response only measure stimulation and have difficulty distinguishing between, for example, joy and anger. It’s also impossible to understand what is triggering a response at any given time.

Indirect methods have also been proposed, including asking participants to estimate how long an activity took. Difficult tasks tend to be overestimated and easy ones are underestimated, but more research is needed before this metric can be used as any proxy for how the user felt during the task.

There is no easy way to measure a user’s emotional reaction to an interface. Currently, the most promising approach is subjective self-reporting. Admittedly, data can only be collected when a question is asked of the user, and the responses may not fully correspond with the actual experience, as they can be skewed by the novelty of the activity or a desire to please the investigator. In any case, emotional experiences can be hard to explain. That said, tools such as post-study questionnaires are convenient, cheap, generalisable and amenable to rapid statistical analysis. Interviews are similarly convenient and can provide a richness of detail that can be absent from quantitative approaches.