Design of experiments

VR disrupts the ordinary perceptual processes of its users. It should be clear from this section that proposed VR systems and experiences need to be evaluated on users to understand whether they are yielding the desired effect while also avoiding unwanted side effects. This amounts to applying the scientific method to make observations, formulate hypotheses, and design experiments that determine their validity. When human subjects are involved, this becomes extremely challenging. How many subjects are enough? What happens if they adapt to the experiment? How does their prior world experience affect the experiment? What if they are slightly sick the day that they try the experiment? What did they eat for breakfast? The answers to these questions could dramatically affect the outcome.

It gets worse. Suppose they already know your hypothesis going into the experiment. This will most likely bias their responses. Also, what will the data from the experiment look like? Will you ask them to fill out a questionnaire, or will you make inferences about their experience from measured data such as head motions, heart rate, and skin conductance? These choices are also critical. See Section 12.4 for more on this topic.

Steven M LaValle 2020-01-06