Abstract
Frequently, two or three alternative interfaces will be empirically tested in a laboratory setting. Often, an argument arises concerning whether the performance or the preference of the user should be taken into account when making decisions about the best interface alternative. Usually, when making decisions regarding time-critical interfaces where mistakes can be costly, user performance outweighs user preference. Should user preference be disregarded at this point? This method attempts to combine users' preference and performance into one linear model. Using preference and performance data from an experiment that tested three alternative prototypes for monitoring air defense engagements, the results showed that one alternative performed better than the others, but was not preferred. After using the Brown-Gibson model for multi-attribute decision making, the resulting linear model showed that a different alternative that did not perform as well could actually be the superior prototype.