Adaptive user interfaces modify their appearance based on a dynamic model of the user. They are designed to tailor a system's interactive behavior to the individual needs of users and altering conditions within the application environment.

This study examined the effects of interface adaptation on user performance in HCI and CMC in an adaptive testing environment. The effects of the software's monitoring of user performance, and adapting difficulty to user performance were the focus of this study. Two competing theories predicted possible effects of adaptive interfaces: 1) Social  facilitation, according to which users with high confidence for the task should perform better when monitored by the software, and users with low task confidence should perform less well; and 2) “choking”, according to which users with high task confidence should perform less well, and users with low task confidence should perform better because they know the interface will adapt to their performance.

A 2 (adaptive vs. non-adaptive) x 2 (high user task confidence vs. low task confidence) x 2 (HCI vs. CMC) laboratory experiment was conducted. Results indicate that for CMC, (human-human interaction), the social facilitation explanation holds true, while results for HCI were consistent with the “choking” explanation.