Assess technology

Observational methods for usability testing

Generally, the type of data you want to collect for usability testing are participants’ comments and behavior such as expressions, concentration level, preferences, actions while performing tasks (including errors), opinions about the technology, and sometimes activity counts (e.g., how many hits for a certain web page).  There are three types of observational methods that can collect usability data. The most effective usability tests use a combination of all three methods.    

Test monitoring

Test monitoring is the most common approach to observe and record test participants’ behavior and especially when there are relatively few participants.  A test monitor directly observes and records participant behavior and comments throughout the usability test.    

Usually, test monitors use a data sheet or database to help guide their recording.  These forms generally include test time, checklists of actions to be completed by participants, comments from participants (or from the test monitor), and narrative descriptions of the event. 

Direct recording

If you have many participants, want to eliminate any possible test monitor bias, or want a literal record of the testing, then directly recording the test is your best choice.  Some recording options include audio and video taping, screen recoding software, or the use of two-way mirrors.  Screen recording software such as Camtasia. may avoid participants feeling uncomfortable with a camera, listening device, or a mirror in their environment.  However, this software does not record participant behavior such as frustration or comments.


A think aloud is a valuable data collection method for usability testing.   In this approach, participants are asked (and reminded) to constantly verbalize their thoughts while performing the testing tasks.  This process allows the participants to freely express their feelings about a product and the difficulties they are encountering while using it. 

Page last updated: Sep 21 2011
Copyright © 2007, The University of Texas at Austin