Internet Based Research
banner

WebLab UX

WebLabUX is a software package that allows researchers and Web site stakeholders to measure user behavior and performance on an instrumented Web site as well as test various designs. WebLabUX allows Web site designers and instructional designers to modify a pre-existing Web site or XHTML/CSS-based instructional help system by: (1) tagging dependent variables of interest (e.g., pages visited, links clicked); (2) instantiating independent variables (e.g., variations of the content design, the navigation design, etc.); (3) composing and inserting survey instruments; and (4) setting up study protocols (e.g., when surveys appear, method for assigning readers to conditions).

After WebLabUX readies the study materials, users can visit the study Web site from a time and place that is natural to them, and perform tasks on the Web site in a natural context.As each user visits the study Web site, WebLabUX can: (1) assign visitors to conditions based on the study protocol; (2) automatically generate design variations (conditions) from a single set of source Web pages (for XHTML/CSS or PHP-based Web sites only); (3) deliver surveys and record results; (4) track user's navigational behavior (e.g., specific links or buttons clicked, page requested, referring page, time on page); (5) link survey results to navigational behavior for each visitor; (6) support longitudinal studies requiring repeated visits to or across study-sections of a Web site; (7) create data files that can be easily exported into Excel, SPSS, R, or other popular packages.

WebLabUX accomplishes the activities listed above in an unobtrusive manner by using server-side scripting and instrumented Web pages. The developer of a Web site instruments the pages by adding tags to the Web pages. Because WebLabUX does not require users in a study to download or install any special software on their computers (as is the case with some commercially available tools), the mechanisms that handle the remote usability study are invisible to the Web site’s users. WebLabUX also includes a set of post-processing tools that detect and flag unexpected user behavior or behaviors that violate study protocols. These scripts allow the Web design and evaluation team to determine, for instance, if users have opened multiple Web browser windows to refer back to specific pages while taking surveys and comprehension tests.

WebLab UX supports the design team in understanding the users of their Web site. We know that some may question the feasibility of remotely studying Web site users through Internet, but our previous work proves the success of the Internet-based approach. We maintain that the involvement of actual users of a Web site in their own environments brings in all of the complexities of the natural world in which users function (e.g., differing computing platforms, connection speeds, and browsers; and varying types of participants with different levels of motivation, interest, topic familiarity, and computer expertise) and that this is a boon to the development of user-appropriate Web sites. Internet-based Web design research can help user studies achieve considerably greater external and ecological validity than studies conducted within the confines of a laboratory.
Contact: Professor Jan H. Spyridakis
Department of Technical Communication
Last updated: February 1, 2008