Summary

This chapter has presented the key issues surrounding the development of a custom Web browser for conducting Web-based experiments. By using this client-side (vis. server-side) approach, to administering treatments and collecting data, the researcher is afforded a high degree of experimental control and precision. The behaviors of the experimental subject can be analyzed and recorded at the level of the mouse click or individual key press. Treatments and measurements timed to the millisecond are easily attained. Flexibility is provided by the fact the software can be customized to meet the specific needs of the research context. The software can be programmed to analyze navigation requests and, in the case of inapplicable requests, to preempt or modify the Web action before the navigation takes place. This allows for experiments to involve the general Web, rather than being restricted to a limited set of experiment-specific pages. Thus, external validity can be greatly increased.

The chapter presented many of the details surrounding the use of the WebBrowser Control, and described the specifics of programming custom handlers for several of the key WebBrowser events. Specifically, event handlers having to do with (1) navigation and downloading activities, (2) emulation of browser interface behavior, and (3) window and session control were discussed.

Beyond these topics, the chapter presented a discussion of some other special software techniques that are pertinent to this research context. Many of the requirements of a custom keyboard handler for this application were detailed, as were those of a custom mouse handling routine. Programmatic control over browser caching behavior was also discussed. Finally, it was illustrated how the custom browser tool can gain access to detailed elements of the HTML page by processing the DOM structure. This technique has the potential to add even more power, flexibility and control to your custom software tool.