Conducting surveys via the Web is popular for some compelling reasons: they
are low-cost, quick to deploy, and results are instantly tabulated. But, there
is a risk (as the saying goes) that “when you have a hammer, everything
looks like a nail.” Not all research objectives are suited to a Web-based
survey. And poorly designed surveys can yield misleading results, leading to
erroneous business and/or design decisions.
If you are considering a Web-based survey, I would recommend reading Doing
Survey Research: A Guide to Quantitative Research Methods by Peter
M. Nardi. It’s highly readable and packed with great real-world research
examples. It covers not only Web- and paper-based surveys, but also those conducted
face-to-face and via telephone. It also puts surveys in context of other research
methods, with Nardi emphasizing the importance of choosing the “right
tool for the right job.”
Update:
Since writing this post, two articles about online surveys have been published that are worth a read:
First is Jakob Nielsen’s Alertbox column “Keep
Online Surveys Short,” which harkens back to the late 1990’s
heyday of Alertbox in terms of focused, practical advice (in this case, on
how to avoid the pitfall of survey bloat).
Offering more in depth advise is Human Factor International’s “Creating
Effective Online Surveys” (subtitled Owning Photoshop Doesn’t
Make You an Artist). It draws primarily upon a
case study by Andrews, Nonnecke and Preece (2003), which emphasizes the
importance of carefully piloting the entire survey process (including data
analysis) before deployment.