Readings
Kuniavsky, ch. 11
There are lots of resources (online and published) on survey research generally. Here are some suggestions. Most emphasize some aspects more than others — at least skim all of them, so that you can pick up from each what it covers more thoroughly than the others do:
Research Methods Knowledge Base: Survey Research. You’ll see lots of similarities to interviews, but some important differences.
University of Reading Statistical Services Center, Guidelines for Planning Effective Surveys. This one has less on question construction than the above but more on other surveying issues.
Statnotes, Survey Research. Especially good on design considerations and item bias. Also good FAQs.
Questionnaires in Usability Engineering: FAQs (3rd ed)
Added 10/13:
The Pew Internet and American Life project is in many ways the gold standard of doing internet-related survey research — not online but via phone, where the sample is more controlled and they can reach non-users. I don’t find a good discussion of their methods for this project, but this web page describes their methods for another project, The Pew Research Center for the People & the Press.
The kinds of surveys done in UX are nowhere near this elaborate, but are in some ways low-effort approximations of the more elaborate and statistically-valid surveys conducted by organizations like Pew.
Look in particular at this page on questionnaire design and this page on survey administration.
Analyzing survey results
University of Reading Statistical Services Center, Approaches to the Analysis of Survey Data
Optional and Useful:
Free Survey Software
There are many. Be sure you check their limits on free or trial use — not just # of questions and respondents, but what sort of data analysis they do, and/or whether you can download data to Excel.
surveymonkey.com
zoomerang.com
polldaddy.com
Questionnaires: Sources of questions and examples of questions (and data analysis)
Large-scale surveys from reputable sources can be great sources of question wording, when they ask questions you need to ask. Presumably such a source will have done an adequate job of developing and testing questions. A survey of a population similar to yours can be a particularly source of questions on demographics.
Even when they don’t ask questions useful to you, looking at good models can help your own survey development.
Their reports may be even more helpful — showing good ways to summarize and present survey data.
Pew Internet and American Life project publishes their surveys. If they have written a question that works for you, use it! Also, by seeing how carefully they word questions and answers you get a sense of the kinds of problems in question wording that they had to deal with (which may help you think about your questions).
A recent example of their work – Online News Survey
UC Berkeley survey of new undergrads – unfortunately, doesn’t address technological capabilities and experience.
Smithsonian -ForeSee Survey — You’ll see this company’s slight-customized surveys on a lot of websites.
Very simply website usability survey from Underage Drinking Enforcement Training Center – as an example
Reports
UC Berkeley survey of new undergrads The reports presented here are minimal. They summarize findings for entire respondent pool; don’t break down by subgroups. Don’t compare years.
Pew Internet and American Life project Their reports tend to be for a popular audience, so don’t always have the detail needed, but they’re good at summarizing key findings.
Filed by Nancy Van House at 7:46 pm under Diary Studies