I’ve been asked why I dislike surveymonkey surveys, I don’t want to reply fully in that forum (because it’s off-topic there) and I want to publish this for your feedback now.
surveymonkey really frustrates me. In short, it’s often confusing, I think it doesn’t follow UK accessiblity laws and I’m unsure if it complies with our privacy laws. It seems like another triumph of marketing over good sense. It’s a free market failure, possibly based on lack of information on the drawbacks of this option.
If a survey is on Surveymonkey, it’s easiest for able-bodied users of full-fat browsers with low security and privacy settings to answer. Using NoScript or a privacy tool? You’ll get an error message (which contradicts the VPAT_SurveyMonkey.pdf description). Unable to see graphics clearly? The forms will be hard to read because they’ve replaced the browser form widgets (such as checkboxes and radio buttons) with javascript-driven images for no good reason.
Those sort of faults will probably bias results a bit – how much will depend on the subject. It’s a particular case of a problem I first mentioned in the last-but-one bullet point on my debian surveys page back in 2006. Basically, I think you should keep open surveys as open as possible. Statistics literature is littered with bizarre spurious results from when the survey method accidentally favours some particular subset of the population.
Finally, surveymonkey surveys can be left with answer requirements which make no sense, such as ticking “Other” requiring a text input, but the text box not accepting input. Is that the fault of the survey owner not being able to work the admin interface, or the interface having some subtle flaws? I’m not completely sure, but I’ve had that experience a few times.
So, can I persuade people to use limeservice, doodle or even Google Docs rather than Surveymonkey, please?
I think Limeservice is better because it’s based in Germany (so EU data protection and privacy laws apply) and is cooperatively-developed FOSS software. Doodle and Google aren’t, but are quicker and easier to use for straw polls.
If you want something to install on your own site and the downloadable limesurvey software seems a bit heavy, then askpeople seems another good option. Our co-op can host either of those if you’re willing to pay. (If someone would be willing to sponsor it, we’d host some free surveys too – let me know.)
Lots of people use surveymonkey, but popularity doesn’t always mean it’s the best tool. Please consider the alternatives and pick the best tool for the job.
36 Responses to Get the Survey Monkey Off Your Back