Bad Web Survey Question of the Day
This morning I was asked to do a survey on backups by a friend of a friend who is doing an MBA project. An interesting topic and one that is close to my heart as a small business owner who has had two issues in the past four years that I only survived because I run regular backups.
One of the basic principals of building Web survey questions is that the options for answering have to cover ALL POSSIBLE responses – not just the responses you think are most likely.
So with that background, here’s the bad survey question of the day:
Q: How long can you survive without the data and applications on your computer?
- A few hours
- A day
- A few days
- More than a few days
- Forever
My answer to this question is: I CAN’T SURVIVE!!!
Seriously, in my business I need my data and applications every moment. But I didn’t have an option to present that opinion, and I probably answered all the other questions in a different way because this is so critical to me – which the surveyers won’t be able to segment.
This question could have been easily fixed by asking “How else could someone answer this question?” With the one exception that I had, it’s a pretty complete list. So maybe they could have realised they were missing my answer if they had just taken a moment to think about completeness.
In general, adding an “Other” category allows you to quickly learn that you have missed something – a great best practice because even the best researcher doesn’t know all the options.
[...] There are many ways to make sure your online survey is efficient and effective. One of them, of course, is to avoid asking bad questions. [...]
Amazing how often we see these types of errors. In another recent example, I saw a survey that used totally out of date household types (marital/children status). They may have been realistic for 99% of the population 20 years ago, but not today!
Great example.
I find some other issues with the metrics the question provides. Technically speaking, it would be really difficult to manage this scale if you want some parametrics. I mean, what´s the gap between a few days and forever, for example? or even funnier, what´s the (semantic) difference btw “more than a few days” and “forever”??
That´s cool, I like this examples, and even though you can manage to do something with this answer the idea is to do the best analytical work you can, no?