I'm wondering about the perennial problem that arises when presenting multiple options, where the user can select none, one, or several options.
Using checkboxes, it's not clear whether a user
- intentionally selected none of the options
- or just ignored the question.
Like so:
download bmml source – Wireframes created with Balsamiq Mockups
See. I don't know whether you don't like any fruit, or whether you just don't want to tell me.
So if I need to know whether the user answered the question, then it makes sense to include a "none of the above" option at the end of the checkbox list. However, there's not a clearly good way to include a "select none" option in a list of checkboxes.
With a small number of options, radio buttons may be a suitable workaround....
...but with a larger number of options, radio buttons quickly become untenable:
So I was quite excited today to stumble across what might be a solution to the problem—a hybrid checkbox/radio control:
When the radio button is selected, the checkboxes become deselected, and when the checkboxes are selected, the radio becomes deselected.
At first glance, this looks like a great solution...but before I start adopting it, I want to check if it may confuse people. Perhaps they view the separate shapes as corresponding with separate questions; or perhaps they don't understand why clicking on some things (but not just any thing) may cause others to become deselected; or perhaps they don't distinguish between checkboxes and radio buttons at all and they're slowed down by wondering why one of the buttons is a different shape from the others.
Has user testing found that this control is commonly understood? Are there particular demographics or psychographics that find this hybrid control problematic?