How should the Don't Know/Not Applicable option be handled in a Likert Scale?

Question

I'm currently evaluating a form design that uses Likert scales questions but it is possible that the user hasn't interacted with the product yet and may not know enough to answer the questions. The current form sets Don't Know/NA as the 0 score on the left end, 1 is equal to Not At All Important and 7 is Very Important, at the right end.

From my research, it looks like the best answer is to separate the Don't Know/NA response and place it to the side of the scale. Is that the best option for handling Don't Know/NA? Does anyone have examples of how they have handled this issue?

Answer 1

It's usually separated, and on the right

Reasons:

• It's separated because it's not part of the scale range. One of the benefits of the Likert Scale is its ability to visually lay out a gradient/range of values for a user (good-to-bad, mild-to-intense, etc). The "no answer" option is not part of this gradient, so it should be separated to preserve the benefit of the visual range.

• It's on the right because designers usually want users to give an answer. Users read left-to-right, so it's better workflow to present users with the values first. The user is more likely to answer than if you place the "don't know" option first.