Have you ever seen a question where you select an answer from a set list like Strongly disagree, Disagree, Neither agree nor disagree, Agree, and Strongly agree? that’s a Likert scale.

Crafting a survey like this is very straightforward and easy, but doing it well is very difficult! For example, this style of survey is lacking the granularity necessary to capture answers like, “I can see both sides” or “Not applicable” or “This question warrants a richer discussion and should not be shoe-horned into a simple yes or no question.” Let’s forget about the problems and instead focus on appropriate usage of the Likert scale.

Think about the last questionnaire like this that you completed. What was your average response? I’ll guess your responses were entirely in one column (more precisely, your standard deviation between responses was low). If you were asked about a crappy restaurant experience you might answer all negatively: that the food was bad, the service was bad, the decor was bad, and the price was bad. If you were asked about your feelings about something you didn’t care about you would probably hover around the middle choice (Neither agree nor disagree in our example above).

Personally, I think that presenting an odd number of options is folly. In my experience, people choose the middle option most often. The middle is safe and easy. But that’s not informative! Questions should not be simplified into Neither agree nor disagree. Everyone feels more strongly one way or another.

The next time you respond to a survey that uses a Likert scale, do this:

  • remember the term “Likert scale” and applaud yourself for exercising your knowledge
  • avoid the simple middle answer–think a second and decide if you really feel a bit more to one side
  • take note of how positive or negative you are