Skip to content

Likert scale

Also called rating scale, agreement scale

A set of ordered answer options, usually four to seven, running between two poles or up from none of something.

Rensis Likert introduced the format in 1932: statements rated on ordered responses, commonly from strongly disagree to strongly agree. A single question is a Likert item; a Likert scale combines several related items.

Bipolar scales run from one pole through a middle to its opposite. Unipolar scales climb from none of a thing to a lot of it, and must not be given a “neutral” midpoint, because there is no neutral amount of “often”. The labels are the instrument: rungs that do not read as roughly even steps make the numbers behave oddly no matter how you analyse them.

An odd number of points offers a middle; an even number forces a side. Either can be right. What is not optional is pairing the scale with somewhere for people the question does not apply to, and keeping that option out of your averages.

Where it goes wrong

Reversing the direction of one item in a battery to catch inattention. It also makes that item behave differently in the analysis, so decide which you care about more.

See also

Ordinal data
Answers with a known order but unknown spacing, which is what almost every rating question produces.
Matrix question
Several statements sharing one set of answer columns, shown as a grid.
Straight-lining
Selecting the same answer across every row of a question battery.

All 35 terms

Start

Put it into practice in your next survey.