The Complete Guide to Likert Scales in Survey Design

The Likert scale is one of the most widely used tools in survey research, and one of the most often misused. Used well, it captures attitudes and opinions with useful precision. Used carelessly, it produces numbers that get analyzed the wrong way. This guide explains what Likert scales are, how to design them, and how to analyze the results honestly.

What Is a Likert Scale?

A Likert scale measures the degree to which someone agrees with a statement, or how strongly they feel along a dimension. Instead of a plain yes or no, respondents choose a point along an ordered range, typically from one extreme to the other.

A classic five-point agreement item looks like this:

• Strongly disagree
• Disagree
• Neither agree nor disagree
• Agree
• Strongly agree

Each statement that uses this format is technically a Likert item. When several related items are combined into a single score, that combined measure is a Likert scale. The scale need not measure agreement; it can capture satisfaction, frequency, importance, or likelihood, as long as the points form a clear order.

How Many Points: 5 vs 7?

The number of points is a real design decision, and both five and seven are common, defensible choices.

Five-point scales

• Simple and fast to read, with low cognitive load.
• Work well on mobile, where space is tight.
• Easy for respondents to map their feeling onto a clear option.

Seven-point scales

• Offer more granularity, letting people express finer distinctions.
• Can be more sensitive to small differences in attitude.
• Take slightly more effort and screen space to present clearly.

As a practical rule, use five points for general-audience or mobile-heavy surveys, and consider seven points when you need finer resolution and your respondents are engaged. Whatever you choose, keep the number of points consistent across similar items so answers stay comparable.

Labeling the Points

Good labels are what make a scale trustworthy. A few guidelines:

• Label every point, not just the ends. Fully labeled scales are interpreted more consistently than scales that only name the extremes.
• Keep the labels balanced. The distance in meaning from "Agree" to "Strongly agree" should feel similar to the distance from "Disagree" to "Strongly disagree."
• Match the wording to the question. An importance question needs importance labels ("Not at all important" to "Extremely important"), not agreement labels.
• Order the points consistently and lay them out left to right (or right to left for RTL languages) in their natural sequence.

Should You Include a Neutral Midpoint?

Whether to offer a middle option like "Neither agree nor disagree" is a genuine trade-off.

Reasons to include it

• Some people truly are neutral, and forcing a side misrepresents them.
• Without a valid neutral choice, undecided respondents may pick at random or abandon the question.

Reasons to consider leaving it out

• A midpoint can become a lazy default for people who do not want to think.
• If your goal is to push for a direction, an even-numbered scale removes the fence to sit on.

A reasonable default is to include a clearly labeled neutral midpoint, and to add an explicit "Don't know" or "Not applicable" option when genuine non-opinions are likely, so true neutrality is not confused with lack of knowledge.

How to Analyze Likert Data

This is where many surveys go wrong. Likert responses are ordinal data: the categories have a clear order, but the gaps between them are not guaranteed to be equal. The step from "Disagree" to "Neutral" may not represent the same change as the step from "Agree" to "Strongly agree."

Reporting a single item

For a single Likert item, the safest summaries respect its ordinal nature:

• Report the full distribution. Show the percentage choosing each option; this is often the most honest and readable view.
• Use the median or mode as the central tendency, since they suit ordinal data.
• Be cautious with the mean. Averaging ordinal codes assumes equal spacing between points, which may not hold. If you report a mean, treat it as a rough indicator and show the distribution alongside it.

Working with a full scale

When you combine several items into one Likert scale and the items are well constructed, the summed or averaged score behaves more like continuous data, and means become more defensible. Even then, report how the scale was built and check that its items hang together.

A few good habits

• Visualize distributions with stacked or diverging bar charts so the balance of opinion is obvious.
• Keep scale direction consistent, and if you reverse-word some items, recode them before combining.
• Avoid over-precise claims; an average of 3.9 versus 4.0 rarely means much on its own.

Conclusion

A Likert scale is simple to answer but easy to misuse. Choose a sensible number of points, label them clearly and in balance, decide deliberately about the neutral midpoint, and analyze the results as the ordinal data they are. Do that, and your scales will measure attitudes faithfully.

In MSN Forms you can build five- or seven-point Likert questions in minutes, then view live distributions and charts that respect the ordinal nature of the data, making it easy to report results the right way.