Learn how to choose the appropriate number of points and labels for Likert scale questions in market research surveys, and what factors and best practices to consider.

Whether you use odd or even number of points depends on how strongly you belief that people have an opinion on something. There is an advantage to getting a "real" non-neutral read and push people to one or the other side of the scale especially if neutral has a double meaning - does it mean I am neutral on something because I didn't experience something or because the experience wasn't particularly good but also not particularly bad. Everyone has an opinion but sometimes the question is whether someone is able to answer that question - i.e. if they didn't experience let's say customer support - rather than offering a neutral option I would prefer to add a Don't know option to exclude these people from the base so my readings are more pure.

How do you choose the appropriate number of points and labels for Likert scale questions?

Likert scale questions are a common way to measure attitudes, opinions, and preferences in market research surveys. They ask respondents to indicate how much they agree or disagree with a statement, using a scale of points and labels. But how do you choose the appropriate number of points and labels for your Likert scale questions? In this article, we will explain some factors to consider and some best practices to follow.

1 Factor 1: Reliability and validity

The first factor to consider is the reliability and validity of your Likert scale questions. Reliability refers to how consistent and accurate your questions are, while validity refers to how well they measure what they are supposed to measure. Generally, the more points and labels you use, the more reliable and valid your questions will be, because you can capture more nuances and variations in responses. However, there is a trade-off between reliability and validity and other factors, such as response rate and ease of analysis.

Add your perspective

2 Factor 2: Response rate and ease of analysis

The second factor to consider is the response rate and ease of analysis of your Likert scale questions. Response rate refers to how many people complete your survey, while ease of analysis refers to how simple and straightforward it is to interpret and report your results. Generally, the fewer points and labels you use, the higher your response rate and ease of analysis will be, because you can reduce the cognitive load and confusion for respondents and simplify the statistical methods and presentation for analysts. However, there is a trade-off between response rate and ease of analysis and other factors, such as reliability and validity.

Add your perspective

3 Factor 3: Context and purpose

The third factor to consider is the context and purpose of your Likert scale questions. Context refers to the topic, audience, and format of your survey, while purpose refers to the goal, objective, and hypothesis of your research. Depending on your context and purpose, you may want to use different numbers and labels for your Likert scale questions. For example, if you are asking about sensitive or controversial issues, you may want to use more points and labels to allow for more neutral or ambivalent responses. If you are asking about simple or straightforward issues, you may want to use fewer points and labels to elicit more clear and decisive responses.

Add your perspective

Simone Holligan, Ph.D.

Researcher | Lecturer
(edited)
Report contribution
As a market researcher one must also be cognizant that skew does and will occur in certain scenarios, and that these data are acceptable for analysis and interpretation. For surveys based on ratings/reviews of a product, service, or experience – say for a local restaurant, an Airbnb stay, an Uber ride, or an employer on Indeed or Workopolis – responses will be skewed in either direction. Those who freely respond either absolutely hate or absolutely love the product, service, or experience you provide. Customers who received or experienced what they had expected all along, (i.e., your product, service, or experience was neither terrible nor outstanding), tend to not provide a rating unless they are driven by an incentive.

Like

4 Best practice 1: Use an odd number of points

One best practice to follow when choosing the number of points for your Likert scale questions is to use an odd number of points, usually between 5 and 7. This allows you to have a midpoint or a neutral option, which can increase the validity and reliability of your questions, as well as the response rate and ease of analysis. A midpoint can also reduce the bias and error that may result from forcing respondents to choose a side or skip the question.

Add your perspective

Rogier Verhulst

Sr. Director at LinkedIn | Market Research Expert
Report contribution
Whether you use odd or even number of points depends on how strongly you belief that people have an opinion on something. There is an advantage to getting a "real" non-neutral read and push people to one or the other side of the scale especially if neutral has a double meaning - does it mean I am neutral on something because I didn't experience something or because the experience wasn't particularly good but also not particularly bad. Everyone has an opinion but sometimes the question is whether someone is able to answer that question - i.e. if they didn't experience let's say customer support - rather than offering a neutral option I would prefer to add a Don't know option to exclude these people from the base so my readings are more pure.

Like
Rob Podhurst

President at Podhurst Associates Marketing Research
Report contribution
An odd number of points is more real world but there is something to be said for the forced choice of an even number of scale points. An even number also forces more variance in the data, which often makes it easier to analyze.

Like

5 Best practice 2: Use consistent and clear labels

Another best practice to follow when choosing the labels for your Likert scale questions is to use consistent and clear labels, preferably with words rather than numbers. This can help you avoid confusion and ambiguity for respondents and analysts, as well as increase the validity and reliability of your questions. You should use labels that match the tone and direction of your statement, and that cover the full range of possible responses. For example, if you are asking about satisfaction, you can use labels such as very satisfied, somewhat satisfied, neither satisfied nor dissatisfied, somewhat dissatisfied, and very dissatisfied.

Add your perspective

George Kuhn

Founder & President @ Drive Research | Market Research Company 📊 | You have questions. We get answers from those who matter most. 🎯 | Visit our website for more advice on how to fuel your strategy using data. 📈
Report contribution
Something to consider here is only labeling the anchor points (e.g., 1 = not at all likely, 5 = very likely). Respondents are smart enough to understand what 2-4 means in this context. I see far too survey fails where writers go overboard labeling every single Likert point on a 7 or 9-point scale and it often creates unbalanced ratings where the mid-point is not the true mid-point (e.g., 1 = bad, 2= fair, 3 = good, 4 = very good, 5 = excellent).

Like

6 Best practice 3: Test and refine your questions

A final best practice to follow when choosing the number of points and labels for your Likert scale questions is to test and refine your questions before launching your survey. You can do this by conducting a pilot survey with a small sample of your target population, and asking for feedback on the clarity, relevance, and difficulty of your questions. You can also analyze the distribution, frequency, and variability of your responses, and look for any patterns, trends, or outliers. Based on the results, you can adjust the number of points and labels, or rewrite the statement, to improve the quality and accuracy of your questions.

Add your perspective

How do you choose the appropriate number of points and labels for Likert scale questions?

1

2

3

4

5

6

1 Factor 1: Reliability and validity

2 Factor 2: Response rate and ease of analysis

3 Factor 3: Context and purpose

4 Best practice 1: Use an odd number of points

5 Best practice 2: Use consistent and clear labels

6 Best practice 3: Test and refine your questions

Market Research

Rate this article

Thanks for your feedback

More articles on Market Research

More relevant reading

How do you choose the appropriate number of points and labels for Likert scale questions?

1

2

3

4

5

6

1 Factor 1: Reliability and validity

2 Factor 2: Response rate and ease of analysis

3 Factor 3: Context and purpose

4 Best practice 1: Use an odd number of points

5 Best practice 2: Use consistent and clear labels

6 Best practice 3: Test and refine your questions

Market Research

Rate this article

Thanks for your feedback

Explore Other Skills