Last updated: 7 October 2026
Quick Answer: A Likert scale is a measure that asks how strongly people agree or disagree with a set of statements, usually on 5 or 7 points. Use it for opinions you can phrase as statements, and label every point. Reliability gains stop at about six options, according to research reviewed in Frontiers in Psychology in 2021.
A Likert scale is a way of turning opinions into numbers: you write statements about a topic, ask people how strongly they agree with each, and combine the answers into a score. Social psychologist Rensis Likert introduced it in 1932, and it is now the default format of attitude research.
Its popularity hides a trap. A Likert item tells you how strongly people agree, not why, and the agree-disagree format itself nudges some respondents toward "agree", so the scale works best when it is designed carefully and paired with a reason.
How Does a Likert Scale Work?
A Likert scale works by presenting several statements about one topic and a fixed set of answer options, usually from "strongly disagree" to "strongly agree". Each option gets a number, and the numbers across statements are added or averaged into a single score for the attitude being measured.
The terms get used loosely, so it helps to separate them:
- Likert item: One statement with its agreement options, the "Likert scale question" most surveys mean.
- Likert scale: The full set of related items, combined into one score.
- Likert-type item: Any similar labeled rating, such as frequency or satisfaction, that is not strictly about agreement.
The typical Likert scale is a 5- or 7-point ordinal scale, as Gail Sullivan and Anthony Artino summarize in a 2013 Journal of Graduate Medical Education editorial on analyzing this kind of data.
What Are Some Likert Scale Examples and Questions?
Likert scale examples share one structure: a clear statement or question and a symmetric set of labeled answers. The statement carries the content; the answer labels carry the dimension being rated, whether agreement, frequency, importance, satisfaction or likelihood.
The rows below run from the classic agreement format to the most common Likert-type variants.
| Dimension | Example item | 5-point labels |
|---|---|---|
| Agreement | "The checkout process was easy." | Strongly disagree, Disagree, Neither, Agree, Strongly agree |
| Frequency | "How often do you buy this brand?" | Never, Rarely, Sometimes, Often, Always |
| Importance | "How important is free delivery?" | Not at all, Slightly, Moderately, Very, Extremely |
| Satisfaction | "How satisfied are you with the app?" | Very dissatisfied to Very satisfied |
| Likelihood | "How likely are you to buy this again?" | Very unlikely to Very likely |
Good Likert scale questions ask about one thing at a time. "The product is affordable and well made" mixes two judgments, and a respondent who agrees with one and not the other has no honest answer.
Three more habits keep items clean:
- Keep the scale balanced: Use as many positive options as negative ones, with neutral in the middle.
- Use plain words: "Easy to use" beats "intuitive user experience" for most audiences.
- Mix the direction: Word a few statements negatively so agreement does not always mean approval.
Should You Use a 5-Point or 7-Point Likert Scale?
Use a 5-point Likert scale for most customer and market research, and a 7-point scale when you need finer distinctions from attentive respondents. Adding options improves reliability only up to a point, and labeling every option matters at least as much as the count.
A 2021 review of Likert scale research in Frontiers in Psychology by Andrew Jebb and colleagues summarized the evidence:
- More options helped, then stopped helping. Reliability rose with more response options, but the benefit stopped after six, and 0 to 1,000 sliders showed no gain.
- A neutral midpoint made no measurable difference. Including or removing "neither agree nor disagree" showed no psychometric effect in the study reviewed.
- Full labels beat end labels. Labeling every option gave higher reliability than labeling only the two ends.
The US government settled on five for its own service surveys. Federal customer experience surveys under OMB Circular A-11, Section 280 prefer a 5-point Likert scale for their satisfaction and trust questions, which keeps results comparable across agencies.
Is a Likert Scale Ordinal or Interval?
A single Likert item is ordinal: its answers have an order, but the gaps between them are not known to be equal. A scale built from many items behaves more like interval data, which is why summed Likert scores are routinely averaged.
Sullivan and Artino note that experts long argued for medians and nonparametric tests on Likert data. They also cite evidence that parametric tests, such as t-tests, are robust enough to give answers close to the truth on Likert scores, even when assumptions are bent.
How Do You Analyze Likert Scale Data?
To analyze Likert scale data, start with the share choosing each answer, then summarize with a top-two-box percentage or a median. Average only items that form a tested scale. Shares keep the shape of opinion visible; a single mean can hide a split audience.
A practical order of work:
- Chart the full distribution for each item, ideally as a stacked bar.
- Report the top-two-box share, such as "agree" plus "strongly agree".
- Compare groups with tests suited to the data, such as chi-square for shares.
- Average or sum items only when they measure one attitude and hang together statistically.
Two groups can share a mean of 3.0 and mean opposite things: one may sit at the midpoint, the other may be split between 1 and 5. The distribution shows which.
A concept that scores 40% top-two-box and 35% bottom-two-box is polarizing, not average, and that calls for a different decision from one with 60% sitting on neutral.
How Is a Likert Scale Different From Other Rating Scales?
A Likert scale measures agreement with statements and combines several items into one score, while most other rating scales ask for a single judgment on one dimension. A star rating, a 0 to 10 recommend question or a satisfaction score are rating scales, and some are called Likert-type when their labels mimic the agreement format.
The difference matters for analysis. A true Likert scale has several items that can be checked for consistency, while a single rating has nothing to check it against. A semantic differential is another cousin, placing a 5- or 7-point line between two opposite adjectives, such as "cheap" and "premium". The 0 to 10 Net Promoter question is a single-item rating scale, not a Likert scale, which is why it is scored by bands rather than summed.
When Should You Use a Likert Scale, and When Not?
Use a Likert scale to measure an attitude people can judge as a statement, such as trust in a brand, ease of a task or appeal of a concept. Avoid it for facts and behaviors that people can report directly, such as how many times they shopped last month.
- Good fits: Brand perceptions, ease of use, agreement with claims, attitude change between waves.
- Poor fits: Counts, prices, yes-or-no facts, and anything the respondent has never experienced.
Agree-disagree items carry a known risk. Pew Research Center's guide to writing survey questions notes that less educated and less informed respondents are more likely to agree with statements, an acquiescence bias that grows when an interviewer is present. Its recommended fix is to offer a choice between two alternative statements.
Long batteries of Likert items in a grid invite another problem, straight-lining, where tired respondents tick the same column down the page. Wording problems compound both, as the guide to biased survey questions shows.
How Do Likert Questions Work on Phone and Chat?
Likert questions work on phone and chat when the scale is short enough to hold in memory and every point is spoken or written as a label. A respondent can see seven boxes on a screen, but hearing seven labels read aloud is harder, which is why voice surveys often use five points or fewer.
Those channels matter because they reach people a web form misses. WhatsApp-native interviews present each item inside a chat, with no link and no app, and accept a typed number or a voice note. AI phone interviews read the labels aloud for people who prefer to talk. Alchemic publishes 60+ languages including Spanish and Hindi, so scale labels can be asked in the respondent's own language rather than translated on the fly.
What Do Likert Scores Miss?
Likert scores miss the reason behind the rating. A 3 out of 5 on "this concept is appealing" might mean mild interest, confusion about what the product does, or a price worry, and the number alone cannot tell you which one to fix.
Pairing the scale with a follow-up question closes that gap. In concept testing, Alchemic's AI moderator probes behind each score, asking why a respondent chose 3 rather than 4, so a quote bank and a distribution chart come from the same interview.
Scales also travel imperfectly across cultures and languages, because people differ in how readily they use the extreme points. Once it has the client's brief, Alchemic designs and tailors the discussion guide to handle these risks before fielding, rather than leaving that work to the buyer. When the only goal is a fast, comparable score across many waves, a plain survey tool with a fixed 5-point scale is the simpler choice.