Agree-Disagree Questions: Examples, Writing Rules, and Better Alternatives

Agree-Disagree questions ask respondents how strongly they agree with a statement. Most scales run from strongly disagree to strongly agree.

These questions turn opinions into numbers you can track, compare, and report. HR teams, customer experience teams, and market researchers use them because respondents understand the format at a glance.

The format also fails quietly. Vague wording, the wrong scale length, and a habit of saying “yes” can all skew results. In this guide, we’ll cover examples, writing rules, scale choices, bias fixes, and the cases where another format works better.

What are agree-disagree questions?

Agree-disagree questions are closed-ended survey items that ask respondents how strongly they agree or disagree with a statement. The answer options form an ordered scale, so every response becomes a number you can analyze.

Each item has three parts:

• The statement: One claim, such as “My manager gives me useful feedback.”
• The response scale: Ordered options, usually five, from strongly disagree to strongly agree.
• The scoring rule: A numeric code for each option, such as 1 to 5, used to calculate percentages and averages.

The format is also called an agreement scale. It measures attitudes, opinions, and perceptions, not facts. “I find the software easy to use” measures a perception. “How many times did you log in last week?” measures behavior and needs a different format.

How do agree-disagree questions differ from Likert scales?

An agree-disagree question is one type of Likert-style item. A Likert scale is the broader family of ordered response formats, named after psychologist Rensis Likert. It can measure agreement, frequency, satisfaction, importance, or likelihood.

The terms get mixed up, yet the formats behave differently. This table separates the closest relatives.

Polarity adds one more distinction. A bipolar scale runs between opposites, such as disagree and agree. A unipolar scale runs from none to a lot, such as not at all satisfied to completely satisfied. This guide to unipolar versus bipolar Likert scales explains when each one fits. Agree-disagree scales are bipolar.

What do good agree-disagree question examples look like?

Good agree-disagree questions state one specific, neutral claim that respondents can judge from their own experience. The 15 examples below cover workplace, customer, and research settings. Each one uses the same five options: strongly disagree, disagree, neither agree nor disagree, agree, and strongly agree.

Workplace and training statements

Workplace statements work best when each one maps to a decision a manager can act on.

• “My manager gives me useful feedback on my work.”
• “I have the tools I need to do my job well.”
• “I understand how my role supports company goals.”
• “The onboarding training prepared me for my first month.”
• “I would recommend this company as a place to work.”

A drop on the tools statement points to a resourcing fix. A drop on a vague “I feel engaged” item points nowhere.

Customer experience statements

Each customer statement should tie to one touchpoint, so a low score shows you where to look.

Brand and product research statements

Researchers use agree-disagree questions to test perceptions across large samples.

• “This brand offers good value for the price.”
• “This brand’s advertising feels relevant to me.”
• “I can easily recognize this brand’s products on a shelf.”
• “The new dashboard makes it easier to find what I need.”
• “The product’s price is fair for the features included.”

Willingness-to-pay statements such as “I would pay more for a sustainable version” invite socially desirable answers. Treat those results as a ceiling. Check them with a method where respondents choose between realistic options.

How to write agree-disagree survey questions

Write each statement about one idea, in neutral and specific language, then test it on a small group before launch. Strong statements make every other part of your survey questions easier to get right.

Keep one idea in each statement

A double-barreled statement asks about two ideas at once, so a respondent who feels differently about each cannot answer honestly. Double-barreled questions are one of the most common wording flaws in agree-disagree surveys.

Scan every draft for these red flags.

• The word “and” or “or” joining two separate ideas
• A comma that adds a second clause
• A paired adjective, such as “fast and friendly”

A statement like “My manager is supportive and communicates clearly” needs splitting. Write one statement about support and one about communication.

Use neutral, specific wording

Neutral wording removes cues that push respondents toward agreement. Specific wording makes sure everyone reads the statement the same way. Write at a reading level your whole audience can handle, and skip jargon.

Pilot test before launch

A pilot test checks whether respondents read each statement the way you meant it. Run one every time, even for short surveys.

1. Recruit about 8 to 10 people who match your audience.
2. Ask each person to restate every statement in their own words.
3. Flag any statement that two or more testers read differently.
4. Flag statements that several testers answer with “neither agree nor disagree,” since those are often unclear.
5. Rewrite the flagged items and retest them.

How many points should an agree-disagree scale have?

Five points suit most surveys. Choose seven for finer separation between mild and strong opinions. Choose four to make respondents pick a side. The right length depends on your audience, your analysis plan, and your survey channel.

A 2010 review found seven-point scales usually more reliable than five-point scales. The evidence is mixed, though. The same summary suggests a three-point scale loses little when respondents struggle with longer ones. For ready-made wording at each length, browse these Likert scale examples.

Should you include a neutral option?

Include a neutral option unless you need to force a choice. Removing it tends to add noise. A middle option attracted between 6% and 23% of respondents across studies. Most came from weak agreement and weak disagreement, not from people with no view. A large analysis combining many studies also found that a neutral midpoint lowers random error.

One caveat applies. In a survey of about 3,000 people in Great Britain, most who chose “neither agree nor disagree” had no real opinion. Researchers call this a hidden don’t know. To test how solid your results are, recode neutral answers as don’t know and see how much they change.

What goes wrong with agree-disagree questions?

The five common problems are acquiescence bias, straight-lining, overuse of the middle option, social desirability bias, and missing reasons. Acquiescence bias is a habit of agreeing no matter what the statement says. Straight-lining means giving the same answer to every statement in a grid. Social desirability bias means choosing answers that look acceptable rather than true.

How acquiescence bias works

Acquiescence bias is the tendency to agree with a statement regardless of what it says. It inflates positive results and hits agree-disagree questions hardest.

Two forces drive it. People tend to be polite, and people tend to save effort. Satisficing means giving an answer that is good enough rather than accurate, to avoid thinking hard. The Encyclopedia of Survey Research Methods says acquiescence is likely strongest among respondents low in ability and motivation. It is also likely to be stronger on harder questions.

Methodologists consider this format prone to bias. Some people agree more readily with a statement than with a balanced version of the same idea. Agree-disagree answers are less reliable than item-specific answers in survey experiments.

How to reduce acquiescence bias and straight-lining

You cannot remove these biases entirely, but a few design choices shrink them.

• Mix statement direction.
  Pair “I get useful feedback” with “Feedback I receive rarely helps me improve.” Avoid the word “not,” because agree-disagree items with negation had lower reliability.
  
• Keep grids short.
  Respondents who rush also straight-line more in grid questions, especially less-educated respondents.
  
• Review before you delete.
  It can be valid when a respondent truly feels the same about every item. Compare the pattern with completion time before removing anyone.
  
• Add a follow-up for extremes.
  Ask an open-ended “why” after strong disagreement, so you capture the reason and not only the score.

When should you use an item-specific scale instead?

An item-specific scale uses answer labels tied to the question itself. Labels might run from not at all useful to extremely useful. Use an agree-disagree scale to track broad attitudes over time. Switch to an item-specific scale when you need a precise rating of one thing.

Compare two versions of the same question. “The training was useful” is the agree-disagree version. “How useful was the training?” is the item-specific version, with five labels from not at all useful to extremely useful. The second version asks respondents to judge usefulness directly, instead of agreeing with the survey writer’s claim.

A University of Nebraska-Lincoln study compared the formats in 1,002 returned mail surveys. On half the positively worded questions, a self-description scale produced lower averages and more varied answers. The evidence is not one-sided, though.

Item-specific questions are as reliable and valid as agree-disagree ones. That held even among respondents most likely to acquiesce.

Choose agree-disagree for many attitudes you benchmark against earlier waves. Choose item-specific for one critical metric, a quality rating, or any frequency or behavior question.

How do you analyze agree-disagree results?

To analyze agree-disagree results, code each answer as a number and report the percentage who agree. Show the full distribution next to any average. These steps keep the numbers honest.

1. Code the scale from 1 (strongly disagree) to 5 (strongly agree). Reverse-code negatively worded statements, which means flipping their numbers so a high score always points the same direction.
2. Calculate percent favorable, the share of respondents who chose agree or strongly agree.
3. Show the full distribution so readers see disagreement and neutrality, not just the top.
4. Report the mean if your audience expects one. Read it as a summary of ordered categories, not exact distances between options.
5. Compare segments, such as teams, regions, or customer groups, and track the same statement across survey waves.

Worked example with 200 responses

Here is how the numbers work for one statement answered by 200 people.

The percent favorable is 70%, because 50 plus 90 equals 140 out of 200. The mean is 3.75, from 750 total points divided by 200 people. The 15% neutral share sits inside the 6% to 23% range in the World Bank summary, so nothing looks off.

Data checks to run before you report

Run these four checks before sharing any agree-disagree results.

• Neutral share.
  A share far above the 6% to 23% range suggests the statement is unclear or irrelevant to many respondents.
  
• Straight-line rate.
  Count respondents who gave one answer to every statement in a grid of five or more rows. Compare that group with completion time before you remove anyone.
  
• Direction consistency.
  Confirm that reverse-coded statements move opposite to positively worded ones. If they do not, respondents may be misreading them.
  
• Segment gaps.
  Look for groups with unusually high agreement across unrelated items, which can signal an acquiescence pattern.

How to build agree-disagree questions in QuestionPro

In QuestionPro, you build a battery of agree-disagree statements with a Matrix Question. Carousel Layout and answer display order then protect data quality. A battery is a set of statements that share one response scale. A matrix question lists those statements as rows against shared scale columns.

1. Add a Matrix Question in QuestionPro Survey Software. Enter each statement as a row and your scale labels as columns.
2. Open the Matrix Question settings and switch on the Cards Carousel toggle. Each statement now appears as its own card, one row at a time, instead of one large grid.
3. Preview the question to confirm every row shows as a separate card.
4. For a standalone agree-disagree question, open its settings and choose Alternate Flip under Answer Display Order. QuestionPro then reverses the scale order for every other respondent. Skip Random for an ordered scale, since shuffling agree and disagree would confuse people.

Carousel Layout changes only how the matrix displays. You can switch it on for an existing matrix without changing the question structure, answer options, or logic.

Treat agree-disagree questions as a measurement choice, not a default

Agree-disagree questions earn their place when you want a fast, comparable read on many attitudes at once. They stop earning it when one answer carries a big decision. They also fall short when you need to know why people feel the way they do.

The strongest surveys mix formats. Use agree-disagree for breadth, item-specific scales for the metrics that matter most, and open-ended questions for reasons. Write every statement as if one respondent will misread it, then test until none do.

Frequently Asked Questions (FAQs)