Attention Check Questions: 15 Examples and How to Use Them

Attention Check Questions: 15 Examples and How to Use Them

Attention Check Questions: 15 Examples and How to Use Them

Written by

Written by

Patricio Luna

Patricio Luna

Published on

Published on

What Is an Attention Check Question?

An attention check is a survey question with one obviously correct answer, placed among the real questions to find respondents who are not reading. People who get it wrong are flagged or removed, so their answers do not distort your results.

Want to check this for your own idea?

Free to start. No card needed.

Some people rush through surveys. They click the first option, pick the same column all the way down a grid, or answer at random to reach the end.

A handful of these responses can move a result by several points. An attention check is the simplest way to find them.

The check has to be fair. A good one is easy for anyone who reads it and impossible to pass by chance most of the time. A bad one fails honest people and tells you nothing.

Key point: an attention check measures attention. It should never test memory, knowledge or patience.

The Five Kinds of Attention Checks

1. Instructed Answer

The question tells respondents exactly which answer to choose. This is the clearest and most widely used kind.

2. Positional Instruction

The question asks for an answer by its position in the list, so it works in any language and cannot be passed by recognizing a keyword.

3. Bogus Statement

A statement that cannot be true for anyone, mixed in with real ones. Agreeing with it signals careless answering.

4. Consistency Check

The same fact asked twice in different ways. Contradictory answers are a warning sign.

5. Trap Option

A fake answer inside a real question, such as a brand that does not exist.

15 Attention Check Question Examples

Instructed Answer

1. "To confirm you are reading this survey carefully, please select 'I am reading carefully' from the options below."
Answers: I am skimming quickly / I am guessing my answers / I am reading carefully / I am not paying attention.

2. "For this question, please select 'Somewhat agree'."
Use the same answer scale as the questions around it, for example one of these Likert scales.

3. In a grid of statements, add the row: "Please select 'Disagree' for this row."
This catches people who click one column all the way down.

Positional Instruction

4. "For this question only, please select the third option from the list below to show you are reading carefully."
Answers: First option / Second option / Third option / Fourth option.

5. "Please choose the last answer in this list."

6. "Two of the words below are colors. Please select both."
Answers: Table / Blue / Tuesday / Green.

Bogus Statement

7. "I have never used the internet." (Agree or disagree, in an online survey.)

8. "I was born on February 30th." (True or false.)

9. "In the past month I have visited every country in the world." (Agree or disagree.)

Consistency Check

10. Ask for age at the start and year of birth near the end.

11. Ask "Have you bought [category] in the last 12 months?" early, and later "When did you last buy [category]?" with "More than a year ago" as an option.

12. Ask the same thing in both directions, far apart: "The app is easy to use" and later "The app is difficult to use".

Trap Option

13. "Which of these brands have you heard of?" with one invented brand name in the list.

14. "Which of these have you done in the past week?" with one impossible activity among the real ones.

15. After showing a product description: "Which of these best describes the product you just read about?" with one correct summary and three clearly wrong ones.

Where to Place an Attention Check

  • In the middle, not at the very start. Attention drops as a survey goes on, so a check on the first screen catches almost no one.

  • Inside the section that matters most. If the pricing questions decide your study, put the check among them.

  • Not as the last question. By then the damage is done and you learn nothing about the answers before it.

  • In the same format as its neighbors. A check that looks different from every other question is spotted without being read.

How Many Attention Checks to Use

  • One for a survey of up to about 15 questions.

  • Two for longer surveys, spaced apart and of different kinds.

  • No more than three. Each extra check annoys careful respondents, and annoyed respondents give worse answers.

What to Do When Someone Fails

Treat the kinds differently.

Instructed answer and positional checks are unambiguous. A respondent who fails one was not reading. Removing them is standard practice.

Bogus statements, consistency checks and trap options can be failed honestly. People misread, misremember or genuinely confuse brand names. Use these as flags and remove a response only when a second signal agrees.

The other signals worth combining with a failed check:

  • Speeding: finishing far faster than the questions can be read.

  • Straightlining: the same answer for every row of a matrix question.

  • Nonsense open answers: random characters, or text pasted from elsewhere.

  • Repeat submissions: the same person answering more than once.

Decide your rule before you look at the results, and report how many responses you removed and why.

Attention checks clean up who answered. They do not fix who you asked: that depends on your sampling method.

Four Mistakes That Fail Honest Respondents

1. Testing memory. "What was the third question about?" measures recall. Attentive people forget details too.

2. Testing knowledge. "What is the capital of Australia?" excludes people for what they know, which biases your sample.

3. Vague wording. If a careful reader could defend two answers, the check is broken.

4. Tricks. A long paragraph with an instruction hidden in the last line punishes people for reading the way people normally read. Say plainly what you are asking.

If more than a small share of respondents fail a check, look at the check before you blame the respondents. Questions that work well around a check are in these 50 survey question examples.

Attention Checks in SegmentOS

Every SegmentOS template includes an attention check, placed in the middle of the questionnaire. Two versions are used: an instructed answer ("please select 'I am reading carefully'") and a positional one ("please select the third option"). Employee surveys use a softer wording of the first.

Respondents who fail are removed, and the results show a data quality summary of what was taken out: screening and attention checks, speeding, and repeat submissions. Speeding detection can be switched on to disqualify anyone who answers faster than a minimum time per question.

Key Takeaways

  • Definition: an attention check has one obviously correct answer and finds respondents who are not reading.

  • Use an instructed answer or a positional check when you plan to remove people who fail.

  • Treat bogus statements, consistency checks and trap options as flags, not as proof.

  • One check is enough for a short survey, placed in the middle.

  • Never test memory or knowledge. A check that fails honest people damages your sample.

What do you want to learn?

Say it in a sentence. SegmentOS writes the survey, and real people answer it.

Free to start. No card needed.

In this article

No headings found on page

Stop guessing. Ask your actual buyers.

Stop guessing. Ask your actual buyers.

Build a study, reach verified respondents in your market, and get answers back in days.

Free plan available · No card required

Frequently Asked Questions (FAQ)

Frequently Asked Questions (FAQ)

What Is an Example of an Attention Check Question?

"To confirm you are reading this survey carefully, please select 'I am reading carefully' from the options below." Anyone who reads it passes, and anyone clicking at random usually fails.

How Many Attention Checks Should a Survey Have?

One for a survey of up to about 15 questions and two for longer ones. More than three irritates careful respondents and lowers the quality of their answers.

Should I Remove Respondents Who Fail an Attention Check?

Remove them when the check was an instructed answer, because failing it means they were not reading. For softer checks, such as consistency questions, remove a response only when another signal like speeding agrees.

Where Should an Attention Check Go in a Survey?

In the middle, inside the section that matters most. A check on the first screen catches almost no one, and a check at the end tells you nothing about the answers before it.

What Makes a Bad Attention Check?

A bad check tests memory or knowledge, has more than one defensible answer, or hides its instruction in a long paragraph. Each of these fails honest respondents.

Insights & Updates

Explore articles, resources, and ideas where we share updates about the product, thoughts on technology, and lessons learned while building along the way.

Insights & Updates

Explore articles, resources, and ideas where we share updates about the product.