Decision Making
Choice Blindness Task
Also called: choice blindness paradigm, face choice blindness, choice manipulation task, Johansson and Hall face task
Measure whether people notice when a choice's outcome is swapped: participants pick the more appealing face of each pair and give reasons, and on 3 pairs see the other face, scored by the swaps they report.
6-8 minUntimed
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Preview any test free after you sign up.
- Papers, 2021 to 2025
- 33
- Evidence grade
- D
- Registered clinical studies
- 0
What it measures
Measures whether people notice when the outcome of a choice is not what they chose. Participants pick the more appealing face of 12 pairs and say why; on 3 pairs the other face is shown as their choice. The main score is the share of those swaps they report.
Constructs
- Choice blindness
- Monitoring of one's own choices
- Introspection and confabulation
- Choice-induced preference change
Research uses
- Decision-making and self-knowledge research
- Studies of introspection and confabulation
- Preference change research
- Group comparison within a study
How the paradigm works
The Choice Blindness Task measures whether people notice when the outcome of a choice is not what they chose. Johansson, Hall, Sikström and Olsson (2005) asked participants to choose the more attractive of two faces and then to explain the choice; on some trials they were secretly handed the face they had not chosen. Most of these swaps went unnoticed, and people gave reasons for a choice they had not made. This computer version follows Remington and colleagues (2024): each pair shows for 3 seconds (2 to 5 as a setting), then two cards cover the faces, and the participant clicks or taps the card of the face they find more appealing and says how sure they are, from 1 to 7. The face then appears again with ten reasons to tick, such as eyes or overall look. On 3 of the 12 pairs it is the other face of the pair. Ticking "I did not mean to choose this face", which the instructions offer for a wrong click, reports the swap.
More about the paradigm
The faces are 24 photographs from the Face Research Lab London Set, white women and men aged 18 to 54, paired by gender and age: 6 pairs of women and 6 of men. The same people appear in the composite face task and never in the face memory tests. The order of the pairs, which face is on the left and which 3 pairs are swapped come from the session's seed: never one of the first 3 pairs, and never two in a row. One practice pair of drawn faces comes first. A second round shows every pair again for a new choice, without swaps or reasons (Johansson, Hall, Tärning, Sikström and Chater, 2014). Two questions end the task: whether anything seemed odd, then, after a sentence saying that faces were swapped, whether the participant noticed. Studies usually explain the procedure afterwards.
The main score is the detection rate: the share of the 3 swaps reported while they happened. In the original study no more than 26% of swaps were detected in any way, and in a computer version about 1 swap in 6 was reported at the time (Taya, Gupta, Farber and Mullette-Gillman, 2014). The answers at the end, the not-chosen answer on unswapped pairs, certainty and the number of reasons ticked are also reported. In the second round, consistency is the share of pairs where the first choice is chosen again; preference change is consistency on unswapped pairs minus consistency on swapped pairs (93% against 57% in Johansson and colleagues, 2014). Choices under 100 ms count as anticipations. With three swaps, one person's score is coarse, so the scores suit comparisons between groups. Scores support research and do not diagnose.
Scores
- Detection Rate%
- Share of the swapped pairs on which "I did not mean to choose this face" was ticked.
- Higher means more swaps noticed at the time
- Swaps Reportedcount
- Swapped pairs on which the not-chosen answer was ticked.
- Informational
- Swaps Showncount
- Swapped pairs answered: 3 in a complete session.
- Informational
- Not-Chosen Answers Without a Swapcount
- Pairs where the chosen face was shown and the not-chosen answer was ticked, as for a wrong click.
- 0 expected; 3 or more are flagged
- Something Seemed Odd0/1
- 1 when the participant answered Yes to whether anything seemed odd, asked before swaps are mentioned.
- Informational
- Swap Noticed, Asked0/1
- 1 when the participant answered Yes to whether they noticed swaps, asked after a sentence saying that faces were swapped.
- A cued report; can include guesses
- Consistency, Unswapped Pairs%
- In the second round, the share of unswapped pairs where the first choice was chosen again.
- Informational
- Consistency, Swapped Pairs%
- In the second round, the share of swapped pairs where the first choice, not the face shown after the swap, was chosen again.
- Lower means the swapped face came to be preferred
- Preference Change%
- Consistency on unswapped pairs minus consistency on swapped pairs.
- Positive means the swaps changed later choices
- Mean Certaintyscore
- Mean certainty after first-round choices, from 1 (not sure) to 7 (completely sure).
- Informational
- Certainty on Swapped Pairsscore
- Mean certainty after the choices that were then swapped.
- Informational
- Reasons per Unswapped Pairscore
- Mean number of reasons ticked when the chosen face was shown.
- Informational
- Reasons per Unnoticed Swapscore
- Mean number of reasons ticked for the face shown after a swap that was not reported.
- Reasons given for a choice not made
- Median Choice Timems
- Median time from the cards appearing to the first-round choice, over choices at or after 100 ms.
- Informational
- Choices Under 100 mscount
- Choices made less than 100 ms after the cards appeared, in either round.
- 0 expected
- Pairscount
- First-round choices scored.
- Informational
Use and evidence
33 papers from 2021 to 2025 name this task in their title or abstract, down 27% compared with 2016 to 2020. That ranks it 288 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 10 |
| 2024 | 5 |
| 2023 | 12 |
| 2022 | 3 |
| 2021 | 3 |
| 2020 | 10 |
| 2019 | 9 |
| 2018 | 7 |
| 2017 | 9 |
| 2016 | 10 |
| 2015 | 16 |
| 2014 | 10 |
| 2013 | 14 |
| 2012 | 8 |
| 2011 | 6 |
| 2010 | 2 |
| 2009 | 3 |
| 2008 | 3 |
| 2007 | 4 |
| 2006 | 6 |
| 2005 | 1 |
| 2004 | 0 |
| 2003 | 0 |
| 2002 | 0 |
| 2001 | 0 |
| 2000 | 1 |
Use grade D
30 to 99 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 0
- Papers on reliability or validity
- 9
- Registered clinical studies using it as an outcome
- 0
How the grades work
Use: A is 1,000 or more papers from 2021 to 2025, B 300 to 999, C 100 to 299, D 30 to 99 and E fewer than 30. Evidence: A means at least 25 papers report its test-retest reliability and at least 25 registered clinical studies use it as an outcome; B, at least 8 test-retest papers, and either 5 registered clinical studies or 150 reliability or validity papers; C, at least 2 test-retest papers, 30 reliability or validity papers or 3 registered clinical studies; D, little published measurement evidence found. The counts measure use. They include every paper that names the task in its title or abstract. Grades describe the published research on this task, not ConductCognition's version of it.
Source: ConductCognition test census 0.14.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Faces
- 24 grayscale photographs from the Face Research Lab London Set, white women and men paired by gender and age: 6 pairs of women and 6 of men. No screen names or rates a face.
- Each pair
- Shown for 3 seconds (standard; 2 to 5), then two cards cover the faces. The participant clicks or taps the card of the face they find more appealing, then rates how sure they are from 1 to 7. No time limit for either.
- Reasons
- The chosen face appears again with 10 reasons to tick, such as eyes or overall look, and the answer "I did not mean to choose this face", which the instructions offer for a wrong click.
- Swaps
- On 3 of the 12 pairs the other face is shown instead: never in the first 3 pairs and never two in a row. The swapped pairs, the order and the sides are drawn from the session's seed and saved.
- Second round
- All 12 pairs again in a new order for a new choice, without swaps or reasons (standard; can be turned off).
- Questions
- Whether anything seemed odd, then, after a sentence saying that faces were swapped, whether the participant noticed.
- Practice
- One pair of drawn faces, never swapped.
- Scores
- The share of swaps reported at the time (detection rate), the answers to the questions, second-round consistency on swapped and unswapped pairs, certainty and the reasons ticked.
Johansson and colleagues (2005) swapped printed photographs by sleight of hand and recorded spoken reasons. This computer version follows Remington and colleagues (2024), who showed the pairs on cards that turned over and offered reasons as tick boxes, with "I did not intend to choose this face" as the report of a swap; the reasons here were written for this task. The photographs come from the London Set (CC BY 4.0), the people of the composite face task. The task shows people a face they did not choose, so studies usually explain the procedure afterwards. Three swaps give a coarse score for one person, so compare groups. The scores support research and do not diagnose.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: the scored run's card choices in both rounds. No face is the right choice, so accuracy is not judged against chance. The same side on at least 90% of choices is flagged for review. A run with no manipulated pair answered is invalid; a run short of its planned pairs and the not-chosen answer on 3 or more pairs where the chosen face was shown are flagged. Choices under 200 ms on at least half the pairs invalidate.
Why these rules
There is no correct option to compare with chance, so accuracy is not judged against chance. A preference does not depend on the side a face is shown on, so choosing one side nearly every time suggests the faces were not compared; it is flagged, never excluded. The main score needs the manipulated pairs. The not-chosen answer is also how a misclick is reported, so a few on unswapped pairs are expected; many suggest the answer was misread. An answer under 200 ms is faster than a simple visual reaction, so it cannot reflect the item. Following response-time effort screening for untimed tests (Wise and Kong 2005), the session is invalid only when at least half the answers are that fast.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Time each pair is shown (seconds) | 3 | 2 to 5 |
| Second round | On | On or off |
Yearly plans can change settings within these limits. Free and pay-as-you-go studies use the standard settings.
Need a setting this test does not have?
Devices and languages
- Untimed test: runs on desktops, laptops and tablets. Participant devices
- Languages: English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese.
- Available on: Every plan, including the free plan.
- No percentile. Scores and quality flags only. How norms work
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the Choice Blindness Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Johansson P, Hall L, Sikström S, Olsson A, 2005), with the standard settings (time each pair is shown (seconds): 3; second round: On). The main score was detection rate (share of the swapped pairs on which "I did not mean to choose this face" was ticked).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Johansson, P., Hall, L., Sikström, S., & Olsson, A. (2005). Failure to detect mismatches between intention and outcome in a simple decision task. Science, 310(5745), 116-119. https://doi.org/10.1126/science.1111709The paradigm: swapped faces, and reasons given for a choice not made.
- Johansson, P., Hall, L., Tärning, B., Sikström, S., & Chater, N. (2014). Choice blindness and preference change: You will like this paper better if you (believe you) chose to read it! Journal of Behavioral Decision Making, 27(3), 281-289. https://doi.org/10.1002/bdm.1807The second round of choices and preference change.
- Taya, F., Gupta, S., Farber, I., & Mullette-Gillman, O. A. (2014). Manipulation detection and preference alterations in a choice blindness paradigm. PLoS ONE, 9(9), e108515. https://doi.org/10.1371/journal.pone.0108515A computer version with certainty ratings and questions at the end.
- Remington, A., White, H., Fairnie, J., Sideropoulos, V., Hall, L., & Johansson, P. (2024). Choice blindness in autistic and non-autistic people. Journal of Cognitive Psychology, 36(4), 493-501. https://doi.org/10.1080/20445911.2024.2356283Cards that cover the faces, reasons as tick boxes, and the answer that reports a swap.
- DeBruine, L., & Jones, B. (2017). Face Research Lab London Set [Data set]. figshare. https://doi.org/10.6084/m9.figshare.5047666.v5
Source papers
- Johansson, Hall, Sikström and Olsson (2005): 120 participants, 15 pairs each with 3 swaps; no more than 26% of swaps detected in any way. Taya and colleagues (2014): 16.9% of 160 swaps reported at the time on a computer. Johansson and colleagues (2014): second-round consistency of 93% on unswapped and 57% on swapped pairs.Paradigm source, Adults
Questions
How long does the Choice Blindness Task take?
About 6 to 8 minutes, as listed in the test library.
What device does a participant need?
Untimed test: runs on desktops, laptops and tablets. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Detection Rate, Swaps Reported, Swaps Shown, Not-Chosen Answers Without a Swap, Something Seemed Odd, Swap Noticed, Asked, Consistency, Unswapped Pairs, Consistency, Swapped Pairs, Preference Change, Mean Certainty, Certainty on Swapped Pairs, Reasons per Unswapped Pair, Reasons per Unnoticed Swap, Median Choice Time, Choices Under 100 ms and Pairs. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Choice Blindness Task used in research?
33 papers from 2021 to 2025 name it in their title or abstract, down 27% compared with 2016 to 2020. That ranks it 288 of 395 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.14.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Choice Blindness Task have?
In the published research, 0 papers report its test-retest reliability and 0 registered clinical studies use it as an outcome (evidence grade D in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Time each pair is shown (seconds) (2 to 5) and second round (On or off). Settings can be changed on yearly plans; free and pay-as-you-go studies use the standard settings.
Which languages can participants use?
Participants can take this test in English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese. English is on every plan, and the other languages come with yearly plans. Have a native speaker read the machine-translated text before launch.
Is this a clinical test?
No. ConductCognition is for research use only. It runs, scores and exports the test; clinical interpretation is outside the platform.
Every test: the full test list.
Run this test in your next study
Start free with 10 free participants on any test, then use participant credits or a yearly plan.