Attention
Face-in-the-Crowd Task
Also called: face in the crowd paradigm, anger superiority effect, happiness superiority effect, emotional face search task, threat detection task
Measure how quickly an angry or a happy face is found among neutral ones: participants judge whether nine drawn faces are all the same or one is different.
6-8 minTimed: desktop or laptop with a mouse or trackpad
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
- 42
- Evidence grade
- D
- Registered clinical studies
- 2
What it measures
Measures how quickly an angry or a happy face is found among neutral ones. Nine drawn faces appear in a square; participants press F if all are the same and J if one face is different. The main score is the anger-superiority effect: the time to find a happy face minus the time to find an angry one.
Constructs
- Attention to emotional faces
- Visual search
- Threat detection
Research uses
- Emotion and attention research
- Anxiety research
- Within-study group comparison
How the paradigm works
The Face-in-the-Crowd Task measures how quickly an angry or a happy face is found among neutral ones. Hansen and Hansen (1988) found that an angry face stood out from a crowd faster than a happy one, the anger-superiority effect, and Öhman, Lundqvist and Esteves (2001) reported the same advantage with schematic faces. Here nine drawn faces appear in three rows of three. On a quarter of the trials one face is angry and on a quarter one face is happy, among neutral faces; on the other half all nine are neutral. Participants press F if all the faces are the same and J if one is different, as quickly and accurately as they can. The instructions never name an emotion.
More about the paradigm
The faces are line drawings made by our code: a round outline, two eyes, two brows and a mouth. The angry face has brows tilted 20 degrees down toward the nose and a mouth that bends down; the happy face has the same tilt and bend the other way. Each changes the neutral face by the same amount, so neither differs from the crowd by more or larger lines than the other, a known source of bias with schematic faces (Coelho, Cloete and Wallis, 2010). No photographs are used. A 500 ms cross comes before each crowd, and the crowd stays until a response or 3,000 ms. The standard session is 216 trials in a seeded random order, with each emotional face in each of the nine places 6 times, after 12 practice trials with feedback: about 6 to 8 minutes.
The main score is the anger-superiority effect: mean correct response time with a happy face minus with an angry face. Positive values mean angry faces were found faster; negative values mean happy faces were, the happiness-superiority effect Becker and colleagues (2011) found once low-level differences between photographed faces were removed. The same contrast in accuracy, the effect from medians, and response times and accuracy in each condition, all-neutral crowds included, are also reported. Responses under 100 ms count as anticipations. Which emotion is found faster depends on how the faces are drawn, and a difference between two means is noisy for one person, so the score suits group comparisons.
Scores
- Anger-Superiority Effectms
- Mean correct response time with a happy face minus with an angry face.
- Positive values mean angry faces were found faster; negative values mean happy faces were
- Anger-Superiority Effect, Accuracy%
- Accuracy with an angry face minus accuracy with a happy face.
- Positive values mean fewer errors with angry faces
- Anger-Superiority Effect, Mediansms
- Median correct response time with a happy face minus with an angry face.
- Positive values mean angry faces were found faster
- Response Time With an Angry Facems
- Mean correct response time when one face was angry.
- Lower is faster
- Response Time With a Happy Facems
- Mean correct response time when one face was happy.
- Lower is faster
- Response Time With All Faces the Samems
- Mean correct response time when all nine faces were neutral.
- Lower is faster
- Accuracy%
- Share of all test trials answered with the right key.
- Higher is better
- Accuracy With an Angry Face%
- Share of trials with an angry face answered correctly, leaving out responses under 100 ms.
- Higher is better
- Accuracy With a Happy Face%
- Share of trials with a happy face answered correctly, leaving out responses under 100 ms.
- Higher is better
- Accuracy With All Faces the Same%
- Share of all-neutral trials answered correctly, leaving out responses under 100 ms.
- Higher is better
- Mean Correct Response Timems
- Mean response time of correct answers on all test trials.
- Lower is faster
- Trials Used With an Angry Facetrials
- Correct trials with an angry face at or after 100 ms, the trials the effect uses.
- Fewer than 20 is flagged
- Trials Used With a Happy Facetrials
- Correct trials with a happy face at or after 100 ms, the trials the effect uses.
- Fewer than 20 is flagged
- Anticipationscount
- Responses under 100 ms from the crowd's onset.
- Lower is better
- Missed Responsescount
- Trials with no key press within 3,000 ms.
- Lower is better
- Total Trialstrials
- Trials in the scored run.
- Informational
Use and evidence
42 papers from 2021 to 2025 name this task in their title or abstract, down 28% compared with 2016 to 2020. That ranks it 268 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 5 |
| 2024 | 5 |
| 2023 | 10 |
| 2022 | 10 |
| 2021 | 12 |
| 2020 | 12 |
| 2019 | 15 |
| 2018 | 12 |
| 2017 | 7 |
| 2016 | 12 |
| 2015 | 15 |
| 2014 | 19 |
| 2013 | 18 |
| 2012 | 11 |
| 2011 | 15 |
| 2010 | 17 |
| 2009 | 13 |
| 2008 | 10 |
| 2007 | 3 |
| 2006 | 10 |
| 2005 | 6 |
| 2004 | 7 |
| 2003 | 5 |
| 2002 | 1 |
| 2001 | 4 |
| 2000 | 0 |
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
- 2
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 2 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Display
- Nine faces in three rows of three on a dark panel: white line drawings with a round outline, eyes, brows and a mouth.
- Faces
- Neutral: level brows and a straight mouth. Angry: brows tilted 20 degrees down toward the nose and a mouth that bends down. Happy: the same tilt and bend the other way.
- Trial mix
- 216 trials (standard; 144 to 288): a quarter with one angry face and a quarter with one happy face among neutral ones, each in each place equally often, and the rest all neutral.
- Keys
- F if all the faces are the same and J if one is different, shown under the faces on every trial.
- Fixation
- 500 ms.
- Faces on screen
- Until a response or 3,000 ms.
- Order
- A seeded random order, saved with the trial data.
- Practice
- 12 trials with feedback for 800 ms. No feedback in the test.
- Score
- Mean correct response time with a happy face minus with an angry face.
Hansen and Hansen (1988) used photographs of faces. This version uses schematic faces among neutral ones, as Öhman, Lundqvist and Esteves (2001) did, drawn so that the angry and the happy face change the neutral one by the same amount. No photographs of emotional expressions are used.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: face_in_the_crowd_test from the scored run, F for all the same or J for one different (chance 50%; half the trials have a different face) within 3,000 ms, correctness from the recorded key. Below chance invalidates and not above chance is flagged. One key on 90% of answers invalidates when accuracy is not above chance. Missed responses and responses under 100 ms follow the timed-test rule. Fewer than 20 correct trials with an angry or a happy face is flagged and withholds the percentile.
Why these rules
Accuracy significantly below chance (one-sided exact test, p < .05) cannot come from guessing; it means the response mapping was reversed or the task was not followed (Slick, Sherman and Iverson 1999 treat below-chance forced-choice performance as definite response bias). Chance-level accuracy can be a real result in impaired participants, so it is only flagged. One option on at least 90% of answered trials, when the correct options are balanced and accuracy is not above chance, means the answers did not depend on the items (the task counterpart of long-string responding in surveys; Curran 2016). Missed responses and responses under 100 ms keep the platform's timed-test rule: more than max(3, 25%) missed responses or more than max(2, 20%) responses under 100 ms invalidate, and any is flagged. A response under 100 ms cannot be a reaction to the stimulus. The effect is a difference between two means, so it needs enough correct trials with each emotion; the limit of 20 is a ConductScience choice.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Number of trials | 216 | 144 to 288 |
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
- Timed test: participants need a desktop or laptop with a mouse or trackpad. 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 Face-in-the-Crowd Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Hansen and Hansen, 1988; Öhman, Lundqvist and Esteves, 2001), with the standard settings (number of trials: 216). The main score was anger-superiority effect (mean correct response time with a happy face minus with an angry face).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Hansen, C. H., & Hansen, R. D. (1988). Finding the face in the crowd: An anger superiority effect. Journal of Personality and Social Psychology, 54(6), 917-924. https://doi.org/10.1037/0022-3514.54.6.917The face-in-the-crowd paradigm and the anger-superiority effect.
- Öhman, A., Lundqvist, D., & Esteves, F. (2001). The face in the crowd revisited: A threat advantage with schematic stimuli. Journal of Personality and Social Psychology, 80(3), 381-396. https://doi.org/10.1037/0022-3514.80.3.381The threat advantage with schematic faces.
- Fox, E., Lester, V., Russo, R., Bowles, R. J., Pichler, A., & Dutton, K. (2000). Facial expressions of emotion: Are angry faces detected more efficiently? Cognition and Emotion, 14(1), 61-92. https://doi.org/10.1080/026999300378996Schematic faces with same or different answers: angry faces among neutral ones were found faster than happy ones.
- Becker, D. V., Anderson, U. S., Mortensen, C. R., Neufeld, S. L., & Neel, R. (2011). The face in the crowd effect unconfounded: Happy faces, not angry faces, are more efficiently detected in single- and multiple-target visual search tasks. Journal of Experimental Psychology: General, 140(4), 637-659. https://doi.org/10.1037/a0024060With low-level differences removed, happy faces were found faster.
- Coelho, C. M., Cloete, S., & Wallis, G. (2010). The face-in-the-crowd effect: When angry faces are just cross(es). Journal of Vision, 10(1), 7. https://doi.org/10.1167/10.1.7Line features alone can speed the search for a schematic face.
- Shasteen, J. R., Sasson, N. J., & Pinkham, A. E. (2014). Eye tracking the face in the crowd task: Why are angry faces found more quickly? PLoS ONE, 9(4), e93914. https://doi.org/10.1371/journal.pone.0093914Nine faces in a 3 by 3 grid, with one key for all the same and one for a different face.
Source papers
- Öhman A, Lundqvist D, Esteves F (2001). The face in the crowd revisited: a threat advantage with schematic stimuli. Journal of Personality and Social Psychology, 80(3):381-396.Paradigm source, Adults
Questions
How long does the Face-in-the-Crowd Task take?
About 6 to 8 minutes, as listed in the test library.
What device does a participant need?
Timed test: participants need a desktop or laptop with a mouse or trackpad. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Anger-Superiority Effect, Anger-Superiority Effect, Accuracy, Anger-Superiority Effect, Medians, Response Time With an Angry Face, Response Time With a Happy Face, Response Time With All Faces the Same, Accuracy, Accuracy With an Angry Face, Accuracy With a Happy Face, Accuracy With All Faces the Same, Mean Correct Response Time, Trials Used With an Angry Face, Trials Used With a Happy Face, Anticipations, Missed Responses and Total Trials. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Face-in-the-Crowd Task used in research?
42 papers from 2021 to 2025 name it in their title or abstract, down 28% compared with 2016 to 2020. That ranks it 268 of 395 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.14.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Face-in-the-Crowd Task have?
In the published research, 0 papers report its test-retest reliability and 2 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?
Number of trials (144 to 288). 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.