Social Cognition
Gaze Direction Task
Also called: cone of direct gaze, cone of gaze, eye contact detection, perceived gaze direction
Measure the range of eye directions that feel like eye contact: participants say whether faces looking straight ahead or slightly to one side are looking at them, scored as the width and center of that range.
5-7 minUntimed
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Try it here first, with no account.
- Papers, 2021 to 2025
- 98
- Evidence grade
- D
- Registered clinical studies
- 0
What it measures
Measures the range of eye directions that feel like eye contact. Faces look straight ahead or slightly to one side, and participants say whether each person is looking at them. Scores give the width and the center of that range.
Constructs
- Gaze perception
- Eye contact
- Social attention
Research uses
- Research on eye contact in social anxiety
- Research on gaze perception in autism and psychosis
- Research on social attention to faces
How the paradigm works
The cone of direct gaze is the range of eye directions that a person still feels as eye contact. Gamer and Hecht (2007) measured it by having observers turn the eyes of a virtual head until it looked at them, or until it stopped looking at them; across their experiments the cone was between 4 and 9 degrees wide, and a turned head pulled its center toward the head's direction. Lasagna and colleagues (2020) ran an online version in which people answered whether each face was looking at them.
More about the paradigm
This version shows London Set photographs of eight people, one woman and one man from each of the set's four recorded groups, at 13 gaze directions from 15 degrees to the left to 15 degrees to the right in steps of 2.5. Our code moves both irises sideways inside each eye opening by the amount a standard eye turns them for that angle (the iris 10 mm from the eye's center of turn, on a face whose pupils are 63 mm apart). Each face appears once at each direction, in a seeded order that never shows the same person twice in a row, for 104 judgments; the short form uses four of the people. Three practice faces come first. Answers are Yes or No, with no time limit, and the button order is counterbalanced. A setting adds each face with the head turned 10 degrees by our code, half of the people to each side.
The primary score is the cone width in degrees: the 2.5-degree step times the summed share of Yes answers across the 13 directions, which equals the distance between the two 50% edges of a symmetric response curve. The cone center is the share-weighted mean direction of Yes answers. With turned heads, the turned-head width and the shift of the center toward the side the head points are reported. A score needs two usable answers at every direction.
The degrees describe the eye turn drawn by our code, not a measured gaze angle, and the judgments describe the pictures, not the people pictured. Norms are off; compare people within a study.
Scores
- Cone widthdegrees
- Width of the range of gaze directions judged as looking at the participant, with the head straight: 2.5 degrees times the summed share of Yes answers over the 13 directions.
- Wider means more eye directions felt like eye contact.
- Cone centerdegrees
- Share-weighted mean direction of Yes answers with the head straight. Positive values lie toward the participant's right.
- Informational
- Straight faces judged looking%
- Percent of faces looking straight ahead that were judged as looking at the participant.
- Quality check
- Turned-head cone widthdegrees
- Cone width for faces with the head turned 10 degrees, when that setting is on.
- Informational
- Head-turn shiftdegrees
- Center of the turned-head cone, signed toward the side the head points. Positive values mean the range moves the way the head turns.
- Informational
- Judgments answeredcount
- Answers to faces shown on screen, excluding answers under 100 ms.
- Quality check
- Planned judgmentscount
- 52, 104 or 208 test judgments, by the faces and head-turn settings.
- Informational
- Median response timems
- Median answer time, excluding answers under 100 ms.
- Informational
- Anticipationscount
- Presses under 100 ms, including those the task ignores.
- Quality check
Use and evidence
98 papers from 2021 to 2025 name this task in their title or abstract, down 8% compared with 2016 to 2020. That ranks it 273 of 670 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 26 |
| 2024 | 12 |
| 2023 | 14 |
| 2022 | 25 |
| 2021 | 21 |
| 2020 | 17 |
| 2019 | 21 |
| 2018 | 21 |
| 2017 | 30 |
| 2016 | 17 |
| 2015 | 19 |
| 2014 | 13 |
| 2013 | 22 |
| 2012 | 20 |
| 2011 | 19 |
| 2010 | 21 |
| 2009 | 14 |
| 2008 | 12 |
| 2007 | 11 |
| 2006 | 10 |
| 2005 | 11 |
| 2004 | 7 |
| 2003 | 7 |
| 2002 | 4 |
| 2001 | 1 |
| 2000 | 5 |
Use grade D
30 to 99 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 1
- Papers on reliability or validity
- 17
- Registered clinical studies using it as an outcome
- 0
Measured on ConductCognition
Pooled across the studies that feed our platform norms, by device group. A figure needs 50 or more people.
No pooled figure yet.
Note on the counts: Yes or no judgments of London Set faces whose irises our code moves; degrees of drawn eye turn; judgments describe the pictures, not the people. Norms off..
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 0.48.0, counted 12 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Faces
- Neutral front London Set photographs of 8 people, one woman and one man from each of the set's four recorded groups, or 4 of them in the short form. All are grayscale, cropped to an oval and matched in luminance and contrast. A separate person appears in practice.
- Gaze
- 13 directions from 15 degrees left to 15 degrees right, in steps of 2.5. Our code moves both irises sideways inside each eye opening by the amount a standard eye turns them; the lids and eye corners stay put.
- Judgments
- 104 in the standard form: each face once at each direction, in a seeded order that never shows the same person twice in a row. The short form has 52. Three practice faces come first.
- Answer
- Yes or No to Is this person looking at you, by click, tap or the F and J keys, with no time limit. The button order is counterbalanced, and a 400 ms blank separates the faces.
- Turned heads
- A setting adds each face with the head turned 10 degrees by our code, half of the people to each side, at the same 13 directions. It doubles the judgments.
- Scoring
- Cone width is 2.5 degrees times the summed share of Yes answers over the 13 directions, which equals the distance between the two 50% edges of a symmetric curve. Cone center is the share-weighted mean direction. Each needs two usable answers at every direction.
The degrees describe the eye turn drawn by our code, not a measured gaze angle, and the judgments describe the pictures, not the people pictured. No answer is right or wrong. Norms are off. Scores support research and do not diagnose. The 8 test people and the practice person also appear in Face Memory Test (3), Face-Name Memory Task (5), Face Matching Task (9), Chimeric Faces Task (3), Part-Whole Face Task (6), Thatcher Illusion Task (1), Two-Tone Pictures Task (3), Verbal Overshadowing Task (5), Face Impressions Task (9), Real or AI Face Task (9), Uncanny Valley Task (5), Face Aftereffect Task (2), Group Faces Task (9). Counts include the other tasks' practice pictures and foils. Separate sessions or counterbalance task order when combining face tasks.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: gaze_direction_test from the scored run. Whether a face looks at the participant has no correct answer. An incomplete run, faces looking straight ahead judged as looking at the participant less than half the time, and a range wider than the levels shown are flagged. Answers under 200 ms on at least half the faces invalidate.
Why these rules
There is no correct option to compare with chance, so accuracy is not judged against chance. The width and center describe judgments within this session. 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 |
|---|---|---|
| Faces | 8 faces, 104 judgments (standard) | 8 faces, 104 judgments (standard), 4 faces, 52 judgments |
| Turned heads | Off | 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, tablets and phones. 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 Gaze Direction Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Gamer and Hecht, 2007), with the standard settings (faces: 8 faces, 104 judgments; turned heads: Off). The main score was cone width (width of the range of gaze directions judged as looking at the participant, with the head straight: 2.5 degrees times the summed share of Yes answers over the 13 directions).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Gamer, M., & Hecht, H. (2007). Are you looking at me? Measuring the cone of gaze. Journal of Experimental Psychology: Human Perception and Performance, 33(3), 705-715. https://doi.org/10.1037/0096-1523.33.3.705The cone of gaze, its width and center, and the 10-degree head turn.
- Gibson, J. J., & Pick, A. D. (1963). Perception of another person's looking behavior. The American Journal of Psychology, 76(3), 386-394. https://doi.org/10.2307/1419779Judgments of whether a looker is looking at the observer.
- Mareschal, I., Calder, A. J., & Clifford, C. W. G. (2013). Humans have an expectation that gaze is directed toward them. Current Biology, 23(8), 717-721. https://doi.org/10.1016/j.cub.2013.03.030Uncertain gaze is seen as directed at the observer.
- Lasagna, C. A., McLaughlin, M. M., Deng, W. Y., Whiting, E. L., & Tso, I. F. (2020). Deconstructing eye contact perception: Measuring perceptual precision and self-referential tendency using an online psychophysical eye contact detection task. PLOS ONE, 15(3), e0230258. https://doi.org/10.1371/journal.pone.0230258Online yes or no judgments of faces at graded gaze angles, with forward and turned heads.
Source papers
Questions
How long does the Gaze Direction Task take?
About 5 to 7 minutes, as listed in the test library.
What device does a participant need?
Untimed test: runs on desktops, laptops, tablets and phones. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Cone width, Cone center, Straight faces judged looking, Turned-head cone width, Head-turn shift, Judgments answered, Planned judgments, Median response time and Anticipations. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Gaze Direction Task used in research?
98 papers from 2021 to 2025 name it in their title or abstract, down 8% compared with 2016 to 2020. That ranks it 273 of 670 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.48.0, counted 12 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Gaze Direction Task have?
In the published research, 1 paper reports 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?
Faces (8 faces, 104 judgments (standard), 4 faces, 52 judgments) and turned heads (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.