Attention
Visual Illusions Test
Also called: optical illusion test, geometric illusions, illusion sensitivity task, Muller-Lyer illusion, Ebbinghaus illusion, Titchener circles, Ponzo illusion
Measure how strongly classic illusions change judged size: Muller-Lyer, Ebbinghaus and Ponzo figures, with staircases that find where two sizes look equal.
8-9 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
- 234
- Evidence grade
- C
- Registered clinical studies
- 1
What it measures
A measure of how strongly three classic illusions change judged size. In the Muller-Lyer, Ebbinghaus and Ponzo figures, participants choose the longer line or the larger circle, and staircases find the size at which the two look equal. The same pairs with nothing around them measure any side bias. A larger illusion is not a better score.
Constructs
- Size perception
- Context effects in perception
- Individual differences in illusion strength
Research uses
- Perception research
- Within-study group comparison
- Individual differences in illusion strength
How the paradigm works
The visual illusions test measures how much three classic figures change the judged size of a line or a circle. Each screen shows two lines or two circles, a standard and a comparison, and participants choose the line that looks longer or the circle that looks larger. In the Muller-Lyer figure the standard line has outward fins and the comparison line has inward fins. In the Ebbinghaus figure the standard circle is ringed by small circles and the comparison circle by large ones. In the Ponzo figure the two lines sit between converging lines, the standard near the narrow end. Each illusion is one part of the test.
More about the paradigm
A staircase changes the comparison's size after each answer until the two look equal (Levitt, 1971). Each figure has two staircases mixed together, one starting 25% below the standard and one 25% above, and the same pair with nothing around it has two more, which measure any side or position bias. The standard's side is random, and the picture stays until the answer, with no time limit. At 14 trials per staircase, with 4 plainly unequal plain pairs in each part, the standard session is 180 trials in about 8 minutes, after 6 practice pairs with feedback and no illusion. Figures are sized from the screen on every trial and the sizes drawn are saved in whole device pixels, so an illusion is a ratio that needs no calibration.
The main score is the mean across the illusions of the comparison's size at the point of equality, as a percent of the standard, minus the same point on the plain pair. Each illusion's percent, each plain pair's bias, how far apart the two staircases ended, accuracy on the plainly unequal pairs, trials used and response time are also reported. A larger illusion means the surrounding shapes changed judged size more; it is not better or worse performance. Scores support research and do not diagnose.
Scores
- Mean Illusion Percent%
- Mean across the illusions of the size at which the comparison looks equal to the standard, as a percent of the standard, minus the same point on the plain pair.
- Larger values mean a stronger illusion, which is not a better or worse score
- Muller-Lyer Illusion%
- Point of equality on the Muller-Lyer figure minus the point on the plain lines, as a percent of the standard line.
- Positive values mean the line with inward fins had to be longer to look equal
- Ebbinghaus Illusion%
- Point of equality on the Ebbinghaus figure minus the point on the plain circles, as a percent of the standard circle's diameter.
- Positive values mean the circle ringed by large circles had to be larger to look equal
- Ponzo Illusion%
- Point of equality on the Ponzo figure minus the point on the plain lines, as a percent of the standard line.
- Positive values mean the line at the wide end had to be longer to look equal
- Muller-Lyer Plain Pair Bias%
- Point of equality on the two lines with no fins, as a percent of the standard line.
- Near 0 means no bias between the two positions
- Ebbinghaus Plain Pair Bias%
- Point of equality on the two circles with no rings, as a percent of the standard circle's diameter.
- Near 0 means no bias between the two positions
- Ponzo Plain Pair Bias%
- Point of equality on the two lines with no converging lines, as a percent of the standard line.
- Near 0 means no bias between the two positions
- Muller-Lyer Staircase Gap%
- How far apart the two staircases on the Muller-Lyer figure ended, in percentage points of the standard.
- Smaller means the staircases agreed
- Ebbinghaus Staircase Gap%
- How far apart the two staircases on the Ebbinghaus figure ended, in percentage points of the standard.
- Smaller means the staircases agreed
- Ponzo Staircase Gap%
- How far apart the two staircases on the Ponzo figure ended, in percentage points of the standard.
- Smaller means the staircases agreed
- Muller-Lyer Plain Pair Staircase Gap%
- How far apart the two staircases on the plain lines of this part ended, in percentage points of the standard.
- Smaller means the staircases agreed
- Ebbinghaus Plain Pair Staircase Gap%
- How far apart the two staircases on the plain circles ended, in percentage points of the standard.
- Smaller means the staircases agreed
- Ponzo Plain Pair Staircase Gap%
- How far apart the two staircases on the plain lines of this part ended, in percentage points of the standard.
- Smaller means the staircases agreed
- Illusions Runcount
- Illusions in the scored run.
- Informational
- Illusions Scoredcount
- Illusions with a point of equality on both the figure and the plain pair. The main score averages these.
- Fewer than the illusions run is flagged for review
- Largest Staircase Gap%
- The largest gap between two staircases on one figure or plain pair.
- Over 20 percentage points is flagged for review
- Staircase Pairs Without a Point of Equalitycount
- Pairs of staircases in which at least one turned too few times for a point of equality.
- Any is flagged for review
- Fewest Reversalscount
- Reversals of the staircase that turned the fewest times.
- Informational
- Accuracy on Plainly Unequal Pairs%
- Share correct on the plain pairs that differ by 30%.
- Under 60% is invalid and under 75% is flagged for review
- Plainly Unequal Pairs Answeredtrials
- Plainly unequal pairs answered in the scored run.
- Informational
- Median Response Timems
- Median time from the picture's onset to the answer.
- Informational; the test has no time limit
- Anticipationscount
- Presses under 100 ms from the picture's onset. They are not taken as answers.
- Lower is better
- One Pixel as a Percent of the Standard%
- The largest size change one device pixel made on this screen.
- Over the staircase's 2.5% final step is flagged for review
- Planned Trialstrials
- Trials the session planned to show.
- Informational
- Total Trialstrials
- Trials in the scored run.
- Fewer than planned is flagged for review
Use and evidence
234 papers from 2021 to 2025 name this task in their title or abstract, up 4% compared with 2016 to 2020. That ranks it 107 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 57 |
| 2024 | 46 |
| 2023 | 42 |
| 2022 | 45 |
| 2021 | 44 |
| 2020 | 51 |
| 2019 | 41 |
| 2018 | 41 |
| 2017 | 54 |
| 2016 | 39 |
| 2015 | 50 |
| 2014 | 48 |
| 2013 | 45 |
| 2012 | 36 |
| 2011 | 30 |
| 2010 | 67 |
| 2009 | 38 |
| 2008 | 22 |
| 2007 | 31 |
| 2006 | 22 |
| 2005 | 26 |
| 2004 | 27 |
| 2003 | 13 |
| 2002 | 22 |
| 2001 | 29 |
| 2000 | 19 |
Use grade C
100 to 299 papers from 2021 to 2025.
Evidence grade C
At least 2 test-retest papers, 30 reliability or validity papers or 3 registered clinical studies.
- Papers on test-retest reliability
- 6
- Papers on reliability or validity
- 37
- Registered clinical studies using it as an outcome
- 1
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.
- Figures
- Muller-Lyer: two lines, one with outward fins and one with inward fins. Ebbinghaus: two circles, one ringed by 8 small circles and one by 6 large circles. Ponzo: two lines between converging lines, the standard near the narrow end.
- Plain pairs
- The same two lines or circles in the same places with nothing around them, which measure any bias between the two positions.
- Answer
- Click or tap a figure, or press an arrow key. The picture stays until the answer, with no time limit and 400 ms between trials. The standard takes each position on half of each staircase's trials.
- Staircases
- Four in each illusion, mixed at random: two on the figure and two on the plain pair, one starting 25% below the standard and one 25% above. Choosing the comparison makes it smaller on that staircase's next trial and choosing the standard makes it larger, in steps of 10%, then 5% after the first reversal and 2.5% after the third.
- Point of equality
- The mean of a staircase's reversals, leaving out the first and keeping an even count, then the mean of the two staircases on that figure. The illusion is this point on the figure minus this point on the plain pair.
- Trials
- 14 trials per staircase (standard; 12 to 24) and 4 plainly unequal plain pairs, 30% apart, in each illusion: 60 trials per illusion and 180 in all, in about 8 minutes. Each illusion is its own part, in a seeded order, with a rest between parts.
- Settings
- Any one or two of the illusions can run alone. The strength setting changes the angle between the Muller-Lyer fins (120, 60 or 30 degrees) and the size of the large Ebbinghaus circles (3, 3.5 or 4 against a center of 2.7). The Ponzo figure stays the same at every strength.
- Practice
- 6 plain pairs 30% apart, with feedback for 1,000 ms and no illusion.
- Screen
- The figure is sized from the screen on every trial, and the sizes drawn are saved in whole device pixels. An illusion is a ratio, so no calibration to centimeters is needed.
The fins follow Dewar (1967) and the rings follow Cretenoud et al. (2019). The Ponzo proportions are our own. Sources mostly used adjustment; this test uses two-choice staircases, with no time limit. A larger illusion is not a better score, so no percentile is reported.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: visual_illusions_test from the scored run, a choice between two figures with no time limit. Only the plainly unequal plain pairs have a correct answer (chance 50%): below chance invalidates and not above chance is flagged, under 60% correct on at least four of them invalidates and under 75% is flagged. One position on 90% of answers invalidates when accuracy on those pairs is not above chance. Answers under 200 ms on at least half the trials invalidate. A pair of staircases without a point of equality, two staircases on one figure ending more than 20 percentage points apart, a run short of its planned trials, or a screen where one pixel exceeds the 2.5% final step is flagged for review.
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). 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. The plainly unequal pairs differ by 30% with nothing around them, so errors on them mean the task was not followed. Two staircases starting from opposite sides should reach the same point of equality (Cornsweet, 1962), so a wide gap means the answers did not follow the sizes shown.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Trials per staircase | 14 | 12 to 24 |
| Illusions | All three (standard) | All three (standard), Muller-Lyer and Ebbinghaus, Muller-Lyer and Ponzo, Ebbinghaus and Ponzo, Muller-Lyer only, Ebbinghaus only, Ponzo only |
| Strength of the surrounding shapes | Standard | Weaker, Standard, Stronger |
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 Visual Illusions Test in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Coren, Girgus, Erlichman and Hakstian, 1976; Cretenoud et al., 2019), with the standard settings (trials per staircase: 14; illusions: All three; strength of the surrounding shapes: Standard). The main score was mean illusion percent (mean across the illusions of the size at which the comparison looks equal to the standard, as a percent of the standard, minus the same point on the plain pair).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Coren, S., Girgus, J. S., Erlichman, H., & Hakstian, A. R. (1976). An empirical taxonomy of visual illusions. Perception & Psychophysics, 20(2), 129-137. https://doi.org/10.3758/BF03199444
- Cretenoud, A. F., Karimpur, H., Grzeczkowski, L., Francis, G., Hamburger, K., & Herzog, M. H. (2019). Factors underlying visual illusions are illusion-specific but not feature-specific. Journal of Vision, 19(14), 12. https://doi.org/10.1167/19.14.12The sizes and distances of the Ebbinghaus rings, the control with no surrounding circles, and the size of the illusion as a percent of the reference.
- Dewar, R. E. (1967). Stimulus determinants of the magnitude of the Mueller-Lyer illusion. Perceptual and Motor Skills, 24(3), 708-710. https://doi.org/10.2466/pms.1967.24.3.708The length of the fins and the angles between them.
- Levitt, H. (1971). Transformed up-down methods in psychoacoustics. The Journal of the Acoustical Society of America, 49(2B), 467-477. https://doi.org/10.1121/1.1912375The up-down rule, the step sizes, the mixed staircases and the estimate from reversals.
- Cornsweet, T. N. (1962). The staircase-method in psychophysics. The American Journal of Psychology, 75(3), 485-491. https://doi.org/10.2307/1419876Two staircases mixed together, starting from opposite sides.
- Doherty, M. J., Campbell, N. M., Tsuji, H., & Phillips, W. A. (2010). The Ebbinghaus illusion deceives adults but not young children. Developmental Science, 13(5), 714-721. https://doi.org/10.1111/j.1467-7687.2009.00931.xA two-choice size judgment on the Ebbinghaus figure, with a control without the surrounding circles.
- Chouinard, P. A., Unwin, K. L., Landry, O., & Sperandio, I. (2016). Susceptibility to optical illusions varies as a function of the autism-spectrum quotient but not in ways predicted by local-global biases. Journal of Autism and Developmental Disorders, 46(6), 2224-2239. https://doi.org/10.1007/s10803-016-2753-1All three illusions in one session, with a comparison starting 20% to 50% away from the standard.
- Roberts, B., Harris, M. G., & Yates, T. A. (2005). The roles of inducer size and distance in the Ebbinghaus illusion (Titchener circles). Perception, 34(7), 847-856. https://doi.org/10.1068/p5273
- Makowski, D., Lau, Z. J., Pham, T., Paul Boyce, W., & Annabel Chen, S. H. (2021). A parametric framework to generate visual illusions using Python. Perception, 50(11), 950-965. https://doi.org/10.1177/03010066211057347
- Makowski, D., Te, A. S., Kirk, S., Liang, N. Z., & Chen, S. H. A. (2023). A novel visual illusion paradigm provides evidence for a general factor of illusion sensitivity and personality correlates. Scientific Reports, 13, 6594. https://doi.org/10.1038/s41598-023-33148-5Illusions set by parameters, with the real difference as the task's difficulty, in a browser task.
- Yildiz, G. Y., Sperandio, I., Kettle, C., & Chouinard, P. A. (2022). A review on various explanations of Ponzo-like illusions. Psychonomic Bulletin & Review, 29(2), 293-320. https://doi.org/10.3758/s13423-021-02007-7
Source papers
- Cretenoud AF, Karimpur H, Grzeczkowski L, Francis G, Hamburger K, Herzog MH (2019). Factors underlying visual illusions are illusion-specific but not feature-specific. Journal of Vision, 19(14):12.Paradigm source, Adults
Questions
How long does the Visual Illusions Test take?
About 8 to 9 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?
Mean Illusion Percent, Muller-Lyer Illusion, Ebbinghaus Illusion, Ponzo Illusion, Muller-Lyer Plain Pair Bias, Ebbinghaus Plain Pair Bias, Ponzo Plain Pair Bias, Muller-Lyer Staircase Gap, Ebbinghaus Staircase Gap, Ponzo Staircase Gap, Muller-Lyer Plain Pair Staircase Gap, Ebbinghaus Plain Pair Staircase Gap, Ponzo Plain Pair Staircase Gap, Illusions Run, Illusions Scored, Largest Staircase Gap, Staircase Pairs Without a Point of Equality, Fewest Reversals, Accuracy on Plainly Unequal Pairs, Plainly Unequal Pairs Answered, Median Response Time, Anticipations, One Pixel as a Percent of the Standard, Planned Trials and Total Trials. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Visual Illusions Test used in research?
234 papers from 2021 to 2025 name it in their title or abstract, up 4% compared with 2016 to 2020. That ranks it 107 of 395 test families in the ConductCognition test census (use grade C). Source: ConductCognition test census 0.14.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Visual Illusions Test have?
In the published research, 6 papers report its test-retest reliability and 1 registered clinical study uses it as an outcome (evidence grade C in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Trials per staircase (12 to 24), illusions (All three (standard), Muller-Lyer and Ebbinghaus, Muller-Lyer and Ponzo, Ebbinghaus and Ponzo, Muller-Lyer only, Ebbinghaus only, Ponzo only) and strength of the surrounding shapes (Weaker, Standard, Stronger). 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.