Attention
Seeing Causes Task
Also called: perceptual causality, phenomenal causality, causal perception, launching effect, launching events, impression of causality, causality perception, causal launching
Measure how the impression that one disc made another move falls with the delay and the gap between them, with yes or no answers and ratings.
8-10 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. Try it here first, with no account.
- Papers, 2021 to 2025
- 59
- Evidence grade
- D
- Registered clinical studies
- 2
What it measures
A measure of how the impression that one thing made another move depends on small delays and gaps. A yellow disc moves toward a blue disc, stops where it touches it or a little short of it, and the blue disc moves off after a short delay. After each display, participants say yes or no, or rate from 1 to 7, whether the first disc made the second one move. The scores are the delay and the gap at which the impression has fallen halfway. A plain launch and a clearly separate display check the answers. A stronger or weaker impression is not a better score.
Constructs
- Perceptual causality
- Causal perception
- Visual motion perception
- Temporal and spatial integration
Research uses
- Visual perception research
- Social perception research
- Individual differences in causal perception
Scores
- Delay at Half Impressionms
- The delay at which the impression of causing has fallen halfway from its value at no delay to its lowest value, over the delays of 0 to 400 ms at no gap and the standard speed, by linear interpolation between delays. The impression is the mean of the yes share and the rating, each put on a scale of 0 to 1.
- Informational; null when the series falls by less than 0.2 or never reaches halfway
- Gap at Half Impressioncanvas units
- The gap, in screen units, at which the impression has fallen halfway, over the gaps of 0 to 96 at no delay and the standard speed.
- Informational; null when the series falls by less than 0.2 or never reaches halfway
- Fall Over Delay%
- The impression at no delay less the impression at 400 ms, on the scale of 0 to 1.
- Informational; larger means a faster fall
- Fall Over Gap%
- The impression at no gap less the impression at a gap of 96 screen units, on the scale of 0 to 1.
- Informational; larger means a faster fall
- Impression of the Plain Launch%
- The impression for the display with no gap, no delay and the standard speed, on the scale of 0 to 1.
- Most viewers see a push in a plain launch; neither a higher nor a lower value is better
- Impression of the Control%
- The impression for the control, in which the first disc stops 150 screen units away and the second moves after 600 ms.
- Low for most viewers; the plain launch less this is checked
- Yes Share, Plain Launch%
- Share of yes answers to the plain launch in the yes or no part.
- Informational
- Rating, Plain Launchscore
- Mean rating, 1 to 7, of the plain launch in the rating part.
- Informational
- Yes Share, Control%
- Share of yes answers to the control in the yes or no part.
- Informational
- Rating, Controlscore
- Mean rating, 1 to 7, of the control in the rating part.
- Informational
- Yes Share%
- Share of yes answers over every display but the control, in the yes or no part.
- Informational; the same answer on nearly every display is flagged
- Mean Ratingscore
- Mean rating, 1 to 7, over every display but the control, in the rating part.
- Informational
- Effect of Speed%
- The impression at the fast speed less the impression at the slow speed, averaged over the gaps above 0 at no delay.
- Positive means faster discs gave a stronger impression across gaps
- Delays on Time%
- Share of displays in which the delay as shown was within one frame of the delay planned.
- Under 0.9 is flagged for review
- Frame Intervalms
- Median time between screen frames during the displays.
- About 16.7 on a 60 Hz screen
- Dropped Frame Rate%
- Share of screen frames the browser skipped during the displays.
- Over 0.05 is flagged
- Completed Displaystrials
- Displays with a usable answer.
- Fewer than planned is flagged for review
- Planned Displaystrials
- Displays the run set out to show, in both parts.
- Informational
Use and evidence
59 papers from 2021 to 2025 name this task in their title or abstract, up 9% compared with 2016 to 2020. That ranks it 417 of 711 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 22 |
| 2024 | 16 |
| 2023 | 8 |
| 2022 | 2 |
| 2021 | 11 |
| 2020 | 14 |
| 2019 | 6 |
| 2018 | 14 |
| 2017 | 7 |
| 2016 | 13 |
| 2015 | 13 |
| 2014 | 8 |
| 2013 | 8 |
| 2012 | 17 |
| 2011 | 15 |
| 2010 | 12 |
| 2009 | 8 |
| 2008 | 12 |
| 2007 | 7 |
| 2006 | 7 |
| 2005 | 12 |
| 2004 | 11 |
| 2003 | 7 |
| 2002 | 9 |
| 2001 | 4 |
| 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
- 0
- Papers on reliability or validity
- 21
- Registered clinical studies using it as an outcome
- 2
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: a few counted works use the words outside perception.
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 0.58.0, counted 16 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Display
- A yellow disc moves along a line toward a blue disc on a dark 640 by 220 panel, stops where its edge touches the blue disc or short of it, and after a delay the blue disc moves 140 screen units at the same speed. Every position is drawn from the time since the display's first frame.
- Displays
- 22: seven delays of 0, 33, 67, 100, 150, 250 and 400 ms with no gap, five gaps of 0, 12, 24, 48 and 96 screen units with no delay at three speeds, 200, 320 and 480 screen units a second, and a control with a gap of 150 and a delay of 600 ms.
- Parts
- Two parts with the same displays: yes or no to whether the first disc made the second one move, and a rating from 1 (not at all) to 7 (very strongly) of how strongly. Which part comes first is drawn at random. Each part shows the 22 displays 2 times (standard; 1 to 3), each time through in a new order: 88 displays.
- Practice
- Two displays, the plain launch and the control, answered as the first part is, with no feedback.
- Timing
- Each display runs on screen frames: 500 ms at rest, the first disc's move, the delay, the second disc's move and 400 ms at rest. The delay as shown is saved and compared with the delay planned.
- Scores
- The delay and the gap at which the impression has fallen halfway, the fall over each series, the impression of the plain launch and of the control, the share of yes answers, the mean rating and the effect of speed.
The discs are drawn in code, and the scores are proportions and ratings on the participant's own screen, so no calibration is needed. A stronger or weaker impression of causing is not a better score. 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: seeing_causes_answer from the scored run, one per display, with the display's timing from seeing_causes_view. An impression of causing is what the participant saw, so it is never right or wrong: accuracy is not judged against chance, and yes or no on any display is a valid answer. The check is the plain launch against the control, a first disc that stops far from a second one that moves after a long wait: when the launch's impression is not at least 0.2 above the control's, the answers are flagged for review. One yes or no answer on 95% or more of at least 20 displays, a run short of its planned displays, delays that were not within a frame of the plan on more than 10% of displays, more than 5% of frames dropped, and a screen below 50 Hz are flagged for review.
Why these rules
There is no correct option to compare with chance, so accuracy is not judged against chance. Whether one disc seems to push another depends on delays of a few frames and gaps of a few pixels (Straube and Chatterjee, 2010; White, 2025), and some people see launches with delays or gaps that others do not, so a long or short half point is a result, not an error. A plain launch looks like a push to most viewers and a display with a far gap and a long wait does not (White, 2025), so answers that do not separate them may not have followed the displays. The score rests on the delays as drawn, so a screen that cannot hold them is flagged, and a hidden tab withholds the percentile.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Times through the displays in each part | 2 | 1 to 3 |
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 Seeing Causes Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Straube and Chatterjee, 2010; Meding, Bruijns, Schölkopf, Berens and Wichmann, 2020; White, 2025), with the standard settings (times through the displays in each part: 2). The main score was delay at half impression (the delay at which the impression of causing has fallen halfway from its value at no delay to its lowest value, over the delays of 0 to 400 ms at no gap and the standard speed, by linear interpolation between delays).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Straube, B., & Chatterjee, A. (2010). Space and time in perceptual causality. Frontiers in Human Neuroscience, 4, 28. https://doi.org/10.3389/fnhum.2010.00028Yes or no judgments of whether one ball caused another to move, with delays of 0 to 267 ms: events judged causal had shorter delays.
- Meding, K., Bruijns, S. A., Schölkopf, B., Berens, P., & Wichmann, F. A. (2020). Phenomenal causality and sensory realism. i-Perception, 11(3), 2041669520927038. https://doi.org/10.1177/2041669520927038Ratings from 1 to 9 of disc launches with delays of 0 to 400 ms: the impression falls with delay and levels off at the longest delays.
- White, P. A. (2025). Michotte's research on perceptual impressions of causality: a registered replication study. Royal Society Open Science, 12(9), 250244. https://doi.org/10.1098/rsos.250244Launching ratings against delay and against gap at three speeds: they fall quickly over small delays and gaps and do not reach zero.
Source papers
- Straube B, Chatterjee A (2010). Space and time in perceptual causality. Frontiers in Human Neuroscience, 4:28.N 16, Adults
Questions
How long does the Seeing Causes Task take?
About 8 to 10 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?
Delay at Half Impression, Gap at Half Impression, Fall Over Delay, Fall Over Gap, Impression of the Plain Launch, Impression of the Control, Yes Share, Plain Launch, Rating, Plain Launch, Yes Share, Control, Rating, Control, Yes Share, Mean Rating, Effect of Speed, Delays on Time, Frame Interval, Dropped Frame Rate, Completed Displays and Planned Displays. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Seeing Causes Task used in research?
59 papers from 2021 to 2025 name it in their title or abstract, up 9% compared with 2016 to 2020. That ranks it 417 of 711 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.58.0, counted 16 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Seeing Causes 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?
Times through the displays in each part (1 to 3). 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.