Attention
Dot Grouping Task
Also called: perceptual grouping, grouping by proximity, grouping by similarity, Gestalt grouping, dot lattices, dot lattice task, attraction function, pure distance law
Measure grouping by proximity and similarity in brief dot lattices: how fast grouping falls with spacing and how much color or shape outweighs it.
7-9 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
- 230
- Evidence grade
- C
- Registered clinical studies
- 0
What it measures
Measures how vision groups dots into rows. A lattice of dots flashes for 300 ms, and participants choose which way its rows ran. The spacing differs between directions, and some lattices set spacing against color or shape. Scores how fast grouping falls as dots spread apart and how much spacing a color or shape difference outweighs.
Constructs
- Perceptual grouping
- Grouping by proximity
- Grouping by similarity
- Perceptual organization
Research uses
- Perceptual organization research
- Visual perception research
- Individual differences in Gestalt perception
How the paradigm works
The Dot Grouping Task measures how vision groups dots into rows, the Gestalt laws of grouping by proximity and by similarity made quantitative. Each trial shows a cross for 500 ms, then a lattice of dots inside a circle for 300 ms and three 200 ms masks of random dots, timed in whole screen frames. In a dot lattice the dots line up along four directions at once, along the two shortest spacings a and b and the two diagonals, so one pattern can be seen as rows running more than one way. Four icons then show those directions, each a circle with one diameter, in an order drawn for the trial, and the participant clicks the one that matches the rows they saw or presses its number, with no time limit.
More about the paradigm
Each round shows 31 lattices once, in a new order. In 12, the directions differ only in spacing, with b from 1 to 1.45 times a, across all six lattice types. In 16 rectangular lattices, the rows along a or along b are each made of one kind of dot, by color or by shape, so similarity works with proximity or against it. Three clear lattices with b twice a check that the icons were read. Each lattice is turned to a random angle and shifted, so the screen's edges never frame its rows. Four rounds (standard; 4 to 8) give 124 trials after 4 practice lattices with feedback.
Scores
- Attraction Slopescore
- Alpha of the additive grouping model, fitted by maximum likelihood to the usable answers: how much the log odds of seeing rows along a direction fall for each step of its spacing beyond the closest, in units of the closest (Kubovy, Holcombe and Wagemans, 1998). A small ridge keeps it finite when every answer follows the closest dots.
- Higher means grouping follows the closest dots more strongly; neither end is better
- Color Weightscore
- The fitted log odds that rows of one color add to a direction, from the color lattices.
- Above 0 means rows of one color drew answers; 0 means color added nothing
- Shape Weightscore
- The fitted log odds that rows of one shape add to a direction, from the shape lattices.
- Above 0 means rows of one shape drew answers; 0 means shape added nothing
- Color Balance Ratioscore
- 1 plus the color weight divided by the attraction slope: the ratio of a wider spacing to the closest at which rows of one color balance rows of closer dots. Null unless both are above 0.
- Higher means color outweighs more spacing
- Shape Balance Ratioscore
- 1 plus the shape weight divided by the attraction slope: the ratio of a wider spacing to the closest at which rows of one shape balance rows of closer dots. Null unless both are above 0.
- Higher means shape outweighs more spacing
- Closest Rows by Spacing Ratio%
- On the lattices that differ only in spacing, the share of answers along a direction of the closest spacing, keyed by b/a (1, 1.15, 1.3 and 1.45). At 1, two or three directions tie as closest.
- Rises as the closest spacing stands out more
- Color over Spacing%
- On color lattices whose one-color rows ran along the wider spacing, the share of answers that chose those rows.
- Higher means color won over spacing more often
- Shape over Spacing%
- On shape lattices whose one-shape rows ran along the wider spacing, the share of answers that chose those rows.
- Higher means shape won over spacing more often
- Clear Lattices%
- Share of the clear lattices, with b twice a, answered along the closest dots, where almost everyone sees the rows.
- Chance is 0.25; not above chance is flagged
- Median Answer Timems
- Median time from the icons appearing to the answer, on usable answers.
- Informational; the answer has no time limit
- Anticipationscount
- Answers given within 100 ms of the icons appearing, which leave the scores.
- More than max(2, 20%) is invalid
- Early Key Pressescount
- 1 to 4 keys pressed before the icons appeared, which are not taken as answers.
- Informational
- Frame Intervalms
- Median time between screen frames during the trials.
- About 16.7 on a 60 Hz screen
- Refresh RateHz
- Screen refresh rate from the frame interval.
- Informational
- Lattices on Time%
- Share of answered trials whose lattice ran within one frame of 300 ms; only these are scored.
- Under 0.9 is flagged for review
- Completed Trialstrials
- Trials answered in the scored run.
- Fewer than planned is flagged for review
- Planned Trialstrials
- Trials the run set out to show.
- Informational
Use and evidence
230 papers from 2021 to 2025 name this task in their title or abstract, down 20% compared with 2016 to 2020. That ranks it 148 of 711 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 65 |
| 2024 | 36 |
| 2023 | 36 |
| 2022 | 48 |
| 2021 | 45 |
| 2020 | 46 |
| 2019 | 50 |
| 2018 | 62 |
| 2017 | 67 |
| 2016 | 63 |
| 2015 | 74 |
| 2014 | 79 |
| 2013 | 66 |
| 2012 | 79 |
| 2011 | 73 |
| 2010 | 152 |
| 2009 | 52 |
| 2008 | 53 |
| 2007 | 51 |
| 2006 | 50 |
| 2005 | 50 |
| 2004 | 73 |
| 2003 | 44 |
| 2002 | 57 |
| 2001 | 38 |
| 2000 | 31 |
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
- 0
- Papers on reliability or validity
- 31
- 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: counts studies of grouping with any task, imaging included.
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 13 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Lattices
- 31 a round: 12 that differ only in spacing, with b 1, 1.15, 1.3 or 1.45 times a, across all six lattice types; 16 rectangular lattices whose rows along a or along b are each of one color or one shape; and 3 clear lattices with b twice a.
- Display
- A cross for 500 ms, then the lattice for 300 ms inside a 360 px circle, with dots 36 px apart at the closest, then three 200 ms masks of random dots, all timed in whole screen frames. Each lattice is turned to a random angle and shifted.
- Answer
- Four icons, a circle with one diameter for each direction of the lattice, in an order drawn for each trial. The participant clicks one or presses 1 to 4, with no time limit.
- Rounds
- 4 rounds (standard; 4 to 8), each lattice once a round in a new order: 124 trials, with a break halfway.
- Practice
- 4 clear lattices, each followed for 1,500 ms by the way its rows ran.
- Scores
- The choices are fitted by maximum likelihood to the additive model of Kubovy, Holcombe and Wagemans (1998) and Kubovy and van den Berg (2008): an attraction slope for spacing and a weight for color and for shape, with the spacing ratio at which each balances proximity.
The lattices are drawn in code, and the scores rest on ratios of distances, so no screen calibration is needed. Neither a stronger nor a weaker pull of the closest dots is 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: dot_grouping_test from the run with the most trials, ties going to the latest run; the answer is read from the saved key and icon order. Only the clear lattices (b/a of 2) have a right answer, along a (chance 1 in 4): below chance invalidates and not above chance is flagged; one icon position on 90% of answers invalidates when they are not above chance. A short run, fewer than 90% of lattices within one frame of 300 ms, or no fitted slope is flagged. Responses under 100 ms follow the timed-test rule.
Why these rules
On the clear lattices almost everyone sees rows along the closest dots, so answers there show whether the icons were read and the task followed. The icons' order is drawn for each trial, so a person who keeps choosing one position is not reporting the lattice. The 300 ms exposure is part of the design, so lattices shown off plan leave the score.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Rounds of lattices | 4 | 4 to 8 |
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 Dot Grouping Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Kubovy and Wagemans, 1995; Kubovy, Holcombe and Wagemans, 1998; Kubovy and van den Berg, 2008), with the standard settings (rounds of lattices: 4). The main score was attraction slope (alpha of the additive grouping model, fitted by maximum likelihood to the usable answers: how much the log odds of seeing rows along a direction fall for each step of its spacing beyond the closest, in units of the closest (Kubovy, Holcombe and Wagemans, 1998)).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Kubovy, M., & Wagemans, J. (1995). Grouping by proximity and multistability in dot lattices: A quantitative Gestalt theory. Psychological Science, 6(4), 225-234. https://doi.org/10.1111/j.1467-9280.1995.tb00597.xDot lattices of six types shown for 300 ms, with an answer screen of four orientations.
- Kubovy, M., Holcombe, A. O., & Wagemans, J. (1998). On the lawfulness of grouping by proximity. Cognitive Psychology, 35(1), 71-98. https://doi.org/10.1006/cogp.1997.0673The pure distance law: the odds of each organization fall with relative distance, whatever the lattice type; masks of random dots after the lattice.
- Kubovy, M., & van den Berg, M. (2008). The whole is equal to the sum of its parts: A probabilistic model of grouping by proximity and similarity in regular patterns. Psychological Review, 115(1), 131-154. https://doi.org/10.1037/0033-295X.115.1.131Proximity and similarity in rectangular lattices add in log odds.
- Wagemans, J., Elder, J. H., Kubovy, M., Palmer, S. E., Peterson, M. A., Singh, M., & von der Heydt, R. (2012). A century of Gestalt psychology in visual perception: I. Perceptual grouping and figure-ground organization. Psychological Bulletin, 138(6), 1172-1217. https://doi.org/10.1037/a0029333Review of the grouping principles and of how dot lattices measure them.
Source papers
- Kubovy M, Holcombe AO, Wagemans J (1998). On the lawfulness of grouping by proximity. Cognitive Psychology, 35(1):71-98.Paradigm source, Adults
Questions
How long does the Dot Grouping Task take?
About 7 to 9 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?
Attraction Slope, Color Weight, Shape Weight, Color Balance Ratio, Shape Balance Ratio, Closest Rows by Spacing Ratio, Color over Spacing, Shape over Spacing, Clear Lattices, Median Answer Time, Anticipations, Early Key Presses, Frame Interval, Refresh Rate, Lattices on Time, Completed Trials and Planned Trials. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Dot Grouping Task used in research?
230 papers from 2021 to 2025 name it in their title or abstract, down 20% compared with 2016 to 2020. That ranks it 148 of 711 test families in the ConductCognition test census (use grade C). Source: ConductCognition test census 0.58.0, counted 13 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Dot Grouping 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 C in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Rounds of lattices (4 to 8). 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.