Attention
Multitasking Panels Task
Also called: multi-attribute task battery, MATB, MATB-II, OpenMATB, multitasking battery, multiple attribute task battery
Measure how well people keep three tasks going at once: a drifting circle kept centered with the mouse while lights, gauges and fuel pumps need keys, at low and high workload.
11-14 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
- 131
- Evidence grade
- C
- Registered clinical studies
- 3
What it measures
Measures how well people keep three tasks going at once. Participants keep a drifting circle centered with the mouse while resetting lights and gauges and switching fuel pumps with keys, in parts at low and high workload. The main score is the mean share of all three kept in order.
Constructs
- Multitasking
- Divided attention
- Mental workload
Research uses
- Workload and human factors research
- Attention and fatigue research
- Within-study group comparison
How the paradigm works
The Multitasking Panels Task measures how well people keep several tasks going at the same time, in the form of the multi-attribute task battery of Comstock and Arnegard (1992) and its later versions (Santiago-Espada and colleagues, 2011; Cegarra and colleagues, 2020), with panels of our own and without the battery's radio task. Three panels run at once. On the left, a circle drifts away from the middle by itself, and the participant keeps it in a central ring with the mouse (compensatory tracking). On the right, a light that should stay green and one that should stay off, and four gauges whose pointers should stay in a middle band, are reset with keys when they change: Q and W for the lights, A, S, D and F for the gauges. Anything not reset goes back to normal after 10 seconds. Below them, two tanks that use fuel all the time are kept between their lines by switching four pumps with the number keys; a pump that fails comes back, switched off, after 20 seconds.
More about the paradigm
Workload is set by the rate of events, as in many studies of the battery (Pontiggia and colleagues, 2024): 3 events a minute at low workload and 12 at high, where the circle also drifts 1.5 times as fast. The standard session has 4 parts of 2 minutes, two at each workload in a balanced order drawn from the session's seed (low, high, high, low or the reverse), after a one-minute practice with all three panels and a summary, about 11 to 14 minutes in all. Every event is set before the session, and nothing a participant does changes a later event, the drift or the difficulty: there is no staircase and no training mode. The panels are one canvas redrawn on every screen frame. Desktop and laptop only, with a mouse and a keyboard.
The main score is the multitasking score: for each part, the mean of three shares (the time the circle was in the ring, the lights and gauges reset within 10 seconds, and the time both tanks were between their lines), averaged over the parts, in percent. Higher means all three tasks were kept in order more of the time. The score at each workload and their difference, the tracking error, reset times, misses and keys pressed when nothing was wrong, the tanks' distance from the middle of their band and the pump presses are also reported. Tracking depends on the mouse or trackpad and on the screen, so compare groups tested on similar equipment. Scores support research and do not diagnose.
Scores
- Multitasking Score%
- For each part, the mean of the time the circle was in the ring, the share of light and gauge events reset in time and the time both tanks were between their lines; averaged over the parts.
- Higher means all three tasks were kept in order more of the time
- Score at Low Workload%
- The multitasking score over the parts at low workload, 3 events a minute.
- Higher is better
- Score at High Workload%
- The multitasking score over the parts at high workload, 12 events a minute with a faster drift.
- Higher is better
- Workload Effect%
- Score at low workload minus score at high workload, in percentage points.
- Above 0 means the higher workload lowered the score
- Time in the Ring%
- Time with the circle's center within the middle ring, as a percent of the time tracked, over all parts.
- Higher is closer tracking
- Tracking Errorscore
- Root mean square distance of the circle from the middle, in half sides of the panel, weighted by time.
- Lower is closer tracking
- Resets in Time%
- Share of light and gauge events reset with their key within 10 seconds.
- Higher is better
- Reset Timems
- Median time from an event's start to its reset, for resets at or after 100 ms.
- Lower is faster
- Missed Eventscount
- Light and gauge events not reset within 10 seconds.
- Lower is better
- Unneeded Resetscount
- Light and gauge keys pressed when nothing was wrong.
- Lower is better
- Tanks Between Their Lines%
- Time each tank was between its lines, as a percent of the time tracked, over both tanks and all parts.
- Higher is better
- Tank Distance from the Middle%
- Mean distance of the tanks' levels from the middle of their band, in percent of a tank, weighted by time.
- Lower is better
- Pump Pressescount
- Number keys pressed to switch the pumps.
- Informational
- Presses on Failed Pumpscount
- Pump keys pressed while that pump had failed, which do nothing.
- Informational
- Parts Without Movementcount
- Parts in which the pointer did not move.
- Lower supports interpretation
- Pausescount
- Times a part paused because the pointer left the window, the tab was hidden or the window changed size.
- Lower supports interpretation
- Long Frames%
- Share of screen frames that took longer than 50 ms.
- Above 0.05 is flagged for review
- Anticipationscount
- Resets under 100 ms from an event's start, which cannot be reactions to it.
- Lower is better
- Completed Partscount
- Parts run in the scored run.
- Fewer than planned is informational
- Planned Partscount
- Parts the run set out to show.
- Informational
Use and evidence
131 papers from 2021 to 2025 name this task in their title or abstract, up 56% compared with 2016 to 2020. That ranks it 185 of 495 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 42 |
| 2024 | 32 |
| 2023 | 22 |
| 2022 | 18 |
| 2021 | 17 |
| 2020 | 14 |
| 2019 | 17 |
| 2018 | 20 |
| 2017 | 13 |
| 2016 | 20 |
| 2015 | 16 |
| 2014 | 5 |
| 2013 | 3 |
| 2012 | 5 |
| 2011 | 4 |
| 2010 | 3 |
| 2009 | 2 |
| 2008 | 2 |
| 2007 | 10 |
| 2006 | 2 |
| 2005 | 5 |
| 2004 | 2 |
| 2003 | 1 |
| 2002 | 2 |
| 2001 | 0 |
| 2000 | 4 |
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
- 3
- Papers on reliability or validity
- 62
- Registered clinical studies using it as an outcome
- 3
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 0.22.0, counted 3 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Panels
- Three panels on one screen, redrawn on every frame: tracking on the left, lights and gauges at the top right and tanks and pumps below them, in a frame of at most 1,200 by 760 px.
- Tracking
- A circle drifts by itself, driven by three slow sine waves on each axis (0.023 to 0.091 Hz) with phases drawn from the session's seed, and stays within 0.66 of the panel's half side; the pointer's place on the panel pushes it back. At every frame the task scores whether the circle is within the middle ring (0.22 of the half side) and its distance from the middle.
- Lights and gauges
- A light that should stay green (Q) and one that should stay off (W), and four gauges whose pointers should stay in the middle band (A, S, D, F). An event is reset with its key; one not reset goes back to normal after 10 seconds and counts as missed.
- Tanks and pumps
- Two tanks start at 50% and each uses 0.5% a second; their band runs from 40% to 60%. Pumps 1 and 2 fill them from the supply at 1% a second, and pumps 3 and 4 move 0.5% a second between them. A failed pump moves nothing and comes back switched off after 20 seconds.
- Workload
- 3 events a minute at low workload and 12 at high (gauges 2 and 8, lights 0.5 and 2, pump failures 0.5 and 2 a minute), with the drift 1.5 times as fast at high workload.
- Parts
- 2 parts at each workload (standard; 1 to 3) of 2 minutes each (standard; 2 to 5), in pairs whose order is drawn from the session's seed, with a rest between parts.
- Practice
- One minute at low workload with both lights, two gauges and a failed pump, then a summary of each panel.
- Fixed plan
- Every event is set before the session, and no response changes a later event, the drift or the difficulty: there is no staircase and no training mode.
- Score
- For each part, the mean of the time in the ring, the events reset in time and the time both tanks were between their lines, averaged over the parts, in percent.
The panels and numbers are ours, after the multi-attribute task battery and without its radio task: nothing of NASA's software, screens or scenario files is used, so published scores of the battery do not carry over. Tracking uses the participant's mouse or trackpad and depends on the screen's size and frame rate, so compare groups tested on similar equipment.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: the scored run's blocks. Each light or gauge event is a response that is reset in time or missed. Missed events are part of what the test measures, so they never invalidate on their own; a session that resets nothing is invalid. Responses under 100 ms follow the timed-test rule. A block with no pointer movement, frames over 50 ms on more than 5% of frames, or three or more pauses are flagged.
Why these rules
Missed targets are what this test measures, so missed responses never invalidate on their own. A session with no responses is invalid. An unbroken run of misses covering at least half the response trials (minimum 8) may mean the participant stopped, so the session is flagged and its percentile withheld. There is no correct option to compare with chance, so accuracy is not judged against chance. A reset under 100 ms cannot be a reaction to the event. Tracking is scored frame by frame, so a block with no pointer movement means the tracking panel was left alone, and many long frames mean the panels did not move smoothly on that computer.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Minutes per part | 2 | 2 to 5 |
| Parts at each workload | 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 Multitasking Panels Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Comstock and Arnegard, 1992), with the standard settings (minutes per part: 2; parts at each workload: 2). The main score was multitasking score (for each part, the mean of the time the circle was in the ring, the share of light and gauge events reset in time and the time both tanks were between their lines; averaged over the parts).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Comstock, J. R., & Arnegard, R. J. (1992). The multi-attribute task battery for human operator workload and strategic behavior research (NASA Technical Memorandum 104174). NASA Langley Research Center. The battery: two lights and four scales reset with keys, compensatory tracking, tanks kept near a target level with eight pumps, and a radio task, all on one screen.
- Santiago-Espada, Y., Myer, R. R., Latorella, K. A., & Comstock, J. R. (2011). The Multi-Attribute Task Battery II (MATB-II) software for human performance and workload research: A user's guide (NASA/TM-2011-217164). NASA Langley Research Center. The second version's defaults: scale events time out after 10 seconds and light events after 15, and tracking is recorded as the root mean square distance from the center.
- Cegarra, J., Valéry, B., Avril, E., Calmettes, C., & Navarro, J. (2020). OpenMATB: A Multi-Attribute Task Battery promoting task customization, software extensibility and experiment replicability. Behavior Research Methods, 52(5), 1980-1990. https://doi.org/10.3758/s13428-020-01364-wAn open version of the battery, built so that tasks can be changed and experiments repeated from their scenario files.
- Pontiggia, A., Gomez-Merino, D., Quiquempoix, M., Beauchamps, V., Boffet, A., Fabries, P., Chennaoui, M., & Sauvet, F. (2024). MATB for assessing different mental workload levels. Frontiers in Physiology, 15, 1408242. https://doi.org/10.3389/fphys.2024.1408242Nineteen studies that compared low and high workload; raising the rate of events was among the most used ways to set it, with medians of 3 events a minute at low and 23.5 at high workload.
Source papers
- Comstock JR, Arnegard RJ (1992). The multi-attribute task battery for human operator workload and strategic behavior research. NASA Technical Memorandum 104174. NASA Langley Research Center.Paradigm source, Adults
- Pontiggia A, Gomez-Merino D, Quiquempoix M, Beauchamps V, Boffet A, Fabries P, Chennaoui M, Sauvet F (2024). MATB for assessing different mental workload levels. Frontiers in Physiology, 15:1408242.Paradigm source, Adults
Questions
How long does the Multitasking Panels Task take?
About 11 to 14 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?
Multitasking Score, Score at Low Workload, Score at High Workload, Workload Effect, Time in the Ring, Tracking Error, Resets in Time, Reset Time, Missed Events, Unneeded Resets, Tanks Between Their Lines, Tank Distance from the Middle, Pump Presses, Presses on Failed Pumps, Parts Without Movement, Pauses, Long Frames, Anticipations, Completed Parts and Planned Parts. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Multitasking Panels Task used in research?
131 papers from 2021 to 2025 name it in their title or abstract, up 56% compared with 2016 to 2020. That ranks it 185 of 495 test families in the ConductCognition test census (use grade C). Source: ConductCognition test census 0.22.0, counted 3 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Multitasking Panels Task have?
In the published research, 3 papers report its test-retest reliability and 3 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?
Minutes per part (2 to 5) and parts at each workload (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.