Attention
Takeover Task
Also called: take-over request task, takeover time task, takeover from automated driving
Measure how quickly and how well people take back control from a car that drives itself, with takeover time, time to collision and lane exceedances.
8-11 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
- 713
- Evidence grade
- C
- Registered clinical studies
- 0
What it measures
Measures how quickly and how well people take back control from a car that drives itself. While the car drives and they do a search task, a tone and a sign ask them to take over before a broken-down car in their lane. The main score is the takeover time.
Constructs
- Takeover from automation
- Attention switching
- Response timing
Research uses
- Human factors research
- Automated driving studies
- Within-study group comparison
How the paradigm works
The Takeover Task measures how quickly and how well people take back control from a car that drives itself. The car drives along a three-lane highway at 120 km/h behind another car while the participant does a search task on a screen under the road. At a request, a tone sounds and a red sign with a steering wheel appears, the car ahead moves aside, and a broken-down car with its hazard lights on stands in the lane. The participant takes over by braking or by steering into a free lane; the automation ends at the first brake or steering key. This follows the takeover design of Gold and colleagues (2013), in drawn scenes.
More about the paradigm
The standard session has a practice drive, then 6 drives with one request each, after 20 to 40 seconds of automated driving drawn from the session's seed. Half of the drives are in the right lane, where only the middle lane is free, and half in the middle lane. The time budget, the time to reach the broken-down car when the request comes, is 7 seconds as standard, or 5 seconds as a study setting. In the search task, the participant finds the one ring with its gap at the top among 12 and presses the number of its column. The scene pauses and blanks if the window loses focus. A short sound step comes first. Desktop or laptop with a keyboard.
Scores
- Takeover Timems
- Median time from the frame that showed the request to the first brake or steering key, at or after 100 ms, over drives with no pause after the request.
- Lower is faster; read with the avoided rate
- Takeovers%
- Drives with a brake or steering key after the request, divided by completed drives.
- Higher means fewer missed requests
- Broken-Down Cars Avoided%
- Drives that steered past or stopped before the broken-down car, divided by completed drives.
- Higher means more takeovers that kept clear
- Minimum Time to Collisionseconds
- Median over drives of the smallest gap divided by speed while the car was in the broken-down car's path after the request; 0 for a drive that would have reached it.
- Higher means a larger margin
- Lane Exceedancescount
- Drives in which the car's body crossed an outer edge line of the three lanes after the request.
- Lower means steadier lane control
- Braked First%
- Takeovers whose first key was the brake rather than steering.
- Describes the kind of response
- First Brake Timems
- Median time from the request to the first brake key, at or after 100 ms.
- Lower is faster
- First Steering Timems
- Median time from the request to the first steering key, at or after 100 ms.
- Lower is faster
- Search Ratecorrect/min
- Rings found in the search task per minute of automated driving before the requests.
- Higher means more engagement in the side task
- Search Accuracy%
- Correct column answers in the search task divided by all answers.
- Higher means more accurate searches
- Early Inputscount
- Brake or steering keys pressed before the requests, which do not change the drive.
- Lower supports interpretation
- Anticipationscount
- Takeovers under 100 ms after the request.
- Lower supports interpretation
- Pausescount
- Drive interruptions after focus loss, hiding the tab or resizing.
- Lower supports interpretation
- Long Frames%
- Share of animation frames longer than 50 ms.
- Above 0.05 is flagged
- Completed Drivescount
- Completed drives in the scored run.
- Informational
- Planned Drivescount
- Drives planned for the run.
- Informational
Use and evidence
713 papers from 2021 to 2025 name this task in their title or abstract, up 100% compared with 2016 to 2020. That ranks it 36 of 632 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 209 |
| 2024 | 145 |
| 2023 | 112 |
| 2022 | 127 |
| 2021 | 120 |
| 2020 | 98 |
| 2019 | 98 |
| 2018 | 81 |
| 2017 | 52 |
| 2016 | 27 |
| 2015 | 24 |
| 2014 | 5 |
| 2013 | 3 |
| 2012 | 4 |
| 2011 | 10 |
| 2010 | 5 |
| 2009 | 4 |
| 2008 | 5 |
| 2007 | 4 |
| 2006 | 5 |
| 2005 | 9 |
| 2004 | 6 |
| 2003 | 2 |
| 2002 | 2 |
| 2001 | 2 |
| 2000 | 0 |
Use grade B
300 to 999 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
- 85
- 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.
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 0.44.0, counted 12 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Scenes
- A drawn three-lane highway at 120 km/h, seen from behind the participant's car, with a car ahead and a broken-down car with its hazard lights on. No filmed clips, and no contact is ever drawn.
- Drives
- 6 drives as standard, with 3 to 12 available, each with one request after 20 to 40 seconds of automated driving drawn from the session seed. Half are in the right lane and half in the middle lane. A practice drive with feedback comes first.
- Request
- A tone of three 150 ms beeps and a red sign with a steering wheel. The car ahead moves aside and the broken-down car stands in the lane, 7 seconds away as standard or 5 seconds as a setting: the time budget.
- Side task
- While the car drives itself, participants find the one ring with its gap at the top among 12 and press its column, 1 to 4. A new search follows each answer.
- Response
- The first press of the down arrow (brake) or the left or right arrow (steering) ends the automation. Keys pressed before the request change nothing and count as early inputs.
- Scores
- Median takeover time, first brake and first steering times, takeovers, broken-down cars avoided, minimum time to collision, lane exceedances, search rate and accuracy, and frame timing.
Our drawn scenes, shorter drives and side task differ from the published simulator study. Scores describe takeovers in these scenes and do not establish fitness to drive. Read takeover time with the avoided rate and the minimum time to collision. No percentile output.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: the completed test drives of the scored run. A drive with no brake or steering key after the request is a missed response. Incomplete runs, frequent keys before requests, pauses and long frames are flagged; missed responses and responses under 100 ms follow the timed-test rule.
Why these rules
Missed responses and responses under 100 ms keep the platform's timed-test rule: more than max(3, 25%) missed responses or more than max(2, 20%) responses under 100 ms invalidate, and any is flagged. A response under 100 ms cannot be a reaction to the stimulus. The takeover time is the response to the request, so a drive with no response is a missed response. The car keeps its lane until the first key after the request, so a key pressed earlier changes nothing and counts only as an early input.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Drives | 6 | 3 to 12 |
| Time budget (seconds) | 7 seconds (standard) | 7 seconds (standard), 5 seconds |
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 Takeover Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Gold, Damböck, Lorenz and Bengler, 2013), with the standard settings (drives: 6; time budget (seconds): 7 seconds). The main score was takeover time (median time from the frame that showed the request to the first brake or steering key, at or after 100 ms, over drives with no pause after the request).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Gold, C., Damböck, D., Lorenz, L., & Bengler, K. (2013). "Take over!" How long does it take to get the driver back into the loop? Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 57(1), 1938-1942. https://doi.org/10.1177/1541931213571433Automated driving at 120 km/h, a request by tone and icon, time budgets of 5 and 7 seconds, a car ahead that uncovers the obstacle, and takeover by the first brake or steering input; our scenes and side task are original.
- Zhang, B., de Winter, J., Varotto, S., Happee, R., & Martens, M. (2019). Determinants of take-over time from automated driving: A meta-analysis of 129 studies. Transportation Research Part F: Traffic Psychology and Behaviour, 64, 285-307. https://doi.org/10.1016/j.trf.2019.04.020Takeover times across 129 studies: faster in more urgent situations and with auditory requests, slower with visual side tasks and handheld devices.
Source papers
- Gold, C., Damböck, D., Lorenz, L., & Bengler, K. (2013). "Take over!" How long does it take to get the driver back into the loop? Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 57(1), 1938-1942.Paradigm source, Adults
Questions
How long does the Takeover Task take?
About 8 to 11 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?
Takeover Time, Takeovers, Broken-Down Cars Avoided, Minimum Time to Collision, Lane Exceedances, Braked First, First Brake Time, First Steering Time, Search Rate, Search Accuracy, Early Inputs, Anticipations, Pauses, Long Frames, Completed Drives and Planned Drives. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Takeover Task used in research?
713 papers from 2021 to 2025 name it in their title or abstract, up 100% compared with 2016 to 2020. That ranks it 36 of 632 test families in the ConductCognition test census (use grade B). Source: ConductCognition test census 0.44.0, counted 12 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Takeover 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?
Drives (3 to 12) and time budget (seconds) (7 seconds (standard), 5 seconds). 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.