Reasoning
Three Doors Task
Also called: Monty Hall dilemma, Monty Hall problem, three-door problem, three-door dilemma
Measure whether people learn to switch: after each pick of one of three doors, an empty door opens and participants stay or switch, over 60 rounds, scored as the switch rate.
5-7 minUntimed
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
- 45
- Evidence grade
- D
- Registered clinical studies
- 0
What it measures
Measures whether people learn to switch doors when switching pays. Points are behind one of three doors; after each pick the computer opens an empty door, and participants stay or switch. Switching wins two rounds in three. The main score is the switch rate over 60 rounds.
Constructs
- Probabilistic reasoning
- Learning from outcomes
- Choice under uncertainty
- Belief updating
Research uses
- Research on probabilistic reasoning and decision biases
- Studies of learning from repeated outcomes
- Within-study comparison of switching over rounds
How the paradigm works
The Three Doors Task measures whether people learn to switch when switching pays. It is the three-door dilemma posed by Selvin (1975), also known as the Monty Hall dilemma: points are behind one of three doors, the participant picks one, and the computer, which knows where the points are, opens one of the other two, always an empty one. Staying with the first pick wins only when that pick was right, one round in three; switching to the other closed door wins in the other two. Most people stay. Meeting the dilemma once, only 12% switched, and over repeated rounds switching rose to 55% (Granberg and Brown, 1995). Over 200 rounds with a point for each win, switching rose from about 57% to 66% of rounds while pigeons came to switch on 96% (Herbranson and Schroeder, 2010), and a systematic review found no repeated-round study in which experience led people to switch on every round (Saenen and colleagues, 2018). People learned to switch more without coming to understand why (Franco-Watkins, Derks and Dougherty, 2003; Saenen, Van Dooren and Onghena, 2015), and they dwelt more on losses after switching than after staying, which went with less learning (Petrocelli and Harris, 2011).
More about the paradigm
Three numbered doors are drawn by code. In each round the points (10) are placed behind a door by a draw from the session's seed before the pick, and the instructions say so, with the rule that the computer always opens an empty door that is neither the pick nor the one with the points. The participant picks a door by click or tap; 800 ms later the computer opens an empty door, chosen by a second draw when the pick hides the points; two buttons then offer to stay with the pick or switch to the other closed door, with no time limit. Every door then opens for 1,500 ms with a line saying where the points were. There are 60 rounds (30 to 200 in the settings), all scored, with no practice, so the first round is the dilemma met once. About 5 to 7 minutes.
The main score is the switch rate over every round. The first round's choice is reported alone, as are the switch rates over the first and the last third of the rounds and the change between them, how often switches and stays found the points, and how often a lost switch was followed by a stay and a lost stay by a switch. Switching more often earns more points, so a higher rate is the better score; staying is still the common choice, not a sign that the task was not followed. Scores support research and do not diagnose.
Scores
- Switch Rate%
- Share of rounds in which the participant switched to the other closed door.
- Higher earns more points: switching wins two rounds in three
- First-Round Switch0/1
- 1 when the first round was a switch, 0 when it was a stay: the dilemma met once, before any outcome.
- Informational
- Switch Rate, First Third%
- Share of switches over the first third of the planned rounds.
- Informational
- Switch Rate, Last Third%
- Share of switches over the last third of the planned rounds.
- Informational
- Change in Switch Rate%
- Switch rate over the last third of the rounds minus the first third.
- Above 0 means more switching with experience
- Switches That Won%
- Share of switches that found the points; by design about two in three.
- Informational: what the participant saw
- Stays That Won%
- Share of stays that found the points; by design about one in three.
- Informational: what the participant saw
- Stay After a Lost Switch%
- After a switch that found no points, the share of next rounds in which the participant stayed.
- Informational
- Switch After a Lost Stay%
- After a stay that found no points, the share of next rounds in which the participant switched.
- Informational
- Points Wonpoints
- Points over all rounds, 10 for each round in which the final door had them.
- Informational
- Median Choice Timems
- Median time from the empty door opening to the stay or switch choice, at or after 100 ms.
- Informational
- Anticipationscount
- Presses within 100 ms of a step appearing. They are not taken as answers.
- Quality check
- Rounds Playedcount
- Rounds in the scored run.
- Fewer than planned is flagged for review
- Planned Roundscount
- Rounds set for the session.
- Informational
Use and evidence
45 papers from 2021 to 2025 name this task in their title or abstract, up 25% compared with 2016 to 2020. That ranks it 312 of 495 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 17 |
| 2024 | 8 |
| 2023 | 8 |
| 2022 | 9 |
| 2021 | 3 |
| 2020 | 3 |
| 2019 | 7 |
| 2018 | 12 |
| 2017 | 7 |
| 2016 | 7 |
| 2015 | 12 |
| 2014 | 12 |
| 2013 | 13 |
| 2012 | 6 |
| 2011 | 11 |
| 2010 | 4 |
| 2009 | 8 |
| 2008 | 10 |
| 2007 | 8 |
| 2006 | 9 |
| 2005 | 6 |
| 2004 | 9 |
| 2003 | 11 |
| 2002 | 4 |
| 2001 | 4 |
| 2000 | 4 |
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
- 5
- Registered clinical studies using it as an outcome
- 0
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.
- Doors
- Three numbered doors drawn by code. In each round 10 points are placed behind one door by a draw from the session's seed before the pick, and the instructions say so.
- Each round
- The participant picks a door by click or tap. After 800 ms the computer opens one of the other two doors, always an empty one: never the pick and never the door with the points. When the pick hides the points, a second draw chooses which empty door opens. Two buttons then offer to stay with the pick or switch to the other closed door, with no time limit.
- Result
- Every door opens for 1,500 ms with a line saying where the points were, then a 400 ms blank. Presses under 100 ms are not taken.
- Rounds
- 60 (standard; 30 to 200), all scored, with no practice, so the first round is the dilemma met once. Points are hypothetical by default; a setting changes the instruction line to say the study pays.
- Scores
- The switch rate over every round, the first round's choice, the switch rates over the first and last third of the rounds and the change between them, how often switches and stays won, and what followed a lost switch and a lost stay.
Switching wins two rounds in three, so a higher switch rate earns more points. Published rates come from other displays, payments and numbers of rounds, and Herbranson and Schroeder (2010) never described the dilemma, while these instructions state the computer's rule, so rates are not directly comparable. No screen states the chances of the doors. 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: three_doors_round from the scored run, one stay or switch per round with no time limit. Switching wins more often, but staying is the choice most people make, so accuracy is not judged against chance and staying or switching on every round is accepted. A run short of its planned rounds is flagged for review. Choices under 200 ms on at least half the rounds invalidate.
Why these rules
There is no correct option to compare with chance, so accuracy is not judged against chance. Staying is the choice the task measures, not a mistake to screen out: most people stay when they first meet the three doors (Granberg and Brown, 1995), and switching rises only slowly over many rounds without reaching every round (Saenen and colleagues, 2018). 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.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Rounds | 60 | 30 to 200 |
| Payment line | Hypothetical points (standard) | Hypothetical points (standard), The study pays for points |
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.
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the Three Doors Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Selvin, 1975; Granberg & Brown, 1995; Herbranson & Schroeder, 2010), with the standard settings (rounds: 60; payment line: Hypothetical points). The main score was switch rate (share of rounds in which the participant switched to the other closed door).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Granberg, D., & Brown, T. A. (1995). The Monty Hall dilemma. Personality and Social Psychology Bulletin, 21(7), 711-723. https://doi.org/10.1177/014616729521700612% of people switched on first meeting the dilemma, and switching rose to 55% across repeated trials.
- Herbranson, W. T., & Schroeder, J. (2010). Are birds smarter than mathematicians? Pigeons (Columba livia) perform optimally on a version of the Monty Hall Dilemma. Journal of Comparative Psychology, 124(1), 1-13. https://doi.org/10.1037/a0017703200 rounds with a point for each win: people's switching rose from about 57% to 66% while pigeons came to switch on 96% of rounds.
- Klein, E. D., Evans, T. A., Schultz, N. B., & Beran, M. J. (2013). Learning how to "make a deal": Human (Homo sapiens) and monkey (Macaca mulatta) performance when repeatedly faced with the Monty Hall Dilemma. Journal of Comparative Psychology, 127(1), 103-108. https://doi.org/10.1037/a0029057Up to 500 rounds: about half of the people and about half of the monkeys came to favor switching.
- Petrocelli, J. V., & Harris, A. K. (2011). Learning inhibition in the Monty Hall problem: The role of dysfunctional counterfactual prescriptions. Personality and Social Psychology Bulletin, 37(10), 1297-1311. https://doi.org/10.1177/014616721141024560 rounds: people dwelt more on losses after switching than after staying, and those who did learned less.
- Saenen, L., Heyvaert, M., Van Dooren, W., Schaeken, W., & Onghena, P. (2018). Why humans fail in solving the Monty Hall dilemma: A systematic review. Psychologica Belgica, 58(1), 128-158. https://doi.org/10.5334/pb.274Across repeated-round studies, switching rose with experience but never to every round.
Source papers
- Granberg D, Brown TA (1995). The Monty Hall dilemma. Personality and Social Psychology Bulletin, 21(7):711-723.Paradigm source, Adults
- Herbranson WT, Schroeder J (2010). Are birds smarter than mathematicians? Pigeons (Columba livia) perform optimally on a version of the Monty Hall Dilemma. Journal of Comparative Psychology, 124(1):1-13.N 12, Students
Questions
How long does the Three Doors Task take?
About 5 to 7 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?
Switch Rate, First-Round Switch, Switch Rate, First Third, Switch Rate, Last Third, Change in Switch Rate, Switches That Won, Stays That Won, Stay After a Lost Switch, Switch After a Lost Stay, Points Won, Median Choice Time, Anticipations, Rounds Played and Planned Rounds. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Three Doors Task used in research?
45 papers from 2021 to 2025 name it in their title or abstract, up 25% compared with 2016 to 2020. That ranks it 312 of 495 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.22.0, counted 3 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Three Doors 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 D in the ConductCognition test census).
Which settings can a researcher change?
Rounds (30 to 200) and payment line (Hypothetical points (standard), The study pays for points). 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.