Executive Function
Planning Fallacy Task
Also called: planning fallacy paradigm, task duration prediction, completion time prediction, time prediction bias, prediction of performance time
Predict how long a set of picture puzzles will take, then do them with no clock on screen. Measures how far the prediction misses the time the puzzles took.
6-9 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
- 70
- Evidence grade
- D
- Registered clinical studies
- 0
What it measures
Measures how well people predict their own time on a task. After a practice picture puzzle, participants predict how long a set of the same puzzles will take, do the set with no clock on screen, then say how long it took. The main score is predicted time over actual time.
Constructs
- Time prediction
- Planning
- Memory for duration
- Optimism in prediction
Research uses
- Judgment and decision making research
- Research on planning and time management
- Research on memory for durations
- Within-study group comparison
How the paradigm works
The Planning Fallacy Task measures how well people predict how long their own work will take. Buehler, Griffin and Ross (1994) describe the planning fallacy as a confident belief that one's own project will go as planned, even while knowing that most similar projects have run late. Many studies of it follow real projects or ask project managers; this task is the laboratory form, in which a person predicts a short task and then does it while the time is measured, as in Roy and Christenfeld (2007, 2008).
More about the paradigm
The task is a set of picture puzzles. Each puzzle is a picture cut into 3 by 3 tiles that start in the wrong places; tapping or clicking two tiles swaps them, and a puzzle is finished when every tile is back in its place. A dark frame around each picture gives every tile a look of its own, so each puzzle has one solution that can be seen. After one practice puzzle, the participant reads that a set of the same puzzles comes next, 10 in the standard session (6 to 16), and types how long the whole set will take, in minutes and seconds. The puzzles then run one after another with no clock on screen, each needing 7 swaps, and the session records the time from the start of the first puzzle to the end of the last. Afterwards the participant types how long the set took, as in Roy and Christenfeld's memory condition. A puzzle not finished in 3 minutes ends unsolved.
The main score is the prediction ratio: the predicted time divided by the actual time, so 1 is an exact prediction, values below 1 are underestimates and values above 1 overestimates. Its base-10 logarithm, the log proportional error of Roy and Christenfeld, treats an underestimate and an overestimate by the same factor alike. The remembered time is reported the same way, with whether the set finished within the predicted time and the practice puzzle's time.
The direction of the error depends on the task. Buehler, Griffin and Ross found that students took longer than they predicted on their theses and assignments. In the laboratory, Roy and Christenfeld found that people overestimated a task new to them, that after some experience their predictions and their memories of the time both tended to fall short, and that tasks under about 2 minutes were overestimated and longer ones underestimated. Halkjelsvik and Jørgensen (2012) found underestimation the more frequent result in engineering and management studies but not in psychology studies. Here the practice puzzle gives some experience before the prediction, and the standard set takes several minutes. Scores support research and do not diagnose.
Scores
- Prediction Ratioscore
- Predicted time for the set divided by the time it took.
- 1 is exact; below 1 an underestimate, above 1 an overestimate; no better direction
- Log Prediction Ratioscore
- Base-10 logarithm of the prediction ratio (log proportional error).
- 0 is exact; negative an underestimate, positive an overestimate
- Predicted Timeseconds
- The time typed before the set, from the start of the first puzzle to the end of the last.
- Informational
- Actual Timeseconds
- Time from the start of the first puzzle to the end of the last.
- Informational
- Remembered Timeseconds
- The time typed after the set for how long it took.
- Informational
- Remembered Ratioscore
- Remembered time divided by the time the set took.
- 1 is exact; below 1 the set is remembered as shorter than it was
- Finished Within Prediction0/1
- 1 when the set took no longer than predicted, else 0.
- Read across a group as a share
- Practice Puzzle Timeseconds
- Time from the start of the practice puzzle to its last swap.
- The experience a prediction can draw on
- Median Puzzle Timeseconds
- Median time from a puzzle's start to its last swap, over the set.
- Informational
- Extra Swapscount
- Swaps beyond the fewest needed, over the set.
- Context for the time
- Puzzles Solvedcount
- Puzzles of the set finished before the time limit.
- Informational
- Puzzles Timed Outcount
- Puzzles of the set that reached the 3-minute limit.
- Lower is better; any is flagged
- Time Hiddenseconds
- Time the page was hidden during the set; it counts in the actual time.
- Any is flagged
- Puzzles Recordedcount
- Puzzles of the set recorded in the scored run.
- Informational
Use and evidence
70 papers from 2021 to 2025 name this task in their title or abstract, up 21% compared with 2016 to 2020. That ranks it 263 of 495 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 17 |
| 2024 | 19 |
| 2023 | 13 |
| 2022 | 7 |
| 2021 | 14 |
| 2020 | 11 |
| 2019 | 13 |
| 2018 | 6 |
| 2017 | 13 |
| 2016 | 15 |
| 2015 | 18 |
| 2014 | 13 |
| 2013 | 6 |
| 2012 | 9 |
| 2011 | 8 |
| 2010 | 11 |
| 2009 | 12 |
| 2008 | 6 |
| 2007 | 5 |
| 2006 | 2 |
| 2005 | 4 |
| 2004 | 5 |
| 2003 | 2 |
| 2002 | 5 |
| 2001 | 1 |
| 2000 | 2 |
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
- 13
- Registered clinical studies using it as an outcome
- 0
Note on the counts: includes project management papers that run no task.
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.
- Puzzles
- Each puzzle is a Noto Emoji picture in a dark frame, cut into 3 by 3 tiles that start in the wrong places and need 7 swaps to solve. Tapping or clicking one tile and then another swaps them. The frame gives each tile a look of its own, so every puzzle has one solution that can be seen. A puzzle not finished in 3 minutes ends unsolved.
- Practice
- One puzzle needing 5 swaps comes before the prediction, so the prediction can draw on some experience of the puzzles. Its time is recorded and never shown.
- Prediction
- Typed in minutes and seconds: how long the set will take, from the start of the first puzzle to the end of the last. The set has 10 puzzles (standard; 6 to 16).
- The set
- The puzzles run one after another, each starting as the one before ends, with no clock on screen. Every swap is recorded with its time, and so is any time the page is hidden.
- Memory
- After the last puzzle, the participant types how long the set took.
- Scores
- The prediction ratio: predicted time over actual time, below 1 for an underestimate and above 1 for an overestimate. Also its base-10 logarithm, the remembered time over actual time, and whether the set finished within the prediction.
Many papers counted under the planning fallacy follow real projects or ask project managers and run no task. This is the laboratory form: a short task predicted and then timed. Whether people under- or overestimate depends on how long the task is and on their experience of it, so compare groups or conditions rather than reading one person's ratio. The puzzles, the frame and the start orders are our own. These are research measures: no result calls anyone a poor planner, and they do not diagnose.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: planning_fallacy_puzzle from the scored run, one per puzzle, with that run's planning_fallacy_prediction and planning_fallacy_memory rows. A prediction has no right answer and every puzzle can be solved, so accuracy is not judged against chance. No prediction invalidates. First taps under 200 ms on at least half the puzzles invalidate. A prediction under 2 seconds or over 10 minutes a puzzle, a puzzle that reached the time limit, fewer puzzles than planned and a page hidden during the set are flagged; the last two withhold the percentile.
Why these rules
There is no correct option to compare with chance, so accuracy is not judged against chance. The score compares a prediction with the time the set took, so it needs both, and the set's time is known only when every puzzle was recorded. A puzzle stopped by the time limit, or a page left hidden, adds time the person did not spend on the puzzles. A prediction far outside any time the puzzles could take is unlikely to be an estimate; the limits are ConductScience choices. 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.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Puzzles in the set | 10 | 6 to 16 |
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.
- 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 Planning Fallacy Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Buehler R, Griffin D, Ross M, 1994; Roy MM, Christenfeld NJS, 2007), with the standard settings (puzzles in the set: 10). The main score was prediction ratio (predicted time for the set divided by the time it took).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Buehler, R., Griffin, D., & Ross, M. (1994). Exploring the "planning fallacy": Why people underestimate their task completion times. Journal of Personality and Social Psychology, 67(3), 366-381. https://doi.org/10.1037/0022-3514.67.3.366The planning fallacy: predictions of one's own completion times that are too optimistic.
- Roy, M. M., & Christenfeld, N. J. S. (2007). Bias in memory predicts bias in estimation of future task duration. Memory & Cognition, 35(3), 557-564. https://doi.org/10.3758/BF03193294A laboratory task predicted in minutes and seconds and then timed, the memory estimate after it, and the log proportional error.
- Roy, M. M., & Christenfeld, N. J. S. (2008). Effect of task length on remembered and predicted duration. Psychonomic Bulletin & Review, 15(1), 202-207. https://doi.org/10.3758/PBR.15.1.202Short tasks overestimated and longer ones underestimated, in prediction and in memory.
- Roy, M. M., Christenfeld, N. J. S., & McKenzie, C. R. M. (2005). Underestimating the duration of future events: Memory incorrectly used or memory bias? Psychological Bulletin, 131(5), 738-756. https://doi.org/10.1037/0033-2909.131.5.738Predictions rest on memories of past durations, which run short.
- Halkjelsvik, T., & Jørgensen, M. (2012). From origami to software development: A review of studies on judgment-based predictions of performance time. Psychological Bulletin, 138(2), 238-271. https://doi.org/10.1037/a0025996A review of time predictions in psychology, engineering and management studies.
Source papers
- Buehler R, Griffin D, Ross M (1994). Exploring the "planning fallacy": Why people underestimate their task completion times. Journal of Personality and Social Psychology, 67(3):366-381.Paradigm source
- Roy MM, Christenfeld NJS (2007). Bias in memory predicts bias in estimation of future task duration. Memory & Cognition, 35(3):557-564.Paradigm source
Questions
How long does the Planning Fallacy Task take?
About 6 to 9 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?
Prediction Ratio, Log Prediction Ratio, Predicted Time, Actual Time, Remembered Time, Remembered Ratio, Finished Within Prediction, Practice Puzzle Time, Median Puzzle Time, Extra Swaps, Puzzles Solved, Puzzles Timed Out, Time Hidden and Puzzles Recorded. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Planning Fallacy Task used in research?
70 papers from 2021 to 2025 name it in their title or abstract, up 21% compared with 2016 to 2020. That ranks it 263 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 Planning Fallacy 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).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Puzzles in the set (6 to 16). 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.