Decision Making
Patch Foraging Task
Also called: patch-leaving task, apple foraging task
Measure when people leave a source that is running out: they harvest apple trees or walk to new ones in orchards with short and long walks, scored against the marginal value theorem.
9-11 minUntimed
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Preview any test free after you sign up.
- Papers, 2021 to 2025
- 356
- Evidence grade
- C
- Registered clinical studies
- 7
What it measures
Measures when people leave a source that is running out: they harvest apple trees whose yield shrinks or walk to a new tree, in orchards with short and long walks.
Constructs
- Foraging decisions
- Opportunity cost of time
- Exploration and exploitation
- Reward rate learning
Research uses
- Decision-making research
- Motivation and apathy studies
- Computational psychiatry batteries
- Within-study group comparison
How the paradigm works
The patch foraging task measures how people decide when to leave a source of reward that is running out for a fresh one that takes time to reach. On each turn the participant harvests the apple tree in front of them, which takes 3 seconds, or walks to a new tree. Each harvest of the same tree tends to give fewer apples, and walking takes 6 seconds in one orchard and 9 seconds in the other. Each orchard lasts the same fixed time, and the time taken to answer counts toward the harvest or the walk, so answering faster gains nothing.
More about the paradigm
The task follows experiment 2 of Constantino and Daw (2015). A new tree's first harvest gives a normal draw with mean 10 and SD 1 apples, and after each harvest the tree keeps a share of its yield drawn with mean 0.88 and SD 0.07 (0.94 with shallow depletion). Every tree comes from its own stream of the session's seed, and every yield, share kept and delay is saved. The standard 4 minutes per orchard, after a 45 second practice, keep the task near 10 minutes; the paper used 14 minutes per orchard. An answer later than 2 seconds costs a 3 second pause with no apples. The orchards come in an order drawn from the session's seed.
The marginal value theorem (Charnov, 1976) says that the best rule leaves a tree once the expected next harvest falls below the long-run reward rate times the harvest time. That threshold is computed for each orchard from its settings: 5.72 apples with 6 second walks and 5.04 with 9 second walks at the standard settings. As in the paper, the threshold a participant used is estimated from the mean of the last two harvests before each departure, here expressed as the expected next harvest. The main score is overharvesting: the theorem's threshold minus the participant's, averaged over the two orchards, so positive values mean staying at trees too long. Scores also include the threshold in each orchard, its change between orchards (the theorem predicts leaving earlier when walks are short), harvests per tree, the apples collected and the median answer time. In the sources, people moved their thresholds in the direction the theorem predicts but tended to stay longer than it prescribes (Constantino and Daw, 2015; Le Heron et al., 2020). The scores support research and do not diagnose.
Scores
- Overharvestingpoints
- The marginal value theorem's leaving threshold minus the expected next harvest when a tree was left, in apples, averaged over the two orchards. Each tree's estimate is the mean of its last two harvests times the orchard's mean share kept.
- Positive means staying at trees too long; 0 matches the theorem's rule
- Leaving Threshold, Short Walkspoints
- Mean expected next harvest, in apples, when a tree was left in the orchard with 6 second walks.
- Higher means leaving trees earlier
- Leaving Threshold, Long Walkspoints
- Mean expected next harvest, in apples, when a tree was left in the orchard with 9 second walks.
- Higher means leaving trees earlier
- Threshold Changepoints
- Short-walk threshold minus long-walk threshold. The theorem predicts 0.68 apples at the standard settings.
- Positive means leaving earlier when walks are short, as the theorem predicts
- Harvests per Tree, Short Walkscount
- Mean harvests on the trees left in the orchard with short walks.
- The theorem predicts fewer than with long walks
- Harvests per Tree, Long Walkscount
- Mean harvests on the trees left in the orchard with long walks.
- The theorem predicts more than with short walks
- Apples Collectedpoints
- Apples collected in the two orchards.
- Higher means more reward earned
- Missed Turnscount
- Turns with no answer within 2 seconds.
- Lower supports interpretation
- Answer Timems
- Median time to answer, over answers of at least 100 ms.
- Shows how long choices took
Use and evidence
356 papers from 2021 to 2025 name this task in their title or abstract, up 36% compared with 2016 to 2020. That ranks it 72 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 110 |
| 2024 | 63 |
| 2023 | 68 |
| 2022 | 55 |
| 2021 | 60 |
| 2020 | 48 |
| 2019 | 68 |
| 2018 | 67 |
| 2017 | 37 |
| 2016 | 41 |
| 2015 | 45 |
| 2014 | 41 |
| 2013 | 34 |
| 2012 | 30 |
| 2011 | 26 |
| 2010 | 23 |
| 2009 | 30 |
| 2008 | 16 |
| 2007 | 17 |
| 2006 | 15 |
| 2005 | 17 |
| 2004 | 12 |
| 2003 | 8 |
| 2002 | 5 |
| 2001 | 6 |
| 2000 | 11 |
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
- 2
- Papers on reliability or validity
- 32
- Registered clinical studies using it as an outcome
- 7
Note on the counts: includes animal studies.
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 0.14.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Turns
- Harvest the current tree, which takes 3 s, or walk to a new tree. The time taken to answer counts toward the harvest or the walk.
- Orchards
- Two, with walks of 6 s and 9 s, in an order drawn from the session's seed (standard) or fixed; 4 minutes each (standard, 2 to 14).
- Trees
- A new tree's first harvest gives 10 apples on average (SD 1); after each harvest it keeps a share of its yield with mean 0.88 (standard) or 0.94 and SD 0.07.
- Deadline
- Answers within 2 s; a late answer costs a 3 s pause with no apples.
- Practice
- 45 s in a practice orchard with a 7.5 s walk.
- Optimal rule
- Leave once the expected next harvest falls below the long-run apples per second times 3 s: 5.72 apples with 6 s walks and 5.04 with 9 s walks at the standard settings.
- Apples
- Points only (standard), or paid when the study pays for them.
Constantino and Daw (2015) ran four orchards of 14 minutes each, crossing two walk times with two depletion rates, and paid for apples. This version runs the two walk times at one depletion rate, 4 minutes each as standard, and states the answer deadline, which the paper does not. The main score compares each departure with the marginal value theorem's threshold for that orchard. 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: the scored run's test turns. No choice is correct in advance, and harvesting is the common choice by design, so neither accuracy nor a repeated choice is judged. An orchard with no turns, or with no tree left after two or more harvests, invalidates; fewer than 3 trees behind a threshold, or most trees left after fewer than two harvests, is flagged. Missing more than 25% of turns invalidates and more than 5% is flagged. Answers under 100 ms on at least half the turns invalidate.
Why these rules
There is no correct option to compare with chance, so accuracy is not judged against chance. Several harvests come before each departure, so a run of the same choice is the task, not a sign of inattention. The scores rest on the trees left after at least two harvests: with none in an orchard there is no threshold to compare with the theorem's, and Constantino and Daw (2015) excluded participants who nearly always harvested, even down to no apples. A threshold from fewer than 3 trees is imprecise, and leaving most trees after one harvest or none earns almost nothing, so both are flagged. The paper reports 1.7% of turns too slow, so missing more than a quarter means the task was not being done and more than 1 in 20 is unusual. Every turn costs the same time however fast the answer, so only answers under 100 ms, which cannot follow from the screen, count as too fast; at least half that fast invalidates.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Minutes per orchard | 4 | 2 to 14 |
| Depletion per harvest | Steep, 0.88 kept (standard) | Steep, 0.88 kept (standard), Shallow, 0.94 kept |
| Order of orchards | Drawn from the session's seed (standard) | Drawn from the session's seed (standard), Short walks first, Long walks first |
| Payment line | Hypothetical apples (standard) | Hypothetical apples (standard), The study pays for apples |
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 Patch Foraging Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Charnov, 1976; Constantino & Daw, 2015), with the standard settings (minutes per orchard: 4; depletion per harvest: Steep, 0.88 kept; order of orchards: Drawn from the session's seed; payment line: Hypothetical apples). The main score was overharvesting (the marginal value theorem's leaving threshold minus the expected next harvest when a tree was left, in apples, averaged over the two orchards).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Charnov, E. L. (1976). Optimal foraging, the marginal value theorem. Theoretical Population Biology, 9(2), 129-136. https://doi.org/10.1016/0040-5809(76)90040-X
- Constantino, S. M., & Daw, N. D. (2015). Learning the opportunity cost of time in a patch-foraging task. Cognitive, Affective, & Behavioral Neuroscience, 15(4), 837-853. https://doi.org/10.3758/s13415-015-0350-y
- Le Heron, C., Kolling, N., Plant, O., Kienast, A., Janska, R., Ang, Y.-S., Fallon, S., Husain, M., & Apps, M. A. J. (2020). Dopamine modulates dynamic decision-making during foraging. Journal of Neuroscience, 40(27), 5273-5282. https://doi.org/10.1523/JNEUROSCI.2586-19.2020People leave earlier when the environment is richer, and stay longer than the theorem predicts.
Source papers
- Constantino and Daw (2015): patch foraging with walk time and depletion varied across orchards. Le Heron et al. (2020): foraging in rich and poor environments.Paradigm source, Adults
Questions
How long does the Patch Foraging Task take?
About 9 to 11 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?
Overharvesting, Leaving Threshold, Short Walks, Leaving Threshold, Long Walks, Threshold Change, Harvests per Tree, Short Walks, Harvests per Tree, Long Walks, Apples Collected, Missed Turns and Answer Time. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Patch Foraging Task used in research?
356 papers from 2021 to 2025 name it in their title or abstract, up 36% compared with 2016 to 2020. That ranks it 72 of 395 test families in the ConductCognition test census (use grade B). Source: ConductCognition test census 0.14.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Patch Foraging Task have?
In the published research, 2 papers report its test-retest reliability and 7 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 orchard (2 to 14), depletion per harvest (Steep, 0.88 kept (standard), Shallow, 0.94 kept), order of orchards (Drawn from the session's seed (standard), Short walks first, Long walks first) and payment line (Hypothetical apples (standard), The study pays for apples). 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.