Social Cognition
Director Task
Also called: director paradigm, referential communication task, perspective-taking shelves task
Measure how well people use another person's view: move objects on shelves for a director who cannot see every slot.
8-12 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
- 54
- Evidence grade
- D
- Registered clinical studies
- 0
What it measures
Measures how well people use another person's view to follow instructions. A drawn director, a computer program, asks for objects to be moved on shelves where some slots are hidden from the director; on critical instructions the best fit from the participant's side is an object the director cannot see. The main score is the share of those instructions answered with the wrong object, with a part that has a rule and no director for comparison.
Constructs
- Using another person's perspective (level-1 perspective taking)
- Theory of mind in use
- Inhibiting one's own view (egocentric bias)
Research uses
- Social cognition research
- Adolescent development and aging studies
- Autism and social communication research
- Within-study group comparison
How the paradigm works
The director task measures how well people use another person's view of a scene when following that person's instructions (Keysar, Barr, Balin and Brauner, 2000). Participants see shelves of 4 by 4 slots holding eight objects. A director stands on the other side, and five slots have a gray back that blocks the director's view. The director asks for objects to be moved: participants click or tap an object and press Up or Down to move it one slot. The director is drawn by our code and is a computer program, and the screens say so.
More about the paradigm
The task follows the computer version of Dumontheil, Apperly and Blakemore (2010). On experimental instructions the words name a relation, such as the small apple or the bottom car, and the object that fits them best from the participant's side is in a gray slot, where the director cannot see it. The right answer is the best fit among the objects the director can see. Control instructions use the same shelves with that hidden object replaced by an unrelated one, and filler instructions name an object with no other of its kind. Each of 16 sets of shelves gets one experimental or control instruction and two fillers, in a seeded order, with 10 seconds for each move as standard. Before the test, participants answer questions about which objects the director can and cannot see, with feedback, and practice three moves. A second part repeats the task with no director and a rule to ignore objects in gray slots, as in the original study; it can be turned off. Objects are pictures of everyday things at three sizes, and the arrangements are dealt by stated rules from the session's seed. The instructions are short sentences written in each of the platform's languages.
The main score is the share of experimental instructions with the director that were answered by moving another object than the one meant, almost always the hidden one. Errors on control instructions, the difference between the two, the same scores in the part with no director, the director's errors minus the rule's, median times to choose the object meant and filler accuracy are also reported. A first choice of the hidden object that was corrected before the move is counted separately. Clicks under 100 ms after an instruction appears are counted as anticipations and not taken as choices. Scores support research and do not diagnose.
Scores
- Errors on Critical Instructions%
- With the director, the share of experimental instructions answered by moving another object than the one meant.
- Lower is better
- Hidden Object Chosen%
- With the director, the share of experimental instructions answered by moving the object the director could not see.
- Lower is better
- Errors on Control Instructions%
- With the director, errors on the same instructions when the hidden object was replaced by an unrelated one.
- Lower is better
- Error Cost of the Hidden Object%
- Experimental minus control errors with the director, in percentage points.
- Lower means the director's view was used more
- Errors on Critical Instructions, No Director%
- In the part with no director, the share of experimental instructions answered with another object than the one meant. Empty when that part is off.
- Lower is better
- Errors on Control Instructions, No Director%
- In the part with no director, errors on control instructions. Empty when that part is off.
- Lower is better
- Director Minus Rule Errors%
- Experimental errors with the director minus those with the rule, in percentage points.
- Informational; the comparison of Dumontheil and colleagues (2010)
- Median Choice Time, Criticalms
- With the director, median time from an experimental instruction to choosing the object meant.
- Informational
- Median Choice Time, Controlms
- With the director, median time from a control instruction to choosing the object meant.
- Informational
- Choice Time Costms
- Experimental minus control median choice time with the director.
- Informational
- Filler Accuracy%
- Filler instructions answered with the right object and direction, in both parts.
- Higher is better; under 75% is flagged
- Corrected First Choicescount
- Experimental instructions with the director where the hidden object was chosen first and the object meant was moved.
- Informational
- Direction Errorscount
- Instructions where the object meant was moved the wrong way.
- Lower is better
- No Move in Timecount
- Instructions with no move before the time limit.
- Lower is better
- Anticipationscount
- Clicks under 100 ms after an instruction appeared. They are not taken as choices.
- Lower is better
- View Questions Not Passedcount
- Questions about which objects the director can see that were not answered right in three tries.
- Above 0 is flagged
- Planned Instructionstrials
- Test instructions the session planned to show.
- Informational
- Total Instructionstrials
- Test instructions shown in the scored run.
- Fewer than planned is flagged for review
Use and evidence
54 papers from 2021 to 2025 name this task in their title or abstract, up 20% compared with 2016 to 2020. That ranks it 248 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 13 |
| 2024 | 4 |
| 2023 | 15 |
| 2022 | 6 |
| 2021 | 16 |
| 2020 | 10 |
| 2019 | 6 |
| 2018 | 13 |
| 2017 | 7 |
| 2016 | 9 |
| 2015 | 7 |
| 2014 | 7 |
| 2013 | 5 |
| 2012 | 2 |
| 2011 | 2 |
| 2010 | 4 |
| 2009 | 2 |
| 2008 | 3 |
| 2007 | 1 |
| 2006 | 2 |
| 2005 | 2 |
| 2004 | 1 |
| 2003 | 2 |
| 2002 | 0 |
| 2001 | 0 |
| 2000 | 3 |
Use grade D
30 to 99 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 1
- Papers on reliability or validity
- 9
- 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 0.14.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Shelves
- 4 by 4 slots holding 8 pictures of everyday objects at three sizes. 5 slots have a gray back that hides them from a director on the other side. The director is a drawn figure and a computer program, and the screens say so.
- Instructions
- Short sentences in the participant's language, such as "Move the small apple up." Each set of shelves gets 3 instructions: one experimental or control instruction and 2 fillers that name an object with no other of its kind, with the first in a seeded place.
- Experimental
- Three objects of one kind, at three sizes for a size word (small or large) or on three shelves for a place word (top or bottom). The best fit from the participant's side is in a gray slot; the right answer is the best fit among the two the director can see.
- Control
- The same shelves with the hidden object replaced by an unrelated one.
- Moves
- Click or tap an object, then press Up or Down to move it one slot. 10 seconds per instruction (standard; 6 to 20). New shelves show alone for 1,500 ms before their first instruction, and the shelves are set back after each move.
- Parts
- With the director: 16 sets of shelves (8 kinds, each with and without the hidden object) and 48 instructions, with no more than 2 sets of one kind of instruction in a row. Then, as standard, the same with no director and a rule to ignore objects in gray slots.
- Practice
- Questions about which objects the director can and cannot see, with feedback and up to 3 tries each, then 3 practice moves with feedback; 1 question and 2 moves before the part with no director.
Dumontheil and colleagues (2010) gave spoken instructions, moves left and right as well as up and down, and 3.6 s to respond after each instruction. Here the instructions are written in each language, moves go up or down only, which look the same from both sides of the shelves, and each instruction has 10 seconds as standard.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: director_task_trial from the scored run, every instruction in both parts. Each is a choice of one of 8 objects and a direction, so chance is 1 in 8 to 1 in 16. Accuracy below chance invalidates and not above chance is flagged. An instruction with no move within its time limit is a missed response: more than max(3, 25%) invalidate and any is flagged. First choices under 200 ms on at least half the instructions invalidate. Following under 75% of the filler instructions, a question about the director's view never answered right, or a short run is flagged and withholds the percentile.
Why these rules
Accuracy significantly below chance (one-sided exact test, p < .05) cannot come from guessing; it means the response mapping was reversed or the task was not followed (Slick, Sherman and Iverson 1999 treat below-chance forced-choice performance as definite response bias). Chance-level accuracy can be a real result in impaired participants, so it is only flagged. A move is due within each instruction's time limit, so an instruction with no move is a missed response under the platform's rule for response windows. 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. Filler instructions name an object with no other of its kind, so missing many of them means the instructions were not followed (Dumontheil and colleagues, 2010, report under 3% filler errors).
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Time per instruction (seconds) | 10 | 6 to 20 |
| Add the part with no director | On | On or off |
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 and tablets. 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 Director Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Keysar et al., 2000; Dumontheil, Apperly & Blakemore, 2010), with the standard settings (time per instruction (seconds): 10; add the part with no director: On). The main score was errors on critical instructions (with the director, the share of experimental instructions answered by moving another object than the one meant).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Keysar, B., Barr, D. J., Balin, J. A., & Brauner, J. S. (2000). Taking perspective in conversation: The role of mutual knowledge in comprehension. Psychological Science, 11(1), 32-38. https://doi.org/10.1111/1467-9280.00211
- Dumontheil, I., Apperly, I. A., & Blakemore, S.-J. (2010). Online usage of theory of mind continues to develop in late adolescence. Developmental Science, 13(2), 331-338. https://doi.org/10.1111/j.1467-7687.2009.00888.xThe computer version: eight objects, five hidden slots, the three kinds of instruction and the part with no director.
Source papers
- Dumontheil I, Apperly IA, Blakemore SJ (2010). Online usage of theory of mind continues to develop in late adolescence. Developmental Science, 13(2), 331-338.Paradigm source, 7 to 27 years in the source
Questions
How long does the Director Task take?
About 8 to 12 minutes, as listed in the test library.
What device does a participant need?
Untimed test: runs on desktops, laptops and tablets. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Errors on Critical Instructions, Hidden Object Chosen, Errors on Control Instructions, Error Cost of the Hidden Object, Errors on Critical Instructions, No Director, Errors on Control Instructions, No Director, Director Minus Rule Errors, Median Choice Time, Critical, Median Choice Time, Control, Choice Time Cost, Filler Accuracy, Corrected First Choices, Direction Errors, No Move in Time, Anticipations, View Questions Not Passed, Planned Instructions and Total Instructions. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Director Task used in research?
54 papers from 2021 to 2025 name it in their title or abstract, up 20% compared with 2016 to 2020. That ranks it 248 of 395 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.14.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Director Task have?
In the published research, 1 paper reports 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?
Time per instruction (seconds) (6 to 20) and add the part with no director (On or off). 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, with its words in the same language. 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.