Attention
Part-Whole Face Task
Also called: part-whole task, part/whole task, parts and wholes task, part-whole effect, whole-part effect, part-whole recognition task
Measure holistic face recognition: participants see a face for a moment, then pick it from two faces that differ in one feature, or pick that feature alone, scored by how much better features are recognized in the whole face.
7-9 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
- 22
- Evidence grade
- D
- Registered clinical studies
- 1
What it measures
Measures holistic face recognition. Participants see a face for a moment, then pick it from two faces that differ in one feature, or pick that feature shown alone. The main score is the whole-face advantage: how much better a feature is recognized in the whole face than alone.
Constructs
- Holistic face recognition
- Memory for face features
- Face inversion effects
Research uses
- Holistic face processing in research on development, aging and recognition of own-group faces
- Face recognition research in prosopagnosia and autism, at the group level
- Batteries of face tasks with the composite face and face inversion effects
How the paradigm works
The part-whole face task measures holistic face recognition: how much a face's features are remembered as parts of the whole face. Each round shows a face for 1 second, then a scrambled face, then a choice between two pictures with no time limit, by click, tap or the F and J keys. On whole trials the two choices are the face just seen and the same face with other eyes, another nose or another mouth; on part trials they are that feature's two versions alone. Tanaka and Farah (1993) found that people recognize a studied face's features better in the whole face than alone, and not when the faces are scrambled or upside down, or for houses; the eyes and the mouth show the advantage most and the nose least (Tanaka and Simonyi, 2016).
More about the paradigm
The faces are made from photographs of the Face Research Lab London Set (DeBruine and Jones, 2017). Each of 8 sets, one for each self-reported gender and group in the set, joins one person's face outline and skin with the eyes, nose and mouth of two other people of the same gender and group, giving 8 composite faces that differ only in their features. The pictures are in grayscale, in an oval that hides hair, ears and clothing. A session has 96 trials: each set, feature and kind of trial twice. Upside-down trials are a setting that adds the same 96 trials upside down; McKone and colleagues (2013) advise this control before the advantage is read as holistic processing of upright faces. Six practice trials with feedback come first.
The main score is the whole-face advantage: percent right on upright whole trials minus upright part trials, in percentage points. Rezlescu and colleagues (2017) found 7 to 9 points online and in the lab, and Sunday and colleagues (2017) 77.5% right on whole trials against 68.8% on part trials. The task also reports the advantage for each feature, the same scores on upside-down trials when they are run, and the upright advantage minus the upside-down one. Presses within 100 ms of the choice are counted as anticipations and not taken as answers. Scores support research and do not diagnose.
Scores
- Whole-Face Advantage%
- Percent right on upright whole trials minus upright part trials, in percentage points. Positive means features were recognized better in the whole face than alone.
- Informational; clearer over a group than in one person
- Whole Trials Correct%
- Percent right on upright whole trials, where the choice is the face just seen or the same face with one feature changed.
- Informational
- Part Trials Correct%
- Percent right on upright part trials, where the choice is between two versions of one feature alone.
- Informational
- Eyes Advantage%
- The whole-face advantage for the eyes and brows, upright, in percentage points.
- Informational
- Nose Advantage%
- The whole-face advantage for the nose, upright, in percentage points.
- Informational
- Mouth Advantage%
- The whole-face advantage for the mouth, upright, in percentage points.
- Informational
- Whole-Face Advantage Upside Down%
- The whole-face advantage on upside-down trials, in percentage points.
- Informational; blank without upside-down trials
- Whole Trials Correct Upside Down%
- Percent right on upside-down whole trials.
- Informational; blank without upside-down trials
- Part Trials Correct Upside Down%
- Percent right on upside-down part trials.
- Informational; blank without upside-down trials
- Upright-Specific Advantage%
- The upright whole-face advantage minus the upside-down one, in percentage points. Positive means the advantage belongs to upright faces.
- Informational; blank without upside-down trials
- Accuracy%
- Percent right over every answered trial, both kinds and both orientations.
- Quality check
- Median Choice Timems
- Median time from the choice appearing to the answer.
- Informational; the choice has no time limit
- Trials Answeredtrials
- Trials answered in the scored run.
- Fewer than planned is flagged for review
- Planned Trialstrials
- Trials the session planned: 96, or 192 with upside-down trials.
- Informational
- Anticipationscount
- Presses within 100 ms of the choice appearing, not taken as answers.
- Quality check
Use and evidence
22 papers from 2021 to 2025 name this task in their title or abstract, up 38% compared with 2016 to 2020. That ranks it 395 of 495 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 2 |
| 2024 | 4 |
| 2023 | 2 |
| 2022 | 6 |
| 2021 | 8 |
| 2020 | 0 |
| 2019 | 2 |
| 2018 | 6 |
| 2017 | 5 |
| 2016 | 3 |
| 2015 | 2 |
| 2014 | 3 |
| 2013 | 8 |
| 2012 | 5 |
| 2011 | 7 |
| 2010 | 4 |
| 2009 | 2 |
| 2008 | 1 |
| 2007 | 2 |
| 2006 | 0 |
| 2005 | 0 |
| 2004 | 0 |
| 2003 | 1 |
| 2002 | 0 |
| 2001 | 0 |
| 2000 | 1 |
Use grade E
Fewer than 30 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 1
- Papers on reliability or validity
- 7
- Registered clinical studies using it as an outcome
- 1
Note on the counts: a few counted papers use part-whole for other topics.
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 2 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Faces
- Grayscale composite faces made from Face Research Lab London Set photographs, aligned on the eyes and cropped to an oval that hides hair, ears and clothing, up to 280 by 320 px. Each of 8 sets joins one person's face outline and skin with the eyes and brows, the nose and the mouth of two other people of the same self-reported gender and group, 4 sets of women and 4 of men, giving 8 faces that differ only in their features.
- Each trial
- A 500 ms cross, a face for 1,000 ms and a scrambled face for 500 ms. Then two pictures side by side, picked by click, tap, or F and J, with no time limit. Presses within 100 ms of the choice are counted and not taken as answers. A 500 ms blank follows each choice.
- Whole and part trials
- On whole trials the choice is the face just seen and the same face with the other version of its eyes, nose or mouth. On part trials it is that feature's two versions alone, in their places on a black ground.
- Trials
- 96: each set, feature and kind of trial twice, once with each version as the answer, in an order drawn for each session with no set twice in a row and no more than 4 answers in a row on one side. A rest after every 48.
- Upside-down trials
- A setting that adds the same 96 trials upside down, mixed in, for 192 in all.
- Practice
- 6 trials of 2 sets the test does not use, 8 with upside-down trials, with feedback for 1,500 ms.
The procedure follows the sequential part-whole task of Tanaka, Kiefer and Bukach (2004) as DeGutis and colleagues (2013) ran it, with sets of our own made from photographs: one person's face with the features of two others. The main score is the whole-face advantage by subtraction; scores by regression need a sample and can be computed from the stored trials. No screen names or labels a face. Its faces are made from 30 people, who include 25 of the Face Matching Task's 40, 14 of the Face Memory Test's 24 distractor faces, 13 of the Face-name Memory Task's 24 foils and 13 of the Chimeric Faces Task's 18, so in a study with the Face Matching, Face Memory or Face-name Memory test, run this task after them or in a separate session. Four of its people are shown whole by the Thatcher Illusion Task; here they give the composites only their outline, the same in both choices, so the two tasks can run in either order.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: part_whole_faces_test from the scored run, one choice of the face or the part just seen per trial with no time limit (chance 1 in 2). Accuracy significantly below chance invalidates and not above chance is flagged. One side on 90% of choices invalidates when accuracy is not above chance; the answer is on each side in half the trials. Choices under 200 ms on at least half the trials invalidate. A run short of its planned trials is flagged for review.
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. One option on at least 90% of answered trials, when the correct options are balanced and accuracy is not above chance, means the answers did not depend on the items (the task counterpart of long-string responding in surveys; Curran 2016). 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. The face to remember shows for 1 second, so a hidden tab withholds the percentile.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Add upside-down trials | Off | 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.
- 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 Part-Whole Face Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Tanaka & Farah, 1993), with the standard settings (add upside-down trials: Off). The main score was whole-face advantage (percent right on upright whole trials minus upright part trials, in percentage points).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Tanaka, J. W., & Farah, M. J. (1993). Parts and wholes in face recognition. The Quarterly Journal of Experimental Psychology Section A, 46(2), 225-245. https://doi.org/10.1080/14640749308401045
- Tanaka, J. W., Kiefer, M., & Bukach, C. M. (2004). A holistic account of the own-race effect in face recognition: Evidence from a cross-cultural study. Cognition, 93(1), B1-B9. https://doi.org/10.1016/j.cognition.2003.09.011The sequential part-whole task: one face, then two whole faces or two parts.
- DeGutis, J., Mercado, R. J., Wilmer, J., & Rosenblatt, A. (2013). Individual differences in holistic processing predict the own-race advantage in recognition memory. PLoS ONE, 8(4), e58253. https://doi.org/10.1371/journal.pone.0058253The trial: a face for 1 second, a scrambled-face mask for 500 ms, then two faces or two parts side by side until a response.
- Tanaka, J. W., & Simonyi, D. (2016). The "parts and wholes" of face recognition: A review of the literature. Quarterly Journal of Experimental Psychology, 69(10), 1876-1889. https://doi.org/10.1080/17470218.2016.1146780The eyes and the mouth show the whole-face advantage most, the nose least.
- McKone, E., Davies, A. A., Darke, H., Crookes, K., Wickramariyaratne, T., Zappia, S., Fiorentini, C., Favelle, S., Broughton, M., & Fernando, D. (2013). Importance of the inverted control in measuring holistic face processing with the composite effect and part-whole effect. Frontiers in Psychology, 4, 33. https://doi.org/10.3389/fpsyg.2013.00033Why upside-down trials should be run before reading the advantage as holistic processing of upright faces.
- Rezlescu, C., Susilo, T., Wilmer, J. B., & Caramazza, A. (2017). The inversion, part-whole, and composite effects reflect distinct perceptual mechanisms with varied relationships to face recognition. Journal of Experimental Psychology: Human Perception and Performance, 43(12), 1961-1973. https://doi.org/10.1037/xhp0000400Whole-face advantages of 9 points online and 7 points in the lab, and their weak relation to the composite effect.
- Sunday, M. A., Richler, J. J., & Gauthier, I. (2017). Limited evidence of individual differences in holistic processing in different versions of the part-whole paradigm. Attention, Perception, & Psychophysics, 79(5), 1453-1465. https://doi.org/10.3758/s13414-017-1311-zThe advantage is clear in a group but its reliability in one person is low.
- DeBruine, L., & Jones, B. (2017). Face Research Lab London Set [Data set]. figshare. https://doi.org/10.6084/m9.figshare.5047666.v5
Source papers
- Tanaka JW, Farah MJ (1993). Parts and wholes in face recognition. Quarterly Journal of Experimental Psychology Section A, 46(2):225-245.Paradigm source, Adults
- DeBruine L, Jones B (2017). Face Research Lab London Set. figshare. Dataset. doi:10.6084/m9.figshare.5047666.v5Paradigm source, Adults
Questions
How long does the Part-Whole Face Task take?
About 7 to 9 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?
Whole-Face Advantage, Whole Trials Correct, Part Trials Correct, Eyes Advantage, Nose Advantage, Mouth Advantage, Whole-Face Advantage Upside Down, Whole Trials Correct Upside Down, Part Trials Correct Upside Down, Upright-Specific Advantage, Accuracy, Median Choice Time, Trials Answered, Planned Trials and Anticipations. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Part-Whole Face Task used in research?
22 papers from 2021 to 2025 name it in their title or abstract, up 38% compared with 2016 to 2020. That ranks it 395 of 495 test families in the ConductCognition test census (use grade E). Source: ConductCognition test census 0.22.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Part-Whole Face Task have?
In the published research, 1 paper reports its test-retest reliability and 1 registered clinical study uses 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?
Add upside-down trials (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. 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.