Card sorting test
Card Sorting Test (Berg)
Match each new card to one of the four cards. Right or Wrong after each choice is your only clue.
6-10 minUntimedNo account
For researchers
Measure set shifting and perseveration as participants find sorting rules from feedback and adapt when the rule changes.
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
- 1,227
- Evidence grade
- A
- Registered clinical studies
- 410
What the participant sees
One trial, screen by screen, with the timings the test uses.
Until a click on a key card, no time limit
The four key cards and a card to sort
1,000 ms
Right or Wrong
Participants click or tap the key card they choose. There is no time limit.
Example stimuli
Key card 1
1 red triangle
Key card 2
2 green stars
Key card 3
3 yellow crosses
Key card 4
4 blue circles
To sort: 3 red stars
Same color as key card 1, same shape as key card 2, same number as key card 3.
To sort: 1 green circle
Same number as key card 1, same color as key card 2, same shape as key card 4.
Session structure
Color
Category 1
Shape
Category 2
Number
Category 3
Color
Category 4
Shape
Category 5
Number
Category 6
Example data
One session, scored by the code that scores your study's sessions.
Perseverative Errors
16
Errors that follow the rule of the previous category, counted from the first rule change.
Quality check: valid84 scored trials
- Perseverative Errors (%)
- 19%
- Categories Completed
- 6
- Total Errors
- 20
- Failures to Maintain Set
- 0
- Cards to First Category
- 10
- Conceptual Level Responses (%)
- 71.4%
- Color rule
- Shape rule
- Number rule
- Perseverative error, in the old rule's color
- Other error
Trial file
The first 6 of 84 rows, with the column names of your study's export.
| participant_id | test_slug | test_version | scoring_version | trial_index | trial_type | response | rt | correct | time_elapsed | card_color | card_number | card_shape | categories_completed_after | category_completed | category_criterion | category_number | chosen_key | consecutive_correct_after | consecutive_correct_before | deck_size | feedback_ms | finished | matched | paradigm_version | random_seed | rule | run_id | task | task_version | trial_number |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| simulated | card-sorting-berg | 1.0.0 | 1.0.0 | 0 | card-sorting-berg-choice | 2 | 1374 | true | 2374 | yellow | 1 | triangle | 0 | false | 10 | 1 | 3 | 1 | 0 | 128 | 1000 | false | color | 1 | 1406573078 | color | example-run | card_sorting_berg_choice | 1.0.0 | 1 |
| simulated | card-sorting-berg | 1.0.0 | 1.0.0 | 1 | card-sorting-berg-choice | 3 | 1418 | true | 4792 | blue | 4 | cross | 0 | false | 10 | 1 | 4 | 2 | 1 | 128 | 1000 | false | color+number | 1 | 1406573078 | color | example-run | card_sorting_berg_choice | 1.0.0 | 2 |
| simulated | card-sorting-berg | 1.0.0 | 1.0.0 | 2 | card-sorting-berg-choice | 1 | 1717 | true | 7509 | green | 3 | star | 0 | false | 10 | 1 | 2 | 3 | 2 | 128 | 1000 | false | color+shape | 1 | 1406573078 | color | example-run | card_sorting_berg_choice | 1.0.0 | 3 |
| simulated | card-sorting-berg | 1.0.0 | 1.0.0 | 3 | card-sorting-berg-choice | 2 | 1580 | true | 10089 | yellow | 2 | triangle | 0 | false | 10 | 1 | 3 | 4 | 3 | 128 | 1000 | false | color | 1 | 1406573078 | color | example-run | card_sorting_berg_choice | 1.0.0 | 4 |
| simulated | card-sorting-berg | 1.0.0 | 1.0.0 | 4 | card-sorting-berg-choice | 0 | 1282 | true | 12371 | red | 1 | triangle | 0 | false | 10 | 1 | 1 | 5 | 4 | 128 | 1000 | false | color+shape+number | 1 | 1406573078 | color | example-run | card_sorting_berg_choice | 1.0.0 | 5 |
| simulated | card-sorting-berg | 1.0.0 | 1.0.0 | 5 | card-sorting-berg-choice | 0 | 1297 | true | 14668 | red | 2 | circle | 0 | false | 10 | 1 | 1 | 6 | 5 | 128 | 1000 | false | color | 1 | 1406573078 | color | example-run | card_sorting_berg_choice | 1.0.0 | 6 |
Scores and how they are computed
- Perseverative Errorscount
- Errors that follow the rule of the previous category, counted from the first rule change. The scorer replays every rule from the cards and the key cards chosen.
- Lower is better
- Perseverative Errors (%)%
- Perseverative errors per 100 cards sorted.
- Lower is better
- Categories Completedcount
- A category is complete after 10 correct sorts in a row; the rule then changes. Scores run from 0 to 6.
- Higher is better
- Total Errorscount
- Sorts that did not follow the rule in force.
- Lower is better
- Failures to Maintain Setcount
- Errors made after 5 or more correct sorts in a row within a category.
- Lower is better
- Cards to First Categorycount
- Cards sorted up to and including the one that completed the first category.
- Lower is faster learning
- Conceptual Level Responses (%)%
- Correct sorts that are part of runs of three or more correct sorts, per 100 cards.
- Higher is better
What it measures
Measures set shifting and perseveration. Participants sort cards by color, shape or number, learn the rule from Right and Wrong feedback, and have to notice when it changes without warning.
Constructs
- Set shifting
- Perseveration
- Learning from feedback
Research uses
- Neuropsychology and executive-function research
- Psychiatric and neurological group comparisons
- Repeated testing in longitudinal studies
How the paradigm works
The card sorting test measures how people find a rule from feedback and let go of it when it stops working. Four key cards stay at the top of the screen: one red triangle, two green stars, three yellow crosses and four blue circles. Each new card has a color, a shape and a number of symbols, and participants sort it under one of the key cards. They are told only whether each sort was right or wrong.
More about the paradigm
The rule in force is color, then shape, then number, then the same order again. It changes without warning after 10 correct sorts in a row. The test ends after six completed categories or when the deck runs out: 128 cards by default, or 64 in the short form. Each run of 64 cards holds every combination of color, shape and number once, in an order drawn from the session's random seed. There is no time limit, and the cards are buttons that work with a mouse or by touch.
Perseverative errors are errors that follow the rule of the previous category, counted from the first rule change. Scores also include categories completed, total errors, failures to maintain set (an error after 5 or more correct sorts in a row), the cards needed to complete the first category, and the share of correct sorts in runs of three or more. These definitions follow the published task; the scores are not interchangeable with those of the commercial Wisconsin Card Sorting Test manual.
Use and evidence
1,227 papers from 2021 to 2025 name this task in their title or abstract, down 8% compared with 2016 to 2020. That ranks it 20 of 711 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 326 |
| 2024 | 206 |
| 2023 | 238 |
| 2022 | 205 |
| 2021 | 252 |
| 2020 | 237 |
| 2019 | 265 |
| 2018 | 257 |
| 2017 | 265 |
| 2016 | 305 |
| 2015 | 284 |
| 2014 | 277 |
| 2013 | 327 |
| 2012 | 308 |
| 2011 | 286 |
| 2010 | 273 |
| 2009 | 252 |
| 2008 | 279 |
| 2007 | 244 |
| 2006 | 236 |
| 2005 | 179 |
| 2004 | 171 |
| 2003 | 148 |
| 2002 | 171 |
| 2001 | 142 |
| 2000 | 149 |
Use grade A
1,000 or more papers from 2021 to 2025.
Evidence grade A
At least 25 papers report its test-retest reliability and at least 25 registered clinical studies use it as an outcome.
- Papers on test-retest reliability
- 51
- Papers on reliability or validity
- 578
- Registered clinical studies using it as an outcome
- 410
Measured on ConductCognition
Pooled across the studies that feed our platform norms, by device group. A figure needs 50 or more people.
No pooled figure yet.
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 0.58.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Key cards
- One red triangle, two green stars, three yellow crosses and four blue circles.
- Rule
- Color, then shape, then number, then the same order again; it changes without warning after 10 correct sorts in a row.
- End
- After 6 categories or when the deck runs out: 128 cards, or 64 in the short form.
- Deck
- Each run of 64 cards holds every combination of color, shape and number once, in an order drawn from the session's seed.
- Feedback
- Right or Wrong for 1,000 ms after each sort.
- Responses
- Click or tap a key card; no time limit.
Perseverative errors here are errors that follow the previous category's rule. The test follows the published Berg task.
Data you get
Every response is saved as it happens. Download scores for all plans, and every trial on yearly plans or with participant credits.
Trial columns for this test
- trial_number
- Order of the sort in the session.
- card_color
- Color of the card sorted.
- card_shape
- Shape of the card sorted.
- card_number
- Number of symbols on the card sorted.
- chosen_key
- The key card chosen, 1 to 4 from the left.
- matched
- The dimensions the card shares with the key card chosen, such as color+number, or none.
- rule
- The rule in force: color, shape or number.
- category_number
- The category the sort belongs to, 1 to 6.
- correct
- Whether the sort followed the rule in force.
- category_completed
- Whether this sort completed a category.
- consecutive_correct_after
- Correct sorts in a row after this one.
- rt
- Time to choose, in ms.
- random_seed
- Seed of the deck order, to rebuild it.
Columns in every export
- participant_id
- Short participant code, the same in every file of the study.
- external_id
- Your own participant ID, when you add one.
- test_version
- Task version the session ran.
- scoring_version
- Scoring version the scores came from.
- customization_status
- standard, or scored_modified when the study changed a setting. A setting that only changes how a test looks, such as a game's Look setting (the game_look trial column), keeps it standard.
- configuration_json
- The task settings the session ran with.
- look
- plain, or game when the study used the game look. A game-look session's percentile is compared with plain-look norms, and the two looks have not yet been shown to agree.
- look_version
- The version of the game layer the session ran. Empty for a plain session.
- is_valid
- The session's quality status (scores file).
- quality_primary_reason
- The main reason when a session needs review or is invalid (scores file).
- started_at
- When the session started (scores file).
- completed_at
- When the session ended (scores file).
Quality checks
Each session is marked valid, needs review or invalid, with the reason.
Scored rows: the scored run's sorts. Accuracy is not judged against chance. Sorting 90% of cards under one key-card position invalidates when accuracy is not above chance (25%), because the correct position changes from card to card. Answers under 200 ms on at least half the sorts invalidate. Completing no category in 64 or more sorts is flagged.
Why these rules
Perseverating on a rule that is no longer rewarded produces systematic errors, so accuracy at or below chance can be the result the test measures; it is flagged for review and never invalidates. Sorting nearly every card under one key card ignores the cards, so the sorts did not follow the task. 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 |
|---|---|---|
| Deck size | 128 cards (standard) | 128 cards (standard), 64 cards (short form) |
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.
Included in these batteries
Each battery runs this test with others in a fixed order, with standard settings.
- Creativity and Flexible Thinking Battery4 tests, 30 minutes per participant
Cite and describe
Participants completed the Card Sorting Test (Berg) in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Berg, 1948; Grant & Berg, 1948), with the standard settings (deck size: 128 cards). Perseverative errors were the errors that followed the previous category's rule, counted from the first rule change.
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Berg, E. A. (1948). A simple objective technique for measuring flexibility in thinking. Journal of General Psychology, 39(1), 15-22. https://doi.org/10.1080/00221309.1948.9918159
- Grant, D. A., & Berg, E. A. (1948). A behavioral analysis of degree of reinforcement and ease of shifting to new responses in a Weigl-type card-sorting problem. Journal of Experimental Psychology, 38(4), 404-411. https://doi.org/10.1037/h0059831
Source papers
- Berg EA (1948). A simple objective technique for measuring flexibility in thinking. Journal of General Psychology, 39(1):15-22.Paradigm source, Adults
- Grant DA, Berg EA (1948). A behavioral analysis of degree of reinforcement and ease of shifting to new responses in a Weigl-type card-sorting problem. Journal of Experimental Psychology, 38(4):404-411.Paradigm source, Adults
Questions
How long does the Card Sorting Test (Berg) take?
About 6 to 10 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?
Perseverative Errors, Perseverative Errors (%), Categories Completed, Total Errors, Failures to Maintain Set, Cards to First Category and Conceptual Level Responses (%). Each session is also marked valid, needs review or invalid, with the reason.
How are perseverative errors counted?
Errors that follow the rule of the previous category, counted from the first rule change. The scorer replays every rule from the cards and the key cards chosen.
What counts as a completed category?
A category is complete after 10 correct sorts in a row; the rule then changes. Scores run from 0 to 6.
How widely is the Card Sorting Test (Berg) used in research?
1,227 papers from 2021 to 2025 name it in their title or abstract, down 8% compared with 2016 to 2020. That ranks it 20 of 711 test families in the ConductCognition test census (use grade A). Source: ConductCognition test census 0.58.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Card Sorting Test (Berg) have?
In the published research, 51 papers report its test-retest reliability and 410 registered clinical studies use it as an outcome (evidence grade A in the ConductCognition test census).
Which settings can a researcher change?
Deck size (128 cards (standard), 64 cards (short form)). 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.