Social Cognition
Implicit Association Test (IAT)
For research studies. Participants see no personal result. To take an IAT yourself, visit Project Implicit.
Measure association strength with the standard seven-block IAT and the improved D-score algorithm.
5-7 minTimed: desktop or laptop with a mouse or trackpad
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
- 1,497
- Evidence grade
- A
- Registered clinical studies
- 113
What the participant sees
One trial, screen by screen, with the timings the test uses.
250 ms
The category labels over a blank screen
Until the correct key, no time limit
A word
- E
- The groups on the left
- I
- The groups on the right
A wrong key shows a red X until the correct key is pressed; the response time runs to the correct key.
Example stimuli
Flower
Target words are white
Insect
Pleasant word
Attribute words are green
Unpleasant word
Session structure
1
20 · Flowers | Insects
2
20 · Pleasant | Unpleasant
3
20 · Flowers + Pleasant | Insects + Unpleasant
4
40 · Flowers + Pleasant | Insects + Unpleasant
5
40 · Insects | Flowers
6
20 · Insects + Pleasant | Flowers + Unpleasant
7
40 · Insects + Pleasant | Flowers + Unpleasant
Example data
One session, scored by the code that scores your study's sessions.
D Score
0.96
Improved-algorithm D: the pairing difference in latency over its pooled standard deviation, averaged over practice and test parts.
Quality check: valid200 scored trials
- Flowers-Pleasant RT
- 805.8 ms
- Insects-Pleasant RT
- 957 ms
- Error Rate
- 13.3%
- Fast Responses
- 0%
- D, Practice Parts
- 1.00
- D, Test Parts
- 0.91
- Flowers with pleasant
- Insects with pleasant
- Practice blocks
- Mean
- Wrong answer
Trial file
The first 6 of 200 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 | block | block_order | block_role | correct_key | error | first_response | first_rt | inter_trial_ms | left_categories | locale | pairing | paradigm_version | random_seed | right_categories | run_id | stimulus_language | task | trial_in_block | trial_number | word | word_category | word_kind |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| simulated | implicit-association | 1.0.0 | 1.0.0 | 0 | implicit-association-trial | i | 691 | true | 1691 | 1 | compatible-first | target_practice | i | false | i | 691 | 250 | ["flowers"] | en | 1 | 3692577411 | ["insects"] | example-run | en | implicit_association_trial | 1 | 1 | cockroach | insects | target | |
| simulated | implicit-association | 1.0.0 | 1.0.0 | 1 | implicit-association-trial | e | 710 | true | 2401 | 1 | compatible-first | target_practice | e | false | e | 710 | 250 | ["flowers"] | en | 1 | 3692577411 | ["insects"] | example-run | en | implicit_association_trial | 2 | 2 | daffodil | flowers | target | |
| simulated | implicit-association | 1.0.0 | 1.0.0 | 2 | implicit-association-trial | e | 593 | true | 2994 | 1 | compatible-first | target_practice | e | false | e | 593 | 250 | ["flowers"] | en | 1 | 3692577411 | ["insects"] | example-run | en | implicit_association_trial | 3 | 3 | magnolia | flowers | target | |
| simulated | implicit-association | 1.0.0 | 1.0.0 | 3 | implicit-association-trial | e | 599 | true | 3593 | 1 | compatible-first | target_practice | e | false | e | 599 | 250 | ["flowers"] | en | 1 | 3692577411 | ["insects"] | example-run | en | implicit_association_trial | 4 | 4 | lilac | flowers | target | |
| simulated | implicit-association | 1.0.0 | 1.0.0 | 4 | implicit-association-trial | i | 677 | true | 4270 | 1 | compatible-first | target_practice | i | false | i | 677 | 250 | ["flowers"] | en | 1 | 3692577411 | ["insects"] | example-run | en | implicit_association_trial | 5 | 5 | hornet | insects | target | |
| simulated | implicit-association | 1.0.0 | 1.0.0 | 5 | implicit-association-trial | i | 592 | true | 4862 | 1 | compatible-first | target_practice | i | false | i | 592 | 250 | ["flowers"] | en | 1 | 3692577411 | ["insects"] | example-run | en | implicit_association_trial | 6 | 6 | ant | insects | target |
Scores and how they are computed
- D Scorescore
- For the practice pair and the test pair of combined blocks: incompatible minus compatible mean latency over the standard deviation of both blocks' latencies; D is the mean of the two (Greenwald, Nosek and Banaji, 2003). Latencies over 10,000 ms are dropped; more than 10% under 300 ms is flagged.
- Positive values mean faster sorting when flowers share a key with pleasant words
- Flowers-Pleasant RTms
- Mean latency in the combined parts where flowers share a key with pleasant words.
- Lower is faster
- Insects-Pleasant RTms
- Mean latency in the combined parts where insects share a key with pleasant words.
- Lower is faster
- Error Rate%
- The share of words in the combined blocks whose first key was wrong.
- Lower is better
- Fast Responses%
- Share of combined-part responses under 300 ms.
- Above 10% is flagged for review
- D, Practice Partsscore
- D from the practice combined parts (3 and 6).
- Part of the D score
- D, Test Partsscore
- D from the test combined parts (4 and 7).
- Part of the D score
What it measures
Measures how strongly two concepts are linked in memory from how fast people sort words when the concepts share a response key. The default version pairs flowers and insects with pleasant and unpleasant words.
Constructs
- Implicit association strength
- Attitudes
- Speeded categorization
Research uses
- Social cognition and attitude research at the group level
- Teaching demonstrations of implicit measurement
- Method studies of implicit measures
How the paradigm works
The Implicit Association Test measures how strongly two concepts are linked in memory from how fast people sort words when the concepts share a response key. The default version uses flowers and insects with pleasant and unpleasant words, the demonstration pairing of the original paper. Category labels sit at the top left and top right of the screen, and participants press E for the left groups and I for the right groups. A wrong key shows a red X and the participant must press the correct key, so each response time runs to the correct key. A 250 ms blank separates the words.
More about the paradigm
The test has the standard seven parts: flowers and insects alone (20 words), pleasant and unpleasant words alone (20), the first combined pairing (20 practice and 40 test words), flowers and insects with their sides switched (40), and the second combined pairing (20 and 40). Combined parts alternate a flower or insect with a pleasant or unpleasant word. Whether flowers or insects first share a key with pleasant words is drawn from the session's random seed and recorded.
The D score follows the improved algorithm of Greenwald, Nosek and Banaji (2003): latencies over 10,000 ms are dropped, the difference between the two pairings is divided by their pooled standard deviation separately for the practice and test parts, and the two values are averaged. Positive scores mean faster sorting when flowers share a key with pleasant words. Sessions with more than 10% of responses under 300 ms are flagged. A D score is a research measure for group comparisons and says little about one person. The words are English, and sessions run in other languages are flagged for review.
Use and evidence
1,497 papers from 2021 to 2025 name this task in their title or abstract, down 3% compared with 2016 to 2020. That ranks it 16 of 711 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 371 |
| 2024 | 294 |
| 2023 | 265 |
| 2022 | 280 |
| 2021 | 287 |
| 2020 | 330 |
| 2019 | 333 |
| 2018 | 311 |
| 2017 | 272 |
| 2016 | 297 |
| 2015 | 303 |
| 2014 | 265 |
| 2013 | 244 |
| 2012 | 233 |
| 2011 | 212 |
| 2010 | 195 |
| 2009 | 180 |
| 2008 | 127 |
| 2007 | 131 |
| 2006 | 113 |
| 2005 | 61 |
| 2004 | 66 |
| 2003 | 56 |
| 2002 | 31 |
| 2001 | 36 |
| 2000 | 9 |
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
- 68
- Papers on reliability or validity
- 857
- Registered clinical studies using it as an outcome
- 113
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.
- Keys
- E for the groups at the top left and I for the groups at the top right.
- Parts
- Seven parts of 20, 20, 20, 40, 40, 20 and 40 words; the combined pairings are parts 3 and 4 and parts 6 and 7.
- Order
- Which pairing comes first is drawn from the session's seed and saved.
- Errors
- A wrong key shows a red X until the correct key is pressed; response times run to the correct key.
- Interval
- 250 ms between words.
- Scoring
- Improved D algorithm: latencies over 10,000 ms dropped; sessions with more than 10% of responses under 300 ms flagged.
A D score is a research measure for group comparisons and says little about one person. The default content is flowers and insects; the words are English.
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 word in the session.
- block
- The part, 1 to 7.
- block_role
- target_practice, attribute_practice, combined_practice, combined_test or reversed_target_practice.
- pairing
- compatible (flowers with pleasant) or incompatible, for combined parts.
- block_order
- compatible-first or incompatible-first.
- word
- The word shown.
- word_category
- flowers, insects, pleasant or unpleasant.
- correct_key
- e or i.
- first_response
- The first key pressed.
- first_rt
- Time to the first key, in ms.
- error
- Whether the first key was wrong.
- rt
- Time to the correct key, in ms: the latency the D score uses.
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: every implicit_association_trial of the scored run, E or I for the word's category (chance 50%, the two keys are correct equally often in every block), correctness from the first key pressed. Below chance invalidates and not above chance is flagged. One key on 90% of first responses invalidates when accuracy is not above chance. More than 10% of combined-block responses under 300 ms, or a session run in a language other than English, is flagged for review. There is no response window, so nothing is missed.
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). Greenwald, Nosek and Banaji (2003) advise setting aside sessions with more than 10% of latencies under 300 ms, which are too fast to reflect sorting. The word lists are English. A session run in another language shows English words to a participant who may not read English fluently, so it is flagged for review.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
This test runs with fixed settings.
Need a different version of this test?
Devices and languages
- Timed test: participants need a desktop or laptop with a mouse or trackpad. Participant devices
- Instructions and screens: English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese. Words and sentences: English.
- 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 Implicit Association Test (IAT) in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Greenwald, McGhee & Schwartz, 1998; Greenwald, Nosek & Banaji, 2003). D scores were computed with the improved algorithm (Greenwald, Nosek and Banaji, 2003): for the practice and the test pairs of combined blocks, incompatible minus compatible mean latency over the standard deviation of both blocks' latencies, averaged over the two pairs, after dropping latencies over 10,000 ms.
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74(6), 1464-1480. https://doi.org/10.1037/0022-3514.74.6.1464
- Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the Implicit Association Test: I. An improved scoring algorithm. Journal of Personality and Social Psychology, 85(2), 197-216. https://doi.org/10.1037/0022-3514.85.2.197
Source papers
- Greenwald AG, McGhee DE, Schwartz JLK (1998). Measuring individual differences in implicit cognition: the implicit association test. Journal of Personality and Social Psychology, 74(6):1464-1480.Paradigm source, Adults
- Greenwald AG, Nosek BA, Banaji MR (2003). Understanding and using the Implicit Association Test: I. An improved scoring algorithm. Journal of Personality and Social Psychology, 85(2):197-216.Paradigm source, Adults
Questions
How long does the Implicit Association Test (IAT) take?
About 5 to 7 minutes, as listed in the test library.
What device does a participant need?
Timed test: participants need a desktop or laptop with a mouse or trackpad. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
D Score, Flowers-Pleasant RT, Insects-Pleasant RT, Error Rate, Fast Responses, D, Practice Parts and D, Test Parts. Each session is also marked valid, needs review or invalid, with the reason.
How is the D score computed?
For the practice pair and the test pair of combined blocks: incompatible minus compatible mean latency over the standard deviation of both blocks' latencies; D is the mean of the two (Greenwald, Nosek and Banaji, 2003). Latencies over 10,000 ms are dropped; more than 10% under 300 ms is flagged.
How are errors counted?
The share of words in the combined blocks whose first key was wrong.
How widely is the Implicit Association Test (IAT) used in research?
1,497 papers from 2021 to 2025 name it in their title or abstract, down 3% compared with 2016 to 2020. That ranks it 16 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 Implicit Association Test (IAT) have?
In the published research, 68 papers report its test-retest reliability and 113 registered clinical studies use it as an outcome (evidence grade A in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which languages can participants use?
The instructions and screens are in English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese. The words and sentences stay in English, so sessions in other languages are flagged for review. 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.