Reasoning
Letter Pair Reasoning Task
Also called: grammatical reasoning test, grammatical transformation test, Baddeley three-minute reasoning test, AB reasoning test
Measure verbal reasoning against the clock: participants judge short active, passive and negative sentences about the letters A and B as true or false of a letter pair, scored by right answers in the time.
5-6 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
- 8
- Evidence grade
- C
- Registered clinical studies
- 3
What it measures
Measures verbal reasoning against the clock. Participants read short sentences about the order of two letters, such as "A is not followed by B", and press F or J to say whether each is true of the pair below it, AB or BA. The sentences vary in voice and negation and are shown in the participant's language. The score is the right answers in a time limit set before the session.
Constructs
- Verbal reasoning
- Sentence comprehension
- Processing of negation and the passive voice
- Speed of reasoning
Research uses
- Research on how stress, fatigue and drugs affect performance
- Individual-differences studies of reasoning and fluid intelligence
- Within-study group comparison
How the paradigm works
The Letter Pair Reasoning Task measures how quickly and accurately people work out what short sentences say. Each item is a sentence about the order of the letters A and B, such as "A is not followed by B", with the pair AB or BA below it. The participant presses one key if the sentence is true of the pair and the other if it is false, and the next sentence follows at once. The score is the number of right answers before a time limit set in advance.
More about the paradigm
The method is the three-minute reasoning test of Baddeley (1968), built for studies of stress on performance, where a test of higher mental processes had to be short and easy to give. It rests on the finding that positive sentences are understood faster than negative ones, active faster than passive and true faster than false. Baddeley crossed positive or negative, active or passive, precedes or follows, A or B named first and the pair AB or BA. He reported a correlation of .59 with the British Army verbal intelligence test in 29 enlisted men, a mean correlation of .80 between two days in 18 people, and practice that raised enlisted men from 32.9 to 41.9 right over five runs in an hour. The literature calls it the grammatical reasoning test or the grammatical transformation test.
Here the 32 items of that crossing appear one at a time in shuffled rounds, so that wherever the time runs out the items seen are nearly balanced. Eight practice items with feedback come first. The sentences are written in each of the eight participant languages from the same eight forms, with real passives in each. German uses anführen (lead) and verfolgen (pursue), because its verbs of order have no passive, and Japanese uses 追う (chase) for follows, because the passive of its usual verb is not in everyday use. Sentences of the same form do not read alike across languages, so compare scores within a language. The time limit is set before the session and never changes. No published norms are applied to this task. The scores support research and do not diagnose.
Scores
- Right Answerscount
- Sentences judged correctly in the time. The item on screen when the time ran out does not count.
- Higher is better; compare only sessions with the same time limit
- Wrong Answerscount
- Sentences judged wrongly in the time.
- Lower is better
- Right Minus Wrongscore
- Right answers minus wrong answers, so that guessing gains nothing.
- Higher is better
- Answered Itemsitems
- Items answered with a key, anticipations included.
- Under 2 a minute flags the session
- Accuracy%
- Right answers as a share of right and wrong answers.
- Under 0.70 over at least 12 answers invalidates the session
- Median Time per Right Answerms
- Median time from an item to the key, over right answers.
- Lower means faster
- Negation Effectms
- Median time over right answers to negative sentences minus positive ones.
- Positive means negative sentences took longer, the usual result
- Passive Effectms
- Median time over right answers to passive sentences minus active ones.
- Positive means passive sentences took longer, the usual result
- False Effectms
- Median time over right answers to false sentences minus true ones.
- Positive means false sentences took longer, the usual result
- Anticipationscount
- Answers under 100 ms, which count as neither right nor wrong.
- Lower is better
- Time Limitseconds
- The length of the timed block.
- Set by the settings
- Seconds Usedseconds
- How much of the time limit had passed when the last item ended.
- Under 90% of the limit invalidates the session
Use and evidence
8 papers from 2021 to 2025 name this task in their title or abstract. That ranks it 459 of 495 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 2 |
| 2024 | 1 |
| 2023 | 2 |
| 2022 | 1 |
| 2021 | 2 |
| 2020 | 1 |
| 2019 | 2 |
| 2018 | 0 |
| 2017 | 2 |
| 2016 | 0 |
| 2015 | 1 |
| 2014 | 2 |
| 2013 | 0 |
| 2012 | 2 |
| 2011 | 0 |
| 2010 | 0 |
| 2009 | 2 |
| 2008 | 4 |
| 2007 | 0 |
| 2006 | 1 |
| 2005 | 1 |
| 2004 | 1 |
| 2003 | 3 |
| 2002 | 1 |
| 2001 | 2 |
| 2000 | 2 |
Use grade E
Fewer than 30 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
- 1
- Papers on reliability or validity
- 7
- Registered clinical studies using it as an outcome
- 3
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.
- Items
- 8 sentence forms from every combination of verb (precedes or follows), voice and negation, each naming A or B first and paired with AB and BA: 32 items, half true.
- Languages
- The sentences are written in each of the eight participant languages, so the whole test runs in the participant's language. German uses two other verbs and Japanese one, where a verb of order has no everyday passive.
- Trial
- The sentence above the letter pair, until F or J for true or false. The next item follows at once.
- Time limit
- One timed block of 180 seconds (standard; 60 to 300 seconds as a setting, fixed for the session).
- Order
- Shuffled rounds of all 32 items, drawn from the session's seed, with no more than 3 of one answer in a row and no item twice in a row. The seed also draws which key means true.
- Practice
- 8 items, one of each sentence form, four true and four false, untimed, with feedback.
- Scores
- Right answers in the time, then wrong answers, right minus wrong, accuracy, median time, and the differences between negative and positive, passive and active, and false and true sentences.
Baddeley (1968) gave 64 sentences on paper, the same 32 statements twice, with 3 minutes to tick true or false. Here the items appear one at a time on a screen and are answered with two keys, in a new order for each session. Wrong answers do not lower the main score, as on paper, so a session with fewer than 70% right is invalid. No published norms are applied, and 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: letter_pair_reasoning_test from the scored run, F or J for true or false in one timed block. Chance is 1 in 2 and half the items of every round are true. An item cut off when the time runs out is not a missed response. Below-chance accuracy, one key on 90% of answers when accuracy is not above chance, or more than max(2, 20%) answers under 100 ms invalidate. A block that ran for under 90% of its time, or accuracy under 70% over at least 12 answers, invalidates. Fewer than 2 answers a minute is flagged.
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 100 ms cannot be a reaction to the sentence. The score counts right answers and takes nothing off for wrong ones, so a person pressing fast at random would score more than a person reading; that is why accuracy under 70% invalidates here rather than being flagged, the limit Symbol Scan and Pattern and Letter Comparison use. The count is only comparable over the full time. Baghaei and colleagues (2017) report 7 to 41 right in 3 minutes in their sample, so fewer than 2 answers a minute is unusual but can be real. These limits are ConductScience choices. The sentences are shown in the session's language, so a session in any of the eight languages is scored the same way.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Time limit (seconds) | 180 | 60 to 300 |
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
- Timed test: participants need a desktop or laptop with a mouse or trackpad. 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 Letter Pair Reasoning Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Baddeley, 1968), with the standard settings (time limit (seconds): 180). The main score was right answers (sentences judged correctly in the time).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Baddeley, A. D. (1968). A 3 min reasoning test based on grammatical transformation. Psychonomic Science, 10(10), 341-342. https://doi.org/10.3758/BF03331551The method: sentences about the order of A and B, voice, negation and verb crossed, right answers in 3 minutes.
- Baghaei, P., Khoshdel-Niyat, F., & Tabatabaee-Yazdi, M. (2017). The Persian adaptation of Baddeley's 3-min grammatical reasoning test. Psicologia: Reflexão e Crítica, 30, 16. https://doi.org/10.1186/s41155-017-0070-zThe test's 32 statements, and how far precede and follow carry into other languages.
- Baudson, T. G., & Preckel, F. (2016). mini-q: Intelligenzscreening in drei Minuten. Diagnostica, 62(3), 182-197. https://doi.org/10.1026/0012-1924/a000150A German version that replaced precede and follow, which have no German passive.
Source papers
- Baddeley AD (1968). A 3 min reasoning test based on grammatical transformation. Psychonomic Science, 10(10):341-342.Paradigm source, Adults
Questions
How long does the Letter Pair Reasoning Task take?
About 5 to 6 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?
Right Answers, Wrong Answers, Right Minus Wrong, Answered Items, Accuracy, Median Time per Right Answer, Negation Effect, Passive Effect, False Effect, Anticipations, Time Limit and Seconds Used. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Letter Pair Reasoning Task used in research?
8 papers from 2021 to 2025 name it in their title or abstract. That ranks it 459 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 Letter Pair Reasoning Task have?
In the published research, 1 paper reports its test-retest reliability and 3 registered clinical studies use it as an outcome (evidence grade C in the ConductCognition test census).
Which settings can a researcher change?
Time limit (seconds) (60 to 300). 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.