Episodic Memory
Word Pair Learning
Also called: verbal paired associates task, paired-associate learning, word-pair learning task, cued recall of word pairs
Measure how well people learn new links between words: related and unrelated word pairs, recalled from the first word over three rounds and after a break.
10-12 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
- 101
- Evidence grade
- B
- Registered clinical studies
- 48
What it measures
Measures how well people learn pairs of words. Participants study 12 English word pairs, half related and half unrelated, and type the second word of each pair when shown the first, over three rounds and once more after a short break.
Constructs
- Associative memory
- Verbal episodic memory
- Learning across trials
- Delayed recall
Research uses
- Memory and aging research
- Memory research in dementia
- Within-study group comparison
How the paradigm works
Word Pair Learning measures how well people learn new links between words, a form of associative memory. Participants study 12 pairs of English nouns, one pair at a time for 4 seconds each with a 0.5 second gap. Six pairs are related, such as landlord and tenant, and six are unrelated, such as cricket and coral. After the list, each first word appears alone and participants type the word that went with it, with no time limit and no feedback. The list and the test come three times, each time in a new order, and the test takes the list a third at a time so that every pair waits about as long. A 2 minute break of true-or-false arithmetic follows, then a delayed round asks for every pair once more with no study. Two practice pairs with feedback come first. The method is the paired-associate learning of Calkins (1896) in the form of the associate learning test of Wechsler (1945): related and unrelated pairs over three rounds. Researchers can set 2 to 8 seconds per pair and a break of 60 to 600 seconds.
More about the paradigm
The pairs were chosen for this task by stated rules: common English nouns, related pairs whose words lie close in a model of word meaning (GloVe), and unrelated pairs matched to them on word length and frequency, with no word another test shows. Typed answers ignore case, accents, punctuation and a leading article, and one wrong, missing, extra or swapped letter is accepted in words of 4 or more letters, unless the answer is another English word.
The primary score is the number of pairs recalled over the three learning rounds, from 0 to 36. Recall in each round and in the delayed round, the gain from the first round to the third, the delayed round as a percent of the third, recall of related and of unrelated pairs, and answers that name a word of another pair are also reported. Unrelated pairs are the harder kind, more so in older adults (Naveh-Benjamin, 2000; Badham and colleagues, 2012) and in Alzheimer's disease (Duchek and colleagues, 1991). The words are English in every session language. Scores support research and do not diagnose.
Scores
- Total Pairs Recalledcount
- Pairs recalled summed over the three learning rounds, 0 to 36 at the standard settings.
- Higher is better
- Recalled, Round 1count
- Pairs recalled in the first round.
- Higher is better
- Recalled, Round 2count
- Pairs recalled in the second round.
- Higher is better
- Recalled, Round 3count
- Pairs recalled in the third round.
- Higher is better
- Recalled, Delayed Roundcount
- Pairs recalled after the break, with no study.
- Higher is better
- Learning Gaincount
- Pairs recalled in the third round minus the first.
- Positive values show learning across rounds
- Percent Retained%
- Pairs recalled in the delayed round as a percent of the third round.
- Higher is better; can exceed 100
- Related Pairs Recalledcount
- Related pairs recalled over the three learning rounds, 0 to 18.
- Higher is better
- Unrelated Pairs Recalledcount
- Unrelated pairs recalled over the three learning rounds, 0 to 18.
- Higher is better
- Related Pairs, Delayed Roundcount
- Related pairs recalled in the delayed round.
- Higher is better
- Unrelated Pairs, Delayed Roundcount
- Unrelated pairs recalled in the delayed round.
- Higher is better
- Wrong-Pair Answerscount
- Answers that name a word of another pair, or the cue itself, over every round.
- Context for the scores
- Typo Matchescount
- Answers accepted with one wrong, missing, extra or swapped letter.
- Context for recall scores
- Blank Answerscount
- Answers left blank, over every round.
- Context for recall scores
- Break Accuracy%
- Share of true-or-false arithmetic answered correctly during the break.
- Shows the break task was done
- Answers Recordedcount
- Recall answers recorded in the run.
- Fewer than planned invalidates the session
- Answers Plannedcount
- Recall answers the session planned: every pair in four rounds.
- Context for the scores
Use and evidence
101 papers from 2021 to 2025 name this task in their title or abstract, up 3% compared with 2016 to 2020. That ranks it 187 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 20 |
| 2024 | 21 |
| 2023 | 25 |
| 2022 | 15 |
| 2021 | 20 |
| 2020 | 22 |
| 2019 | 22 |
| 2018 | 19 |
| 2017 | 21 |
| 2016 | 14 |
| 2015 | 16 |
| 2014 | 17 |
| 2013 | 14 |
| 2012 | 19 |
| 2011 | 25 |
| 2010 | 20 |
| 2009 | 9 |
| 2008 | 11 |
| 2007 | 6 |
| 2006 | 13 |
| 2005 | 11 |
| 2004 | 13 |
| 2003 | 11 |
| 2002 | 4 |
| 2001 | 4 |
| 2000 | 5 |
Use grade C
100 to 299 papers from 2021 to 2025.
Evidence grade B
At least 8 test-retest papers, and either 5 registered clinical studies or 150 reliability or validity papers.
- Papers on test-retest reliability
- 12
- Papers on reliability or validity
- 78
- Registered clinical studies using it as an outcome
- 48
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.
- Pairs
- 12 pairs of common English nouns: 6 related, such as landlord and tenant, and 6 unrelated, each matched to a related pair on the length and frequency of its words. Related words lie close in a model of word meaning (GloVe cosine at least 0.5) and unrelated words far apart (at most 0.2).
- Study
- Every pair once a round, 4 seconds each (standard; 2 to 8) with a 0.5 second gap, in a new order each round with no more than 3 related or unrelated pairs in a row.
- Recall
- After each study list, each first word alone; the participant types the second, with no time limit and no feedback. The test takes the list a third at a time, each third in a new order, so every pair waits about as long.
- Rounds
- 3 rounds of study and recall, then 120 seconds (standard; 60 to 600) of true-or-false arithmetic, then a delayed round of recall with no study.
- Practice
- 2 pairs, one related and one unrelated, studied and recalled, each answer followed by the correct pair for 3,000 ms.
- Answers
- Case, accents, punctuation and a leading article are ignored, and one wrong, missing, extra or swapped letter is accepted in words of 4 or more letters, unless the answer is another English word.
- Selection rule
- Candidate pairs written for this task, chosen by stated rules from word frequencies, word ratings and a model of word meaning, with no word another test shows. No published word-pair list is used.
The pairs were written for this task, so scores are not comparable with norms from published paired-associate tests. The words are English.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: word_pair_learning_recall from the scored run, one typed answer per pair in each of the three learning rounds and the delayed round. Typed answers have no chance level, so accuracy is not judged against chance, and answers are untimed, so none can be missed. Typed answers under 200 ms on at least half of 4 or more answered pairs invalidate. Fewer answers recorded than planned invalidate. Recalling no pairs in the learning rounds, the same answer for 90% of at least 12 answers, no arithmetic answered during the break, or break arithmetic at or below chance over at least 10 answers is flagged for review. A session run in another language is flagged. A hidden tab withholds the percentile.
Why these rules
There is no correct option to compare with chance, so accuracy is not judged against chance. The scores are counts over every round, so a session with missing answers cannot be compared with complete ones. Recalling nothing can be a real result, so it is flagged rather than invalidated, as for verbal learning. One word typed for nearly every cue cannot recall 12 different pairs, but it can be a placeholder typed for forgotten pairs, so it is flagged. The break keeps the pairs out of rehearsal before the delayed round; no arithmetic answered, or true-or-false answers at or below chance, means the break may not have been used for the arithmetic. 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 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. A hidden tab lets the study, delay or response timers run without the participant, so a percentile is withheld after one.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Study time per pair (seconds) | 4 | 2 to 8 |
| Break before the delayed round (seconds) | 120 | 60 to 600 |
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
- 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.
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the Word Pair Learning task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Calkins, 1896; Wechsler, 1945), with the standard settings (study time per pair (seconds): 4; break before the delayed round (seconds): 120). The main score was total pairs recalled (pairs recalled summed over the three learning rounds, 0 to 36 at the standard settings).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Calkins, M. W. (1896). Association: An essay analytic and experimental. The Psychological Review: Monograph Supplements, 1(2), i-56. https://doi.org/10.1037/h0092984The paired-associate method.
- Wechsler, D. (1945). A standardized memory scale for clinical use. The Journal of Psychology, 19(1), 87-95. https://doi.org/10.1080/00223980.1945.9917223Related and unrelated pairs over three rounds of cued recall.
- Duchek, J. M., Cheney, M., Ferraro, F. R., & Storandt, M. (1991). Paired associate learning in senile dementia of the Alzheimer type. Archives of Neurology, 48(10), 1038-1040. https://doi.org/10.1001/archneur.1991.00530220054019Learning of related and unrelated pairs in Alzheimer's disease.
- Uttl, B., Graf, P., & Richter, L. K. (2002). Verbal Paired Associates tests limits on validity and reliability. Archives of Clinical Neuropsychology, 17(6), 567-581. https://doi.org/10.1093/arclin/17.6.567Ceiling effects, and the first round apart from learning across rounds.
- Naveh-Benjamin, M. (2000). Adult age differences in memory performance: Tests of an associative deficit hypothesis. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26(5), 1170-1187. https://doi.org/10.1037/0278-7393.26.5.1170Age and memory for links between items.
- Badham, S. P., Estes, Z., & Maylor, E. A. (2012). Integrative and semantic relations equally alleviate age-related associative memory deficits. Psychology and Aging, 27(1), 141-152. https://doi.org/10.1037/a0023924Related pairs and the age difference.
Source papers
- Calkins MW (1896). Association: an essay analytic and experimental. The Psychological Review: Monograph Supplements, 1(2):i-56.Paradigm source, Adults
- Wechsler D (1945). A standardized memory scale for clinical use. The Journal of Psychology, 19(1):87-95.Paradigm source, Adults
- Uttl B, Graf P, Richter LK (2002). Verbal Paired Associates tests limits on validity and reliability. Archives of Clinical Neuropsychology, 17(6):567-581.Paradigm source, 18 to 91 years
- Naveh-Benjamin M (2000). Adult age differences in memory performance: tests of an associative deficit hypothesis. Journal of Experimental Psychology: Learning, Memory, and Cognition, 26(5):1170-1187.Paradigm source, Young and older adults
- Badham SP, Estes Z, Maylor EA (2012). Integrative and semantic relations equally alleviate age-related associative memory deficits. Psychology and Aging, 27(1):141-152.Paradigm source, Young and older adults
- Duchek JM, Cheney M, Ferraro FR, Storandt M (1991). Paired associate learning in senile dementia of the Alzheimer type. Archives of Neurology, 48(10):1038-1040.Paradigm source, Older adults
Questions
How long does the Word Pair Learning take?
About 10 to 12 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?
Total Pairs Recalled, Recalled, Round 1, Recalled, Round 2, Recalled, Round 3, Recalled, Delayed Round, Learning Gain, Percent Retained, Related Pairs Recalled, Unrelated Pairs Recalled, Related Pairs, Delayed Round, Unrelated Pairs, Delayed Round, Wrong-Pair Answers, Typo Matches, Blank Answers, Break Accuracy, Answers Recorded and Answers Planned. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Word Pair Learning used in research?
101 papers from 2021 to 2025 name it in their title or abstract, up 3% compared with 2016 to 2020. That ranks it 187 of 395 test families in the ConductCognition test census (use grade C). Source: ConductCognition test census 0.14.0, counted 31 October 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Word Pair Learning have?
In the published research, 12 papers report its test-retest reliability and 48 registered clinical studies use it as an outcome (evidence grade B in the ConductCognition test census).
Which settings can a researcher change?
Study time per pair (seconds) (2 to 8) and break before the delayed round (seconds) (60 to 600). Settings can be changed on yearly plans; free and pay-as-you-go studies use the standard settings.
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.