Creativity
Remote Associates Test
Also called: compound remote associates, compound word problems, RAT
Measure convergent thinking: participants find the one word that forms a compound or phrase with each of three cue words, within a time limit.
8-12 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
- 217
- Evidence grade
- C
- Registered clinical studies
- 9
What it measures
Measures convergent thinking. Participants see three words and type the one word that forms a familiar compound word or phrase with each of them, with 30 seconds per problem by default. The main score is the number of problems solved. The words are in English.
Constructs
- Convergent thinking
- Associative search
- Verbal problem solving
Research uses
- Creativity and insight research
- Individual differences in problem solving
- Within-study group comparison
How the paradigm works
The remote associates test comes from Mednick's (1962) account of creative thinking as forming remote associations into new combinations. Because each problem has one answer that links words that seem unrelated, it is usually described as a measure of convergent thinking. Each problem shows three cue words, and the participant types the one word that forms a familiar compound word or two-word phrase with each of them, as in the compound problems of Bowden and Jung-Beeman (2003). The solution may come before or after a cue.
More about the paradigm
The 40 problems were written for this test, in three bands of rough difficulty; no norm is claimed for them. A session shows 30 by default, or 20 or 40 as a setting, drawn evenly from the bands. The session's random seed draws and orders the problems within each band, and the bands run from easier to harder. Each problem has 30 seconds by default, 15 to 60 as a setting, with a countdown bar; 30 seconds is the longest limit in Bowden and Jung-Beeman. An answer counts when it is submitted within the limit. A plural counts as the word, and one spelling slip is accepted in a word of 4 or more letters unless the typed word is another English word. Two practice problems outside the set have feedback; the test has none. The words are English; sessions run in other languages are flagged for review. Scores support research and do not diagnose.
Scores
- Problems Solvedcount
- Problems whose solution was submitted within the time limit.
- Higher means more problems solved
- Percent Solved%
- Problems solved as a percent of those presented.
- Higher means more problems solved; use it to compare sessions of different lengths
- Median Solution Timems
- Median time from a problem's appearance to its submission on solved problems, leaving out answers under 100 ms.
- Lower means faster solutions; it covers solved problems only
- Wrong Answerscount
- Submitted answers that were not the solution.
- Describes how readily answers were offered
- Time-outscount
- Problems with no submitted answer when the time ran out.
- A normal result on hard problems
- Problems Attempteditems
- Problems with a submitted answer, or with text in the box when the time ran out.
- Fewer than 20% of those presented is invalid
- Problems Presenteditems
- Problems shown in the scored run.
- Fewer than 75% of those set is invalid
- Problems Setitems
- Problems the session set out to show.
- 20, 30 or 40
- Answers Under 100 mscount
- Answers submitted within 100 ms.
- Expected to be 0
- Same Answer Countcount
- The largest number of problems given one identical answer.
- 10 or more is flagged for review
- Paste Attemptscount
- Blocked attempts to paste or drop text into the answer box.
- Any is flagged for review
Use and evidence
217 papers from 2021 to 2025 name this task in their title or abstract, down 10% compared with 2016 to 2020. That ranks it 117 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 48 |
| 2024 | 29 |
| 2023 | 47 |
| 2022 | 45 |
| 2021 | 48 |
| 2020 | 53 |
| 2019 | 57 |
| 2018 | 62 |
| 2017 | 32 |
| 2016 | 37 |
| 2015 | 30 |
| 2014 | 34 |
| 2013 | 16 |
| 2012 | 31 |
| 2011 | 15 |
| 2010 | 16 |
| 2009 | 12 |
| 2008 | 9 |
| 2007 | 12 |
| 2006 | 7 |
| 2005 | 3 |
| 2004 | 4 |
| 2003 | 8 |
| 2002 | 2 |
| 2001 | 0 |
| 2000 | 3 |
Use grade C
100 to 299 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
- 4
- Papers on reliability or validity
- 91
- Registered clinical studies using it as an outcome
- 9
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.
- Problems
- 40 problems written for this test: 14 in an easier band, 13 in a middle band and 13 in a harder band, by rough judgment. No norm is claimed.
- Stimulus
- Three lower-case cue words, each forming a familiar compound word or two-word phrase with one solution word. Each solution is used once and shares no stem with its cues.
- Session
- 30 problems by default, or 20 or 40 as a setting, drawn evenly from the bands. The session's seed draws and orders the problems within each band, and the bands run from easier to harder. All of it is saved.
- Time limit
- 30 seconds per problem by default, 15 to 60 as a setting, with a countdown bar. 30 seconds is the longest limit Bowden and Jung-Beeman (2003) report.
- Answer
- Typed with autocorrect off, and submitted with Enter or a button. A plural counts as the word, and one spelling slip is accepted in a word of 4 or more letters unless the typed word is another English word.
- Practice
- 2 problems outside the set, each followed by feedback. The test has no feedback.
The problems were written for this test; no published problem set is used. The words are English; sessions run in other languages are flagged for review.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: remote_associates_problem from the scored run, one typed answer per problem within its time limit. Typed answers have no chance level, so accuracy is not judged against chance, and a time-out is a normal result on a hard problem. Fewer than 75% of the problems set presented invalidates, and so does fewer than 20% of those presented attempted. Typed answers under 200 ms on at least half of 4 or more answers invalidate. One identical answer on 10 or more problems, a blocked paste, or a session run in another language is flagged for review. 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. A low score is the result the test measures and never invalidates, but with almost nothing typed, a low score cannot be told from no effort, so too few attempts invalidate. A score over a much shorter session cannot be compared with a full one. The solutions all differ, so one answer typed for many problems cannot follow them; it is flagged. A blocked paste can mean answers were being looked up; it is flagged. 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 problems are English compounds. 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 |
|---|---|---|
| Problems | 30 problems (standard) | 20 problems, 30 problems (standard), 40 problems |
| Time per problem (seconds) | 30 | 15 to 60 |
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 Remote Associates Test in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Mednick, 1962; Bowden & Jung-Beeman, 2003), with the standard settings (problems: 30 problems; time per problem (seconds): 30). The main score was problems solved (problems whose solution was submitted within the time limit).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Mednick, S. A. (1962). The associative basis of the creative process. Psychological Review, 69(3), 220-232. https://doi.org/10.1037/h0048850
- Bowden, E. M., & Jung-Beeman, M. (2003). Normative data for 144 compound remote associate problems. Behavior Research Methods, Instruments, & Computers, 35(4), 634-639. https://doi.org/10.3758/BF03195543
- Olteţeanu, A.-M., & Falomir, Z. (2015). Pattern Recognition Letters, 67, 81-90. https://doi.org/10.1016/j.patrec.2015.05.015
- Olteţeanu, A.-M., Schultheis, H., & Dyer, J. B. (2018). Behavior Research Methods, 50(5), 1971-1980. https://doi.org/10.3758/s13428-017-0965-8
Source papers
- Bowden EM, Jung-Beeman M (2003). Normative data for 144 compound remote associate problems. Behavior Research Methods, Instruments, & Computers, 35(4):634-639.Paradigm source, Adults
Questions
How long does the Remote Associates Test take?
About 8 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?
Problems Solved, Percent Solved, Median Solution Time, Wrong Answers, Time-outs, Problems Attempted, Problems Presented, Problems Set, Answers Under 100 ms, Same Answer Count and Paste Attempts. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Remote Associates Test used in research?
217 papers from 2021 to 2025 name it in their title or abstract, down 10% compared with 2016 to 2020. That ranks it 117 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 Remote Associates Test have?
In the published research, 4 papers report its test-retest reliability and 9 registered clinical studies use it as an outcome (evidence grade C in the ConductCognition test census).
Which settings can a researcher change?
Problems (20 problems, 30 problems (standard), 40 problems) and time per problem (seconds) (15 to 60). 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.