Reasoning
Arithmetic Stories Task
Also called: arithmetic word problems, math word problems, compare problems, consistent and inconsistent language problems, consistency effect task
Measure arithmetic word problem solving: participants type the answers to short one- and two-step stories whose key word matches or contradicts the operation, scored as accuracy by wording and the share of wrong answers that follow the key word.
6-10 minUntimed
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
- 520
- Evidence grade
- C
- Registered clinical studies
- 1
The red jar holds 34 buttons. It holds 8 more buttons than the blue jar. How many buttons does the blue jar hold?
The red jar holds 34 buttons. It holds 8 more buttons than the blue jar. How many buttons does the blue jar hold?
26What it measures
Measures arithmetic word problem solving. Participants read short stories and type the number that answers each one's question, in one or two steps. Half the stories word their comparison so that its key word (more, less) matches the operation and half so that it contradicts it. The main score is accuracy; the share of wrong answers that follow the key word is scored too.
Constructs
- Arithmetic word problem solving
- Comprehension of relational statements
- Problem representation
- Key-word strategy
Research uses
- Mathematics education research
- Numerical cognition research
- Reading comprehension research
- Comparing consistent and inconsistent wordings within a study
How the paradigm works
Arithmetic word problems measure how people turn a short story into arithmetic. In compare problems a relational sentence states one amount in terms of another, and solvers make more reversal errors, adding where they should subtract or the reverse, when its key word (more than, less than) contradicts the operation the story needs than when it matches it (Lewis and Mayer, 1987). With university students their account held for two-step problems, which put more demand on the solver, and not for one-step ones (Verschaffel, De Corte and Pauwels, 1992), and unsuccessful solvers tend to build their answer from the numbers and key words rather than from a model of the situation (Hegarty, Mayer and Monk, 1995). In this task people read 16 short stories written in each language, 8 one-step and 8 two-step that multiply after the comparison, and type the number that answers each. Every story is worded so that its key word matches the operation in some sessions and contradicts it in others.
Scores
- Accuracy%
- Right answers over the stories shown, 16 in a complete run.
- Higher is better
- One-Step Accuracy%
- Accuracy on the 8 one-step stories.
- Higher is better
- Two-Step Accuracy%
- Accuracy on the 8 two-step stories, which multiply after the comparison.
- Higher is better
- Consistent Wording Accuracy%
- Accuracy on the 8 stories whose key word matches the operation.
- Higher is better
- Inconsistent Wording Accuracy%
- Accuracy on the 8 stories whose key word contradicts the operation.
- Higher is better
- Consistency Effect%
- Consistent minus inconsistent accuracy.
- Higher means the contradicting wording cost more
- Accuracy by Steps and Wording%
- Accuracy on the one-step and two-step stories in each wording, such as twoStep-inconsistent.
- Lower on inconsistent stories means the wording mattered
- Key-Word Answers%
- Share of the wrong typed answers that follow the key word instead of the story: adding for more, or subtracting for less, where the story needs the opposite.
- Higher means more answers built from the key word
- Reversals, Consistent Wording%
- Answers with the comparison reversed, over the consistent stories.
- Usually low
- Reversals, Inconsistent Wording%
- Answers with the comparison reversed, over the inconsistent stories.
- Higher than on consistent stories means the wording drew reversals
- Key-Word Errorscount
- Inconsistent stories answered with the operation their key word suggests.
- Informational
- Reversals Against the Key Wordcount
- Consistent stories answered with the opposite of the operation their key word suggests.
- Informational
- First Step Onlycount
- Two-step stories answered with the result of the comparison alone, in either direction.
- Informational
- Other Errorscount
- Wrong typed numbers that match no named error, such as slips in the arithmetic.
- Informational
- Unanswered Storiescount
- Stories left blank or without a number that can be read.
- Quality check
- Median Answer Timems
- Median time from a story appearing to its answer, for answers of at least 100 ms.
- Informational; the task has no time limit
- Median Time, Consistent Wordingms
- Median time to a right answer on the consistent stories, for answers of at least 100 ms.
- Informational
- Median Time, Inconsistent Wordingms
- Median time to a right answer on the inconsistent stories, for answers of at least 100 ms.
- Longer than on consistent stories means the wording took extra work
- Answers Under 100 mscount
- Answers submitted less than 100 ms after their story appeared.
- 0 expected
- Stories Answeredcount
- Stories with a typed number that can be read.
- Fewer than half the planned stories invalidates
- Storiescount
- Stories shown in the scored run.
- Compare with planned stories
- Planned Storiescount
- 16: 8 one-step and 8 two-step, half of each in each wording.
- Informational
- Most Repeated Answercount
- The largest number of stories given one identical number.
- Half the planned stories or more is flagged
Use and evidence
520 papers from 2021 to 2025 name this task in their title or abstract, up 14% compared with 2016 to 2020. That ranks it 56 of 711 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 130 |
| 2024 | 132 |
| 2023 | 89 |
| 2022 | 98 |
| 2021 | 71 |
| 2020 | 105 |
| 2019 | 127 |
| 2018 | 75 |
| 2017 | 76 |
| 2016 | 75 |
| 2015 | 65 |
| 2014 | 72 |
| 2013 | 79 |
| 2012 | 61 |
| 2011 | 56 |
| 2010 | 61 |
| 2009 | 51 |
| 2008 | 39 |
| 2007 | 35 |
| 2006 | 24 |
| 2005 | 34 |
| 2004 | 26 |
| 2003 | 30 |
| 2002 | 19 |
| 2001 | 16 |
| 2000 | 14 |
Use grade B
300 to 999 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
- 110
- Registered clinical studies using it as an outcome
- 1
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.
Note on the counts: mostly classroom and math education studies.
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 14 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Stories
- 16 stories written for this library in each language: 8 one-step and 8 two-step, which multiply after the comparison, such as how many points 4 purple tickets cost if an orange ticket costs 12 points and a purple one 3 points more.
- Wording
- Each story relates one amount to another by more or less. In the consistent wording the key word matches the operation; in the inconsistent wording it contradicts it, as when the orange ticket costs 3 points more than a purple one. The seed gives half the stories of each kind each wording, so every story appears in both across sessions.
- Answers
- Typed numbers, all whole and under 100 before the second step. One practice story with feedback, then no time limit and no feedback.
- Order
- The 16 stories in an order drawn from the session's seed.
- Scores
- Accuracy (the main score), by steps and by wording, the consistency effect, and the share of wrong answers that follow the key word instead of the story. Reversals and answers that stop after the first step are counted.
The stories and numbers are this library's own. The scores describe answers to these stories, support research and make no grade-level, school-readiness or diagnostic claim.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: word_problems_story from the run with the most of them, latest on ties, one typed number per story. Typed answers have no chance level, so accuracy is not judged against chance, and stories have no time limit, so none can be missed. Fewer than half the planned stories answered invalidates. An incomplete run is flagged, and so is one number typed for half the stories or more, and right answers on fewer than half the stories whose key word matches the operation. Answers under 200 ms on at least half of four or more answered stories invalidate. A hidden tab is flagged.
Why these rules
There is no correct option to compare with chance, so accuracy is not judged against chance. A score over fewer than half the stories cannot be compared with a score over all of them, so it invalidates. Wrong answers that follow the key word are what the test measures and never invalidate. The stories have different answers, so one number typed for half of them cannot follow them; it is flagged. On the consistent stories the key word and the story agree, so missing most of them means the stories or the arithmetic were not followed; 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. Stories stay on screen until they are answered, so a hidden tab does not change what was shown, but it can mean a calculator or help was used; it is flagged.
Running this test
Settings you can change
This test runs with fixed settings.
Need a different version of this test?
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.
Cite and describe
Participants completed the Arithmetic Stories Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Lewis & Mayer, 1987; Verschaffel, De Corte & Pauwels, 1992; Hegarty, Mayer & Monk, 1995). The main score was accuracy (right answers over the stories shown, 16 in a complete run).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Lewis, A. B., & Mayer, R. E. (1987). Students' miscomprehension of relational statements in arithmetic word problems. Journal of Educational Psychology, 79(4), 363-371. https://doi.org/10.1037/0022-0663.79.4.363Compare problems in consistent and inconsistent language, a compare step and then a second step, where college students made more reversal errors in the inconsistent form.
- Verschaffel, L., De Corte, E., & Pauwels, A. (1992). Solving compare problems: An eye movement test of Lewis and Mayer's consistency hypothesis. Journal of Educational Psychology, 84(1), 85-94. https://doi.org/10.1037/0022-0663.84.1.85One-step and two-step compare problems with university students and third graders, where the consistency account held when the task put demands on the solver.
- Hegarty, M., Mayer, R. E., & Monk, C. A. (1995). Comprehension of arithmetic word problems: A comparison of successful and unsuccessful problem solvers. Journal of Educational Psychology, 87(1), 18-32. https://doi.org/10.1037/0022-0663.87.1.18Unsuccessful solvers building their plan from numbers and key words, successful ones from a model of the situation.
Source papers
Questions
How long does the Arithmetic Stories Task 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?
Accuracy, One-Step Accuracy, Two-Step Accuracy, Consistent Wording Accuracy, Inconsistent Wording Accuracy, Consistency Effect, Accuracy by Steps and Wording, Key-Word Answers, Reversals, Consistent Wording, Reversals, Inconsistent Wording, Key-Word Errors, Reversals Against the Key Word, First Step Only, Other Errors, Unanswered Stories, Median Answer Time, Median Time, Consistent Wording, Median Time, Inconsistent Wording, Answers Under 100 ms, Stories Answered, Stories, Planned Stories and Most Repeated Answer. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Arithmetic Stories Task used in research?
520 papers from 2021 to 2025 name it in their title or abstract, up 14% compared with 2016 to 2020. That ranks it 56 of 711 test families in the ConductCognition test census (use grade B). Source: ConductCognition test census 0.58.0, counted 14 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Arithmetic Stories Task have?
In the published research, 1 paper reports its test-retest reliability and 1 registered clinical study uses it as an outcome (evidence grade C in the ConductCognition test census).
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.