Language
Writing Process Task
Also called: keystroke logging task, keystroke logging in writing, writing process logging, pause and burst analysis, writing fluency task
Record how a text is written: participants write to an everyday prompt in a box that times each key press and records each insertion and deletion, scored for pauses, bursts of typing, revisions and speed.
12-14 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
- 186
- Evidence grade
- C
- Registered clinical studies
- 0
What it measures
Records how a short text is written. Participants write to an everyday prompt in a box on the task page, which times each key press and records each insertion and deletion. The main score is the mean burst length, the characters typed between pauses of 2 seconds or more; the text is saved for the researcher and never scored.
Constructs
- Writing fluency
- Planning and revising while writing
- Typing fluency
Research uses
- Writing process research
- Research on first and second language writing
- Within-study group comparison
How the paradigm works
Keystroke logging records when each key is pressed and how the text changes while a person writes, so the process of writing can be studied beside the finished text (Leijten and Van Waes, 2013). Chenoweth and Hayes (2001) found that the strings of words writers produce between pauses or revisions grow longer with experience in a language. Conijn, Roeser and van Zaanen (2019) compared copying, an email and an academic summary and found that the times between words and between sentences and the number of revisions changed with the task, while the times within words did not.
More about the paradigm
This task gives a prompt about an everyday experience, written for each language, and a writing box on the task page, with 10 minutes for a text by default; the participant can press Finish sooner. While the box has focus, the page records the time each key is pressed and released with the kind of key (a character, the space bar, Enter, Backspace, Delete, a cursor key, an input method key or another key), not the key itself, and each change to the text with its time, its input type as the browser reports it, its place, the number of characters removed and the text added. It reads no other page, window or clipboard: a paste or a drop is recorded with its length only. The finished text is saved for the researcher with the session's data; it is never scored and never shown to another participant. A pause is a gap of 2,000 ms or more from one key press to the next, a threshold keystroke-logging studies often use (Roeser and colleagues, 2024), and a burst is the typing between two pauses. The main score is the mean burst length in characters; the task also reports pauses, the time between key presses, writing speed, revisions, revisions inside the text and the share of typing kept in the final text. With an input method, as for Chinese and Japanese, keys are recorded as the browser reports them, often as input method keys, and characters count when a composition is committed; each composition is recorded with its start, end and committed text.
Scores
- Mean burst lengthcount
- Mean characters typed or composed per burst, the typing between two pauses of 2,000 ms or more, over bursts that added at least one character. Pasted text does not count.
- Informational. Longer bursts and shorter pauses have been related to more fluent writing (Conijn and colleagues, 2019); compare within a language, input method and prompt
- Burstscount
- Bursts that added at least one character, over all texts.
- Informational
- Pausescount
- Gaps of 2,000 ms or more from one key press to the next, over all texts. The time before the first key press is not a pause.
- Informational
- Median pausems
- Median length of the pauses.
- Informational
- Time in pauses%
- Time in pauses divided by the time from the first key press to the last, summed over texts.
- Informational
- Median time between key pressesms
- Median time from one key press to the next, pauses included.
- Informational. Depends on typing skill and the input method
- Writing speedcharacters/min
- Characters in the final texts per minute of writing time, from the prompt appearing to the end of each text.
- Informational; compare within a language and input method
- Revisionscount
- Runs of changes that delete text (Backspace, Delete, cutting, or typing over selected text), each run counted once.
- Informational
- Revisions inside the textcount
- Runs of changes made before the end of the text, going back into what was already written, each run counted once.
- Informational
- Deleted characterscount
- Characters removed by all changes.
- Informational
- Share of typing kept%
- Characters in the final texts divided by all characters added, typed, composed or pasted.
- Informational. 1 means nothing added was deleted
- Time to first keyms
- Median time from the prompt appearing to the first key press, over texts with a key press.
- Informational. Includes reading the prompt
- Final characterscount
- Characters in the final texts.
- Informational
- Texts writtencount
- Texts whose final text has a character other than a space.
- Quality check
- Planned textscount
- 1 to 3 texts.
- Informational
- Key pressescount
- Key presses recorded in the writing box, leaving out modifier keys, Tab, Escape and shortcuts with Ctrl or Command; a key held down counts once.
- Quality check; none is invalid and fewer than 100 is flagged
- Pastescount
- Pastes and drops into the writing box. Their text is recorded only as its length.
- 0 expected
- Pasted characterscount
- Characters added by pastes and drops.
- Quality check; more than half the final text is flagged
- Texts with an input methodcount
- Texts in which an input method composed characters, as for Chinese and Japanese. Their keys are recorded as the browser reports them, often as input method keys, and their characters count when a composition is committed, so the burst lengths count committed characters.
- Informational; compare within an input method
- Focus lossescount
- Times the writing box lost focus, for example to another window.
- 0 expected; any is flagged
Use and evidence
186 papers from 2021 to 2025 name this task in their title or abstract, up 34% compared with 2016 to 2020. That ranks it 181 of 711 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 51 |
| 2024 | 33 |
| 2023 | 34 |
| 2022 | 37 |
| 2021 | 31 |
| 2020 | 33 |
| 2019 | 42 |
| 2018 | 29 |
| 2017 | 22 |
| 2016 | 13 |
| 2015 | 25 |
| 2014 | 9 |
| 2013 | 9 |
| 2012 | 17 |
| 2011 | 9 |
| 2010 | 9 |
| 2009 | 11 |
| 2008 | 9 |
| 2007 | 3 |
| 2006 | 13 |
| 2005 | 7 |
| 2004 | 0 |
| 2003 | 2 |
| 2002 | 2 |
| 2001 | 0 |
| 2000 | 0 |
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
- 1
- Papers on reliability or validity
- 31
- Registered clinical studies using it as an outcome
- 0
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: includes translation studies and a few on keyboard monitoring.
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 17 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Prompts
- Three everyday topics written for each language, used in this order: a small everyday problem the participant solved, a place where they like to spend time, and something they learned to do. One text is the standard, and a study can ask for up to three. A practice prompt, a rainy day, comes first.
- Writing box
- A box on the task page with the prompt above it and a clock showing the time left: 10 minutes for each text by default, or 5, 15 or 20. The participant can press Finish sooner. Spelling checks and automatic corrections are off.
- Recording
- While the box has focus: the time each key is pressed and released, with its kind (a character, the space bar, Enter, Backspace, Delete, a cursor key, an input method key or another key) but not the key itself, and each insertion or deletion with its place in the text. Nothing outside the box is read: a paste is recorded as a paste with its length. With an input method, as for Chinese and Japanese, keys are recorded as the browser reports them, and each composition with its start, end and committed text.
- Text
- The finished text is saved with the session's data for the researcher. It is never scored and never shown to another participant.
- Scores
- Mean burst length (the characters typed between two pauses), pauses of 2,000 ms or more, the time between key presses, writing speed, revisions, revisions inside the text and the share of typing kept in the final text.
- Device
- A desktop or laptop with a physical keyboard.
Pauses, bursts and revisions describe how a text was produced, not how good it is. They change with the writing task (Conijn and colleagues, 2019) and can differ by language and input method, so they compare best within a study that gives everyone the same prompts.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: writing_process_test texts (practice excluded). A text has no right answer and no chance level, and the text itself is never scored. No key press in the writing box invalidates. Fewer than 100 key presses, fewer texts written than planned, pasted text making up more than half of the final text, a lost focus and a log that reached its limit are flagged.
Why these rules
There is no correct option to compare with chance, so accuracy is not judged against chance. The scores are the times between key presses and the changes they make, so with no key press there is nothing to score. Pauses and bursts rest on the typing, so short writing, pasted text or time away from the box make them less dependable.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Texts | 1 | 1 to 3 |
| Time per text | 10 minutes (standard) | 5 minutes, 10 minutes (standard), 15 minutes, 20 minutes |
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.
- 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 Writing Process Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Chenoweth and Hayes, 2001; Leijten and Van Waes, 2013; Conijn, Roeser and van Zaanen, 2019), with the standard settings (texts: 1; time per text: 10 minutes). The main score was mean burst length (mean characters typed or composed per burst, the typing between two pauses of 2,000 ms or more, over bursts that added at least one character).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Chenoweth, N. A., & Hayes, J. R. (2001). Fluency in writing: Generating text in L1 and L2. Written Communication, 18(1), 80-98. https://doi.org/10.1177/0741088301018001004Fluency in writing: the strings of words writers produce between pauses or revisions grow longer with experience in a language.
- Conijn, R., Roeser, J., & van Zaanen, M. (2019). Understanding the keystroke log: The effect of writing task on keystroke features. Reading and Writing, 32(9), 2353-2374. https://doi.org/10.1007/s11145-019-09953-8Bursts as typing without a gap of two seconds or more; the times between words and sentences and the revisions change with the writing task, while the times within words do not.
- Roeser, J., De Maeyer, S., Leijten, M., & Van Waes, L. (2024). Modelling typing disfluencies as finite mixture process. Reading and Writing, 37(2), 359-384. https://doi.org/10.1007/s11145-021-10203-zThe pause threshold: keystroke-logging studies often count a gap of 2 seconds or more between key presses as a pause.
Source papers
Questions
How long does the Writing Process Task take?
About 12 to 14 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?
Mean burst length, Bursts, Pauses, Median pause, Time in pauses, Median time between key presses, Writing speed, Revisions, Revisions inside the text, Deleted characters, Share of typing kept, Time to first key, Final characters, Texts written, Planned texts, Key presses, Pastes, Pasted characters, Texts with an input method and Focus losses. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Writing Process Task used in research?
186 papers from 2021 to 2025 name it in their title or abstract, up 34% compared with 2016 to 2020. That ranks it 181 of 711 test families in the ConductCognition test census (use grade C). Source: ConductCognition test census 0.58.0, counted 17 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Writing Process Task have?
In the published research, 1 paper reports its test-retest reliability and 0 registered clinical studies use it as an outcome (evidence grade C in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Texts (1 to 3) and time per text (5 minutes, 10 minutes (standard), 15 minutes, 20 minutes). 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.