Language
Phoneme Monitoring Task
Also called: phoneme detection, initial phoneme monitoring, target phoneme detection, phoneme monitoring paradigm
Measure how fast a speech sound is picked out in spoken sentences: a press the moment a word starts with b.
7-9 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
- 19
- Evidence grade
- D
- Registered clinical studies
- 2
What it measures
Measures how fast a speech sound is picked out while listening to sentences. Participants hear English sentences from a synthetic voice and press the space bar as soon as a word starts with b; a third of the sentences have none, and a few questions check that they listened for meaning. The main score is the median detection time.
Constructs
- Speech perception during sentence comprehension
- Use of sentence context
- Listening speed
Research uses
- Sentence processing research
- Second-language listening research
- Within-study comparison of listening conditions
How the paradigm works
The Phoneme Monitoring Task measures how fast a speech sound is picked out while a person listens to sentences for their meaning. In Foss's (1969) method, listeners press a key as soon as they hear a word that starts with a set sound; the time from that sound to the press grows when the sentence is harder to follow at that point, so detection time is a window on sentence processing as it happens. This version plays English sentences written for the platform, said by a synthetic voice, and the set sound is b. The test starts with the sound setup: the sentences load and the participant turns the device's volume until a steady tone is comfortable. Wired headphones or the device's speakers both work; wireless headphones add a delay the browser may not report.
More about the paradigm
Thirty-two of the 48 test sentences hold one word that starts with b followed by a vowel, after at least five words; the other 16 have no b, so a press is never safe without listening. Each target word comes after a sentence that leads up to it or one that does not, with the same number of words before it: the design Foss and Gernsbacher (1983) left standing after showing that comparing different words in one sentence frame, as Morton and Long (1976) did, lets the words' own sounds pass for a context effect. Two lists swap which sentence each word gets, and every session hears every word once, 16 in each condition, in an order drawn from its seed. Each press is timed from the release of the b, measured on each recording, the point phoneme monitoring studies time a stop from (Frauenfelder and Segui, 1989; Mehta and Cutler, 1988); a press from 100 to 1,500 ms after it counts. After 12 of the sentences a yes or no question asks what the sentence said, which keeps listening on meaning, as in the online replication of Beier and Ferreira (2022). Six practice sentences with feedback come first.
The main score is the median detection time over the target words found. The predictability effect, the median for words after a neutral sentence less the median after a leading one, is reported beside it with the hit rate, false alarms on sentences without a b, early and late presses, and the share of questions answered right. Controlled studies put the effect at tens of milliseconds (23 to 73 ms across Morton and Long, 1976, Mehta and Cutler, 1988, and Nix and colleagues, 1993), smaller than the spread of one person's times, so it is read across a group. Synthetic speech slowed detection in Nix and colleagues (1993) and left its pattern as it was; online detection is slower than in the laboratory. The sentences are English in every session: a session in another language is flagged for review. Scores are research measures and do not diagnose.
Scores
- Median Detection Timems
- Median time from the release of the target's b to the press, over the target words found (presses 100 to 1,500 ms after the release).
- Lower is faster detection
- Median Time, Predictable Wordsms
- The same median for target words after a sentence that leads up to them.
- Informational
- Median Time, Unpredictable Wordsms
- The same median for target words after a sentence that does not lead up to them.
- Informational
- Predictability Effectms
- Median time for unpredictable target words less the median for predictable ones.
- Positive means predictable words were found faster; read across a group
- Hit Rate%
- Share of the 32 target sentences with a press 100 to 1,500 ms after the release of the b.
- Higher is more targets found
- Hit Rate, Predictable Words%
- Share of the predictable target words found.
- Informational
- Hit Rate, Unpredictable Words%
- Share of the unpredictable target words found.
- Informational
- False Alarm Rate%
- Share of the 16 sentences without a b that drew a press.
- Lower is better; presses that do not separate the two kinds of sentence are flagged
- Early Pressescount
- Presses before the release of the b on target sentences.
- Informational
- Anticipationscount
- Target sentences whose first press came under 100 ms after the release, too soon to be a reaction to it.
- Follows the timed-test rule
- Late Responsescount
- Target sentences whose first press came 1,500 ms or more after the release.
- Informational
- Missescount
- Target sentences with no press after the release.
- Follows the timed-test rule
- Question Accuracy%
- Share of the yes or no questions about the sentences answered right.
- Under 0.8 is flagged for review
- Questions Answeredcount
- Questions about the sentences answered in the scored run.
- Informational
- Completed Sentencestrials
- Test sentences played in the scored run.
- Fewer than planned is flagged for review
- Planned Sentencestrials
- Test sentences the run set out to play.
- Informational
- Session in Another Language0/1
- 1 when the session ran in a language other than English, while the sentences are English.
- 1 is flagged for review
- Reported Output Delayms
- Output delay the device reported to the browser; empty when the browser does not report it.
- Informational
Use and evidence
19 papers from 2021 to 2025 name this task in their title or abstract, down 46% compared with 2016 to 2020. That ranks it 403 of 495 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 5 |
| 2024 | 3 |
| 2023 | 1 |
| 2022 | 4 |
| 2021 | 6 |
| 2020 | 9 |
| 2019 | 9 |
| 2018 | 7 |
| 2017 | 3 |
| 2016 | 7 |
| 2015 | 11 |
| 2014 | 5 |
| 2013 | 6 |
| 2012 | 6 |
| 2011 | 11 |
| 2010 | 3 |
| 2009 | 8 |
| 2008 | 7 |
| 2007 | 4 |
| 2006 | 8 |
| 2005 | 3 |
| 2004 | 8 |
| 2003 | 2 |
| 2002 | 7 |
| 2001 | 9 |
| 2000 | 8 |
Use grade E
Fewer than 30 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 0
- Papers on reliability or validity
- 2
- Registered clinical studies using it as an outcome
- 2
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 3 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Sound setup
- The sentences load and the participant turns the device's volume until a steady 1 kHz tone is comfortable. Wired headphones or the device's speakers both work, so there is no headphone check.
- Sentences
- 48 English sentences written for this test, said by a synthetic voice, not recorded from a person; 2.6 to 4.9 s each. In 32 of them one word starts with b followed by a vowel, after at least five words; the other 16 have no b.
- Targets
- Each target word comes after a sentence that leads up to it or one that does not, with the same number of words before it. Two lists swap which sentence each word gets; each session's list comes from its seed, so every session hears every word once, 16 in each condition.
- Trial
- The sentence starts 1,000 ms into the trial, and a press counts from 100 to 1,500 ms after the release of the b, measured on each recording. The trial ends 1,000 ms after the sentence.
- Questions
- A yes or no question about what the sentence said after 12 of the 48 sentences.
- Practice
- 6 sentences with feedback, then the test with a rest halfway.
- Level
- Every sentence 12 dB below the volume-setting tone, a level relative to the volume the participant set.
The voice is synthetic and the sentences are English; sessions run in other languages are flagged for review. Which sentence makes a word predictable was set by the writer and has not been measured with readers. The predictability effect is small in controlled studies, so read it across a group, not for one person. 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: phoneme_monitoring_test from the scored run, a space-bar press on the 32 sentences with a word that starts with b and none on the 16 without; on a target sentence the response is the first press at or after the release of its b, timed from it. Missed targets and responses under 100 ms follow the timed-test rule; presses concentrated on the sentences without a b invalidate and presses that do not separate the two kinds are flagged. Questions answered right less than 80% of the time, a run short of its planned sentences, or a session run in a language other than English is flagged for review.
Why these rules
Missed responses and responses under 100 ms keep the platform's timed-test rule: more than max(3, 25%) missed responses or more than max(2, 20%) responses under 100 ms invalidate, and any is flagged. A response under 100 ms cannot be a reaction to the stimulus. Responses concentrated on trials that needed none (one-sided Fisher exact test, p < .05) mean the instructions were reversed and invalidate; responses that do not separate the two trial types are flagged. The questions check that the sentences were followed for their meaning, as phoneme monitoring asks (Beier and Ferreira, 2022, kept participants with 80% or more right). 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.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
This test runs with fixed settings.
Need a different version of this test?
Devices and languages
- Timed test: participants need a desktop or laptop with a mouse or trackpad. 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.
- 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 Phoneme Monitoring Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Foss, 1969; Foss and Gernsbacher, 1983). The main score was median detection time (median time from the release of the target's b to the press, over the target words found (presses 100 to 1,500 ms after the release)).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Foss, D. J. (1969). Decision processes during sentence comprehension: Effects of lexical item difficulty and position upon decision times. Journal of Verbal Learning and Verbal Behavior, 8(4), 457-462. https://doi.org/10.1016/S0022-5371(69)80089-7
- Foss, D. J., & Gernsbacher, M. A. (1983). Cracking the dual code: Toward a unitary model of phoneme identification. Journal of Verbal Learning and Verbal Behavior, 22(6), 609-632. https://doi.org/10.1016/S0022-5371(83)90365-1Why the same target word follows both kinds of sentence: different words can carry the difference by themselves.
- Morton, J., & Long, J. (1976). Effect of word transitional probability on phoneme identification. Journal of Verbal Learning and Verbal Behavior, 15(1), 43-51. https://doi.org/10.1016/S0022-5371(76)90005-0Faster detection of targets in predictable words.
- Mehta, G., & Cutler, A. (1988). Detection of target phonemes in spontaneous and read speech. Language and Speech, 31(2), 135-156. https://doi.org/10.1177/002383098803100203Timing from the release of the target.
- Nix, A. J., Mehta, G., Dye, J., & Cutler, A. (1993). Phoneme detection as a tool for comparing perception of natural and synthetic speech. Computer Speech & Language, 7(3), 211-228. https://doi.org/10.1006/csla.1993.1011Detection in synthetic speech: slower, with the same pattern of effects.
- Beier, E. J., & Ferreira, F. (2022). Replication of Cutler, A., & Fodor, J. A. (1979). Semantic focus and sentence comprehension. Cognition, 7(1), 49-59. Journal of Memory and Language, 126, 104339. https://doi.org/10.1016/j.jml.2022.104339Phoneme monitoring run online, with comprehension questions.
Source papers
- Morton J, Long J (1976). Effect of word transitional probability on phoneme identification. Journal of Verbal Learning and Verbal Behavior, 15(1):43-51.N 24, Adults
Questions
How long does the Phoneme Monitoring Task take?
About 7 to 9 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?
Median Detection Time, Median Time, Predictable Words, Median Time, Unpredictable Words, Predictability Effect, Hit Rate, Hit Rate, Predictable Words, Hit Rate, Unpredictable Words, False Alarm Rate, Early Presses, Anticipations, Late Responses, Misses, Question Accuracy, Questions Answered, Completed Sentences, Planned Sentences, Session in Another Language and Reported Output Delay. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Phoneme Monitoring Task used in research?
19 papers from 2021 to 2025 name it in their title or abstract, down 46% compared with 2016 to 2020. That ranks it 403 of 495 test families in the ConductCognition test census (use grade E). Source: ConductCognition test census 0.22.0, counted 3 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Phoneme Monitoring Task have?
In the published research, 0 papers report its test-retest reliability and 2 registered clinical studies use it as an outcome (evidence grade D in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
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.