Attention
Listening Effort Task
Also called: dual-task listening effort paradigm, dual-task paradigm for listening effort, listening effort dual task, secondary task reaction time, probe reaction time task, speech in noise dual-task
Measure how much a second task slows while people listen to speech in background talk: a sentence through headphones in quiet or over babble, a circle or a square to answer with a key, and the sentence's key words to choose.
12-15 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
- 735
- Evidence grade
- A
- Registered clinical studies
- 79
What it measures
Measures how much a second task slows down while people listen to speech in background talk. Through headphones, a synthetic voice says an everyday sentence in quiet or in babble at three levels, while a circle or a square appears on the screen and the participant answers it with a key; then the participant chooses the sentence's three key words from nine. The main score is how much slower the shapes were answered with babble than in quiet; the accuracy of the key words at each level comes with it. One way to measure listening effort, with a second task. A research measure, not a hearing test.
Constructs
- Listening effort
- Dual-task cost
- Divided attention
- Speech perception in noise
Research uses
- Hearing and speech perception research
- Attention and effort research
- Within-study group comparison
How the paradigm works
Listening effort measured with a second task. Listening to speech in noise takes mental effort, and a person's attention is limited, so the more of it listening takes, the less is left for a second task done at the same time, and the second task slows. Downs (1982) timed the answers to a visual probe while hearing-impaired listeners did a speech task with and without hearing aids; Sarampalis and colleagues (2009) used a visual reaction time task and a word memory task at several signal-to-noise ratios; Fraser and colleagues (2010) expressed the cost as a proportion of the second task done alone. Gagné, Besser and Lemke (2017) reviewed 29 such studies and found second tasks, timings and scores that differ so much that no one design can be called the best, and Kuchinsky, Gallun and Lee (2024) ask for control conditions and for care over how the two tasks share attention. Wu and colleagues (2016) found the reaction times of the second task longest at the signal-to-noise ratios where 30 to 50 percent of the speech was understood, and shorter again where less was. Studies of listening effort use self-report, behavioral and physiological measures (McGarrigle and colleagues, 2014), most of them pupil size, brain activity or ratings, not a second task; the measures often do not agree with one another (Gagné and colleagues, 2017), and this task is one method among several.
More about the paradigm
Set a comfortable volume and complete the 6-trial headphone check. A failed check stops the task by default; researchers can choose to continue with a quality flag.
First 12 shapes alone, with no sound: a plus sign, then after 1 to 2.5 s a circle or a square; press F for the circle and J for the square, or press its button. Then 3 practice sentences with feedback, in quiet, with babble 8 dB under the sentence and with babble 4 dB over it. Then four parts of 12 sentences each (8 to 14 as a setting), one part for each level, in an order drawn from the session's seed: quiet, and babble 8 dB under the sentence, 2 dB under it and 4 dB over it.
Each trial plays a short everyday sentence of 6 to 10 words, our own, said by the speech kit's synthetic voice in the session's language, with three key words: a person, a thing and a place. The babble is the kit's own, eight streams of its other sentences played at their own rates and summed, so no key word is ever in it; it starts half a second before the sentence and ends half a second after it. 45 to 80 percent of the way through the sentence, a circle or a square appears and stays for up to 1,500 ms; the participant answers it as fast as they can, and the sentence is the main task. When the sound has ended nine words appear, the three key words and six others from other sentences, two of each kind; choose three and press Done, with no time limit.
The sentence plays at -19 dB re the volume-setting tone at every level, so only the babble changes: -27, -21 and -15 dB. The sentence and the babble together stay under 0.9 of full scale in every language. The hardest level puts the babble 4 dB over the sentence, a ratio chosen to leave the key words above chance in a closed set of nine words; the share right at each level is reported, and a level near chance is flagged.
The median time of the correct shape answers is found at each level and for the shapes alone, leaving out answers under 100 ms. The main score is the mean, over the three babble levels, of the median time less the median time in quiet, in ms and as a percentage of the time in quiet. Also reported: each median, the cost of doing both tasks in quiet over the shapes alone, the share of key words right at each level and in all, the shape accuracy and misses, anticipations, completed trials and the sound check.
Scores
- Listening Effort Costms
- Mean over the three babble levels of the median time to answer a correct shape at that level less the median time in quiet.
- Higher means the shapes were answered more slowly with babble than in quiet, as when listening takes more effort; one way to measure effort, not a hearing level
- Listening Effort Cost, Percent%
- The same cost as a percentage of the median time in quiet, as Fraser and colleagues (2010) expressed a dual-task cost.
- Higher means a larger slowing relative to the time in quiet
- Shape Time Alonems
- Median time to answer a correct shape in the 12 shapes alone, with no sound.
- The speed of the second task by itself
- Shape Time in Quietms
- Median time to answer a correct shape while a sentence played in quiet.
- The speed of the second task with listening that needs little effort
- Shape Time, Babble 8 dB Underms
- Median time to answer a correct shape while a sentence played over babble 8 dB under it.
- Compare with the time in quiet
- Shape Time, Babble 2 dB Underms
- Median time to answer a correct shape while a sentence played over babble 2 dB under it.
- Compare with the time in quiet
- Shape Time, Babble 4 dB Overms
- Median time to answer a correct shape while a sentence played over babble 4 dB louder than it.
- Compare with the time in quiet; it can fall again when the sentences are mostly not understood (Wu and colleagues, 2016)
- Two-Task Cost in Quietms
- The median shape time in quiet less the median shape time alone.
- The cost of listening to a sentence at all, with no babble
- Key Words Right%
- Share of the key words, three on each trial, that were chosen, over all scored trials.
- Chance is a third; shows whether the sentences were heard
- Key Words Right in Quiet%
- The same share for the sentences in quiet.
- Below 60% is flagged: the sentences may not have been heard well
- Key Words Right, Babble 8 dB Under%
- The same share for the sentences over babble 8 dB under them.
- Informational
- Key Words Right, Babble 2 dB Under%
- The same share for the sentences over babble 2 dB under them.
- Informational
- Key Words Right, Babble 4 dB Over%
- The same share for the sentences over babble 4 dB louder than them.
- 40% or less is flagged: close to the third that guessing gives
- Shapes Right%
- Share of the shapes answered during the sentences that were answered with the right key or button.
- Close to 1 when the keys were used as asked; below chance invalidates
- Shapes Not Answeredcount
- Shapes during the sentences that were not answered within 1,500 ms.
- Any is flagged; more than a quarter of the shapes invalidates
- Anticipationscount
- Shape answers under 100 ms, in the shapes alone and during the sentences, too fast to be reactions to the shape.
- More than a fifth of the shapes makes the session invalid
- Completed Trialstrials
- Sentences answered in the scored run.
- Fewer than planned is flagged for review
- Planned Trialstrials
- Sentences the run set out to play.
- Informational
- Headphone Check Passed0/1
- 1 when the headphone check passed (at least 5 of 6 right), 0 when it did not.
- 0 stops the test unless the study set it to continue
- Headphone Check Correctcount
- Headphone check trials answered right, of 6.
- 5 or 6 passes
- Reported Output Delayms
- Output delay the device reported to the browser; empty when the browser does not report it.
- Informational
Use and evidence
735 papers from 2021 to 2025 name this task in their title or abstract, up 71% compared with 2016 to 2020. That ranks it 37 of 711 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 213 |
| 2024 | 175 |
| 2023 | 143 |
| 2022 | 101 |
| 2021 | 103 |
| 2020 | 110 |
| 2019 | 82 |
| 2018 | 101 |
| 2017 | 70 |
| 2016 | 66 |
| 2015 | 59 |
| 2014 | 31 |
| 2013 | 40 |
| 2012 | 10 |
| 2011 | 21 |
| 2010 | 17 |
| 2009 | 12 |
| 2008 | 9 |
| 2007 | 4 |
| 2006 | 4 |
| 2005 | 1 |
| 2004 | 1 |
| 2003 | 0 |
| 2002 | 1 |
| 2001 | 1 |
| 2000 | 0 |
Use grade B
300 to 999 papers from 2021 to 2025.
Evidence grade A
At least 25 papers report its test-retest reliability and at least 25 registered clinical studies use it as an outcome.
- Papers on test-retest reliability
- 34
- Papers on reliability or validity
- 158
- Registered clinical studies using it as an outcome
- 79
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: most studies measure pupil size, brain activity or ratings; the dual-task method is one of several.
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 16 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Sound setup
- The participant turns the device's volume until a steady 1 kHz tone is comfortable, then takes a 6-trial headphone check (Woods and colleagues, 2017). By default a failed check ends the test; its result is saved either way.
- Shapes
- A circle or a square appears in a box on the screen and is answered with F for the circle and J for the square, or by pressing its button, as fast as possible; it stays up for 1,500 ms. The shapes come first on their own, 12 with no sound, each 1 to 2.5 s after the last.
- Sentences
- 59 everyday sentences per language of 6 to 10 words, our own, said by the speech kit's synthetic voice, each with three key words: a person, a thing and a place. 56 are test sentences and 3 are practice sentences.
- Babble
- Eight streams of the kit's other sentences, each played at its own rate and summed, so no key word of a sentence is ever in it. It starts half a second before the sentence and ends half a second after it.
- Levels
- Four, each one part of the session, in an order drawn from the session's seed: the sentence alone, and the sentence over babble 8 dB under it, 2 dB under it and 4 dB over it. The sentence plays at -19 dB re the volume-setting tone at every level, so only the babble changes, at -27, -21 and -15 dB.
- Trial
- Press Play to hear a sentence. A shape appears 45 to 80 percent of the way through it, at a time drawn from the seed, and is answered with a key. When the sound has ended nine words appear: the three key words and six others from other sentences, two of each kind. Choose three and press Done, with no time limit.
- Trials
- 12 sentences at each level (standard, 8 to 14): 48 sentences, each used once, with no feedback.
- Practice
- 3 sentences, in quiet, over babble 8 dB under the sentence and over babble 4 dB over it, with feedback for 1,800 ms.
- Score
- The median time of the correct shape answers is found for each level and for the shapes alone, leaving out answers under 100 ms. The main score is the mean, over the three babble levels, of the median time less the median time in quiet, in ms and as a percentage of the time in quiet. Also reported: each median, the cost of doing both tasks in quiet over the shapes alone, the share of key words right at each level and in all (chance is a third), the shape accuracy and misses, and anticipations.
Listening to a sentence and answering a shape are done at the same time, and the sentence is the main task. The shapes are one way to measure listening effort among several: most studies of listening effort use pupil size, brain activity or ratings, and the measures do not always agree. The babble is eight streams of one synthetic voice, a stand-in for several talkers, not recorded talkers. The sentences and the babble play through the participant's own headphones at a volume they chose, and the test needs normal hearing and headphones, wired if possible: wireless headphones can add delay and change the sound. It is a research measure, not a hearing test: it gives no hearing level and no pass or fail, says nothing about hearing aids or any person's hearing, and no percentile is given.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: the shapes alone (listening_effort_baseline) and the shapes answered while a sentence played (listening_effort_test) from the selected run, each a choice of a circle or a square (chance 1 in 2), with 1,500 ms to answer. Below chance invalidates and not above chance is flagged; the same shape on 90% of answers invalidates when accuracy is not above chance; more than max(3, 25%) shapes not answered invalidate and any is flagged; answers under 100 ms count as anticipations, and more than max(2, 20%) invalidate. A headphone check that stopped the test invalidates; a failed check with the test run anyway, fewer than 60% of the key words right in quiet, 40% or fewer at the loudest babble, no cost to find and a run short of its planned sentences are flagged for review.
Why these rules
Accuracy significantly below chance (one-sided exact test, p < .05) cannot come from guessing; it means the response mapping was reversed or the task was not followed (Slick, Sherman and Iverson 1999 treat below-chance forced-choice performance as definite response bias). Chance-level accuracy can be a real result in impaired participants, so it is only flagged. One option on at least 90% of answered trials, when the correct options are balanced and accuracy is not above chance, means the answers did not depend on the items (the task counterpart of long-string responding in surveys; Curran 2016). 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. The shapes are the second task of a dual-task measure of listening effort (Gagne, Besser and Lemke 2017): a person who does not do it, or answers before it appears, gives no measure of how much the listening took. Woods, Siegel, Traer and McDermott (2017) let only listeners who passed the headphone check go on, because a listener on loudspeakers hears the sounds differently; a session that ran after a failed check keeps its scores but not its percentile. The key words are the first task: the shapes' times are read as effort only while the sentences are heard, and Wu and colleagues (2016) found the times fall again when most of the speech is not understood, so a quiet condition that was not heard, or a loudest condition that was near chance, is flagged.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Sentences per level | 12 | 8 to 14 |
| When the headphone check fails | Stop the test | Stop the test, Continue the test |
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. Needs headphones. 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 Listening Effort Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Downs, 1982; Sarampalis, Kalluri, Edwards and Hafter, 2009; Fraser, Gagné, Alepins and Dubois, 2010; Gagné, Besser and Lemke, 2017), with the standard settings (sentences per level: 12; when the headphone check fails: Stop the test). The main score was listening effort cost (mean over the three babble levels of the median time to answer a correct shape at that level less the median time in quiet).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Downs, D. W. (1982). Effects of hearing aid use on speech discrimination and listening effort. Journal of Speech and Hearing Disorders, 47(2), 189-193. https://doi.org/10.1044/jshd.4702.189The time to answer a probe while a speech task is done, as a measure of listening effort.
- Sarampalis, A., Kalluri, S., Edwards, B., & Hafter, E. (2009). Objective measures of listening effort: Effects of background noise and noise reduction. Journal of Speech, Language, and Hearing Research, 52(5), 1230-1240. https://doi.org/10.1044/1092-4388(2009/08-0111)A visual reaction time task and a word memory task as second tasks, at several signal-to-noise ratios.
- Fraser, S., Gagné, J.-P., Alepins, M., & Dubois, P. (2010). Evaluating the effort expended to understand speech in noise using a dual-task paradigm: The effects of providing visual speech cues. Journal of Speech, Language, and Hearing Research, 53(1), 18-33. https://doi.org/10.1044/1092-4388(2009/08-0140)The cost of the second task as a proportion of the second task alone.
- Picou, E. M., & Ricketts, T. A. (2014). The effect of changing the secondary task in dual-task paradigms for measuring listening effort. Ear and Hearing, 35(6), 611-622. https://doi.org/10.1097/aud.0000000000000055A simple visual probe answered by a key press as the second task, one of three compared.
- Wu, Y.-H., Stangl, E., Zhang, X., Perkins, J., & Eilers, E. (2016). Psychometric functions of dual-task paradigms for measuring listening effort. Ear and Hearing, 37(6), 660-670. https://doi.org/10.1097/aud.0000000000000335A shape-like second task answered by a key at a random time in the second half of the sentence, and the finding that reaction times fall again when most of the speech is not understood.
- Gagné, J.-P., Besser, J., & Lemke, U. (2017). Behavioral assessment of listening effort using a dual-task paradigm. Trends in Hearing, 21, 2331216516687287. https://doi.org/10.1177/2331216516687287A review of 29 studies: the first task given priority, the second task alone as a baseline, and the proportional cost.
- Kuchinsky, S. E., Gallun, F. J., & Lee, A. K. C. (2024). Note on the dual-task paradigm and its use to measure listening effort. Trends in Hearing, 28, 23312165241292215. https://doi.org/10.1177/23312165241292215Control conditions, and how the two tasks share attention.
- McGarrigle, R., Munro, K. J., Dawes, P., Stewart, A. J., Moore, D. R., Barry, J. G., & Amitay, S. (2014). Listening effort and fatigue: What exactly are we measuring? A British Society of Audiology Cognition in Hearing Special Interest Group 'white paper'. International Journal of Audiology, 53(7), 433-445. https://doi.org/10.3109/14992027.2014.890296Self-report, behavioral and physiological measures of listening effort, of which this is a behavioral one.
- Simpson, S. A., & Cooke, M. (2005). Consonant identification in N-talker babble is a nonmonotonic function of N. The Journal of the Acoustical Society of America, 118(5), 2775-2778. https://doi.org/10.1121/1.2062650Babble of 8 talkers sits inside the range, 6 to 128 talkers, where consonant identification was lowest.
- Rosen, S., Souza, P., Ekelund, C., & Majeed, A. A. (2013). Listening to speech in a background of other talkers: Effects of talker number and noise vocoding. The Journal of the Acoustical Society of America, 133(4), 2431-2443. https://doi.org/10.1121/1.4794379Performance changed little as the number of talkers in the babble grew past 2 to 4.
- Woods, K. J. P., Siegel, M. H., Traer, J., & McDermott, J. H. (2017). Headphone screening to facilitate web-based auditory experiments. Attention, Perception, & Psychophysics, 79(7), 2064-2072. https://doi.org/10.3758/s13414-017-1361-2The headphone check before the test.
Source papers
- Downs (1982): probe reaction time during speech discrimination. Sarampalis, Kalluri, Edwards and Hafter (2009): a visual reaction time task and a word memory task at several signal-to-noise ratios. Fraser, Gagné, Alepins and Dubois (2010): a proportional dual-task cost. Wu, Stangl, Zhang, Perkins and Eilers (2016): reaction times across signal-to-noise ratios. Gagné, Besser and Lemke (2017) and Kuchinsky, Gallun and Lee (2024): reviews of the design.Paradigm source, Adults
Questions
How long does the Listening Effort Task take?
About 12 to 15 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. Needs headphones. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Listening Effort Cost, Listening Effort Cost, Percent, Shape Time Alone, Shape Time in Quiet, Shape Time, Babble 8 dB Under, Shape Time, Babble 2 dB Under, Shape Time, Babble 4 dB Over, Two-Task Cost in Quiet, Key Words Right, Key Words Right in Quiet, Key Words Right, Babble 8 dB Under, Key Words Right, Babble 2 dB Under, Key Words Right, Babble 4 dB Over, Shapes Right, Shapes Not Answered, Anticipations, Completed Trials, Planned Trials, Headphone Check Passed, Headphone Check Correct and Reported Output Delay. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Listening Effort Task used in research?
735 papers from 2021 to 2025 name it in their title or abstract, up 71% compared with 2016 to 2020. That ranks it 37 of 711 test families in the ConductCognition test census (use grade B). Source: ConductCognition test census 0.58.0, counted 16 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Listening Effort Task have?
In the published research, 34 papers report its test-retest reliability and 79 registered clinical studies use it as an outcome (evidence grade A in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Sentences per level (8 to 14) and when the headphone check fails (Stop the test, Continue the test). 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.
Related tests