Social Cognition
Social Evaluation Learning Task
Also called: SELT, social evaluative learning task, social evaluation learning paradigm
Study how people learn what others think of them: participants learn from feedback whether made-up characters like or dislike them or another person, scored as the self-other difference in errors to criterion.
8-12 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
- 13
- Evidence grade
- D
- Registered clinical studies
- 2
What it measures
Measures learning of positive and negative evaluations about oneself and about another person. Participants meet made-up characters and learn from feedback whether each one likes or dislikes them, or likes or dislikes a person called Sam, by choosing the word the character would use. The main score compares how many fewer errors "likes" takes than "dislikes" to learn about oneself with the same difference about Sam, counting errors before 8 choices in a row follow the character's rule.
Constructs
- Social evaluation learning
- Learning from feedback
- Self-referential processing
- Social cognition
Research uses
- Research on social anxiety and depression
- Antidepressant and mood research
- Comparing learning about oneself with learning about another person
How the paradigm works
The social evaluation learning task measures how people learn what others think of them and of someone else (Button, Kounali, Stapinski and colleagues, 2015). The participant meets made-up characters one at a time. Each character describes either the participant or another person, Sam. In each trial the character starts a sentence, "I think you are…" or "I think Sam is…", and the participant chooses which of a positive and a negative word the character would use, then sees Correct or Incorrect. A character who likes the person described calls the positive word correct, and one who dislikes the person calls the negative word correct, on 8 of every 10 trials; on the other 2 the feedback is switched. After each character the participant rates how much it likes the person described, from 0 to 100. In the source, adults with more fear of negative evaluation chose fewer positive words, mostly about themselves rather than about the other person. Across later studies, people learned that a character likes the person described with fewer errors than that it dislikes them, and this advantage about oneself was smaller with more symptoms of depression or social anxiety (Hoffmann, Hobbs, Moutoussis and Button, 2024).
More about the paradigm
The standard session has four characters: one who likes and one who dislikes the participant, and the same two for Sam. Whether the participant or Sam comes first, and the order of the two characters within each, come from the session's seed. Each character shows the same 32 word pairs (20 to 32 in the settings), each once, in its own order, with the positive word on the left in half of them. The words were written for this task in each of the eight languages: mild words for how a person acts with others, with nothing about the body, looks or groups of people. The words stay as buttons until a choice by click, tap, or the F and J keys, with no time limit, and Correct or Incorrect shows for 1 second. A setting adds the 2015 study's two neutral characters, whose feedback calls the positive word correct on half the trials. Four practice trials come first. About 8 to 12 minutes.
The characters and their feedback are simulated. The instructions say that a computer program chooses their words by set rules, and the end screen says again that the characters were made up and that nothing in the task was about the participant or Sam. Studies usually explain the task again afterwards.
The main score compares learning about oneself with learning about Sam. For each person, the errors made before 8 choices in a row follow the rule of the character who dislikes them, minus the same count for the character who likes them, gives the like advantage; the main score is the participant's like advantage minus Sam's. Errors for each character, the share of positive words chosen, the ratings, and how choices changed after Correct and after Incorrect come next. The scores describe learning in this task as research measures. They do not say what a person expects from others, and they do not diagnose.
Scores
- Self-Other Like Advantagescore
- The like advantage about the participant minus the like advantage about Sam.
- Positive: "likes" took relatively fewer errors about oneself than about Sam. No better direction
- Like Advantage, Selfscore
- Errors to criterion with the character who dislikes the participant minus errors with the character who likes the participant.
- Positive: "likes" took fewer errors to learn
- Like Advantage, Samscore
- Errors to criterion with the character who dislikes Sam minus errors with the character who likes Sam.
- Positive: "likes" took fewer errors to learn
- Errors to Criterion, Likes Youcount
- Negative words chosen with the character who likes the participant before 8 positive choices in a row; every error in the block when the run never came.
- Lower is faster learning
- Errors to Criterion, Dislikes Youcount
- Positive words chosen with the character who dislikes the participant before 8 negative choices in a row; every error in the block when the run never came.
- Lower is faster learning
- Errors to Criterion, Likes Samcount
- Negative words chosen with the character who likes Sam before 8 positive choices in a row; every error in the block when the run never came.
- Lower is faster learning
- Errors to Criterion, Dislikes Samcount
- Positive words chosen with the character who dislikes Sam before 8 negative choices in a row; every error in the block when the run never came.
- Lower is faster learning
- Characters Learnedcount
- Characters who like or dislike, whose choices reached 8 in a row that follow their rule.
- 4 in the standard session means every rule was learned
- Positive Words, Self%
- Share of positive words chosen with the characters who describe the participant.
- Informational
- Positive Words, Sam%
- Share of positive words chosen with the characters who describe Sam.
- Informational
- Positive Words, Likes You%
- Share of positive words chosen with the character who likes the participant.
- Informational; the feedback calls positive words correct on 80% of trials
- Positive Words, Dislikes You%
- Share of positive words chosen with the character who dislikes the participant.
- Informational; the feedback calls positive words correct on 20% of trials
- Positive Words, Likes Sam%
- Share of positive words chosen with the character who likes Sam.
- Informational; the feedback calls positive words correct on 80% of trials
- Positive Words, Dislikes Sam%
- Share of positive words chosen with the character who dislikes Sam.
- Informational; the feedback calls positive words correct on 20% of trials
- Positive Words, Neutral About You%
- Share of positive words chosen with the neutral character who describes the participant, when the setting adds neutral characters.
- Informational; the feedback calls positive words correct on 50% of trials
- Positive Words, Neutral About Sam%
- Share of positive words chosen with the neutral character who describes Sam, when the setting adds neutral characters.
- Informational; the feedback calls positive words correct on 50% of trials
- Rating, Likes Youscore
- How much the participant thought the character who likes them likes them, from 0 (dislikes a lot) to 100 (likes a lot).
- Informational
- Rating, Dislikes Youscore
- How much the participant thought the character who dislikes them likes them, from 0 to 100.
- Informational
- Rating, Likes Samscore
- How much the participant thought the character who likes Sam likes Sam, from 0 to 100.
- Informational
- Rating, Dislikes Samscore
- How much the participant thought the character who dislikes Sam likes Sam, from 0 to 100.
- Informational
- Rating, Neutral About Youscore
- The rating of the neutral character who describes the participant, when the setting adds neutral characters.
- Informational
- Rating, Neutral About Samscore
- The rating of the neutral character who describes Sam, when the setting adds neutral characters.
- Informational
- Repeat After Correct%
- Within a character, the share of choices that kept the word's valence after Correct.
- Informational
- Shift After Incorrect%
- Within a character, the share of choices that changed the word's valence after Incorrect.
- Informational
- Median Choice Timems
- Median time from the words appearing to the choice.
- Informational
- Anticipationscount
- Presses within 100 ms of the words appearing. They are not taken as choices.
- Quality check
- Choices Madecount
- Choices in the scored run.
- Fewer than planned is flagged for review
- Planned Trialscount
- Trials in the session: the trials per character times the characters.
- Informational
Use and evidence
13 papers from 2021 to 2025 name this task in their title or abstract. That ranks it 436 of 495 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 3 |
| 2024 | 4 |
| 2023 | 1 |
| 2022 | 2 |
| 2021 | 3 |
| 2020 | 1 |
| 2019 | 2 |
| 2018 | 1 |
| 2017 | 0 |
| 2016 | 1 |
| 2015 | 1 |
| 2014 | 1 |
| 2013 | 0 |
| 2012 | 0 |
| 2011 | 0 |
| 2010 | 0 |
| 2009 | 0 |
| 2008 | 0 |
| 2007 | 0 |
| 2006 | 0 |
| 2005 | 0 |
| 2004 | 0 |
| 2003 | 0 |
| 2002 | 0 |
| 2001 | 0 |
| 2000 | 0 |
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
- 0
- 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 2 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Characters
- Faces drawn by code for this task, with names, met one at a time. Four in the standard session: one who likes the participant, one who dislikes them, and the same two for Sam, a person the participant does not know. The participant or Sam comes first by the session's seed, and the two characters within each in a seeded order.
- Rules
- A character who likes the person described calls the positive word correct, and one who dislikes them calls the negative word correct, on 8 of every 10 trials; on the other 2, in seeded places, the feedback is switched.
- Words
- 32 pairs of mild words for how a person acts with others, one positive and one negative, written for this task in each of the eight languages. Each character shows the same pairs (standard 32; 20 to 32), each once, in its own order, with the positive word on the left in half of them.
- Trial
- The face and the start of a sentence, "I think you are…" or "I think Sam is…", over two word buttons, until a choice by click, tap, or the F and J keys, with no time limit. Correct or Incorrect shows for 1,000 ms, then a 500 ms blank. Presses under 100 ms are not taken.
- Rating
- After each character, how much it likes the person described, from 0 (dislikes a lot) to 100 (likes a lot), on a bar with no starting mark.
- Neutral characters
- A setting adds the 2015 study's two neutral characters, one about the participant and one about Sam, whose feedback calls the positive word correct on 5 of every 10 trials.
- Practice
- 4 trials that ask for the animal in a pair of words.
The source drew 32 of 64 English word pairs for each persona. This test writes its own 32 pairs in eight languages and shows every character the same pairs, so its values are not directly comparable with published ones.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scored rows: social_evaluation_learning_trial from the scored run, a choice between a positive and a negative word with no time limit. Each character's rule calls one word correct most of the time, so choices that match the rule measure learning, and accuracy is not judged against chance: choosing mostly positive or mostly negative words is a result the test measures. One side on at least 90% of choices is flagged for review; the positive word is on each side equally often. Answers under 200 ms on at least half the choices invalidate. A run short of its planned choices is flagged for review and withholds the percentile.
Why these rules
Learning a character's rule is the measure, and the source found choices that departed from the feedback in some participants and not others (Button and colleagues, 2015), so neither accuracy nor a run of the same word means the task was not followed. The words, not their places, carry the choice, so one side on nearly every choice means the words were not read. 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.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Trials per character | 32 | 20 to 32 |
| Neutral characters | Off | On or off |
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
- 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 Social Evaluation Learning Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Button, Kounali, Stapinski et al., 2015; Button, Browning, Munafò and Lewis, 2012), with the standard settings (trials per character: 32; neutral characters: Off). The main score was self-other like advantage (the like advantage about the participant minus the like advantage about Sam).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Button, K. S., Kounali, D., Stapinski, L., Rapee, R. M., Lewis, G., & Munafò, M. R. (2015). Fear of negative evaluation biases social evaluation inference: Evidence from a probabilistic learning task. PLoS ONE, 10(4), e0119456. https://doi.org/10.1371/journal.pone.0119456The task: personas who like, dislike or are neutral toward the participant or another person, feedback switched on 2 of every 10 trials (5 for the neutral rule), 32 word pairs per persona, the rating after each, and the self-other contrast.
- Button, K. S., Browning, M., Munafò, M. R., & Lewis, G. (2012). Social inference and social anxiety: Evidence of a fear-congruent self-referential learning bias. Journal of Behavior Therapy and Experimental Psychiatry, 43(4), 1082-1087. https://doi.org/10.1016/j.jbtep.2012.05.004The first form, with the rules "I am liked" and "I am disliked".
- Button, K. S., Karwatowska, L., Kounali, D., Munafò, M. R., & Attwood, A. S. (2016). Acute anxiety and social inference: An experimental manipulation with 7.5% carbon dioxide inhalation. Journal of Psychopharmacology, 30(10), 1036-1046. https://doi.org/10.1177/0269881116653105The four-block form without neutral rules, and errors before 8 rule-matching choices in a row, with every error counted when the run never comes.
- Hobbs, C., Beck, M., Denham, F., Pettitt, L., Faraway, J., Munafò, M. R., Sui, J., Kessler, D., & Button, K. S. (2023). Relationship between change in social evaluation learning and mood in early antidepressant treatment: A prospective cohort study in primary care. Journal of Psychopharmacology, 37(3), 303-312. https://doi.org/10.1177/02698811221116928Errors to criterion and the difference between the like and dislike rules about the self, a friend and a stranger, during antidepressant treatment.
- Hoffmann, J. A., Hobbs, C., Moutoussis, M., & Button, K. S. (2024). Lack of optimistic bias during social evaluation learning reflects reduced positive self-beliefs in depression and social anxiety, but via distinct mechanisms. Scientific Reports, 14, 22471. https://doi.org/10.1038/s41598-024-72749-6Errors to criterion and the like-over-dislike difference in a mega-analysis of 450 participants and a preregistered online study of 807.
Source papers
- Button KS, Kounali D, Stapinski L, Rapee RM, Lewis G, Munafò MR (2015). Fear of negative evaluation biases social evaluation inference: Evidence from a probabilistic learning task. PLoS ONE, 10(4):e0119456.N 100, 18 to 49
- Hoffmann JA, Hobbs C, Moutoussis M, Button KS (2024). Lack of optimistic bias during social evaluation learning reflects reduced positive self-beliefs in depression and social anxiety, but via distinct mechanisms. Scientific Reports, 14:22471.N 807, 18 to 65
- No normative data for this task.Paradigm source, N/A
Questions
How long does the Social Evaluation Learning Task 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?
Self-Other Like Advantage, Like Advantage, Self, Like Advantage, Sam, Errors to Criterion, Likes You, Errors to Criterion, Dislikes You, Errors to Criterion, Likes Sam, Errors to Criterion, Dislikes Sam, Characters Learned, Positive Words, Self, Positive Words, Sam, Positive Words, Likes You, Positive Words, Dislikes You, Positive Words, Likes Sam, Positive Words, Dislikes Sam, Positive Words, Neutral About You, Positive Words, Neutral About Sam, Rating, Likes You, Rating, Dislikes You, Rating, Likes Sam, Rating, Dislikes Sam, Rating, Neutral About You, Rating, Neutral About Sam, Repeat After Correct, Shift After Incorrect, Median Choice Time, Anticipations, Choices Made and Planned Trials. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Social Evaluation Learning Task used in research?
13 papers from 2021 to 2025 name it in their title or abstract. That ranks it 436 of 495 test families in the ConductCognition test census (use grade E). Source: ConductCognition test census 0.22.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Social Evaluation Learning 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 settings can a researcher change?
Trials per character (20 to 32) and neutral characters (On or off). 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.