Social Cognition
Sender-Receiver Game
Also called: deception game, cheap talk game, sender-receiver deception game, white lies game, Gneezy deception game
Measure how the consequences of a false message change whether people send one: participants report a die roll to a simulated receiver in eight rounds whose points favor different players, scored by the share of false messages.
5-7 minUntimed
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Preview any test free after you sign up.
- Papers, 2021 to 2025
- 68
- Evidence grade
- D
- Registered clinical studies
- 0
What it measures
Measures how often people send false messages when the message changes what they and another player get. In 8 rounds the participant sees a die and sends a simulated receiver a message about it; the points of each round reward a false message for the participant, the receiver, both or neither.
Constructs
- Lying aversion
- Weighing a lie's consequences for others
- Expectations about whether others will trust a message
- Moral decision making
Research uses
- Behavioral economics research on lying and white lies
- Moral psychology research
- Studies of social preferences and incentives
- Group comparison within a study
How the paradigm works
The Sender-Receiver Game measures how the consequences of a false message change whether people send one (Gneezy, 2005). A sender who alone knows what two options pay sends a message, and a receiver then chooses. Gneezy found that fewer senders lied when the lie cost the receiver more, and more lied when it paid the sender more, so people weigh the harm to others as well as their own gain. This version uses the die form of Erat and Gneezy (2012): in each round a die shows a number from 1 to 6 that only the participant sees, and the participant sends one of six messages, from "The die shows 1" to "The die shows 6". The receiver sees only the message and chooses a number. Option A sets the points when it chooses the number rolled, and option B otherwise, so a false message can help the participant, the receiver, both or neither.
More about the paradigm
Option A always gives each player 50 points. Option B raises or lowers each player's points by 5 or 30, in eight payoff pairs played once each in an order drawn from the session's seed: in three, a followed false message helps the participant at the receiver's cost, as in Gneezy (2005); in three, it helps the receiver alone or both players, the white lies of Erat and Gneezy (2012); in one, it helps only the participant; in one, it costs both. After each message the participant says whether they expect the receiver to choose the number in the message (after Sutter, 2009). The receiver is a computer program that follows set rules, and the standard instructions say so: it chooses the number in the message with probability 0.8, near the 78% of Gneezy's receivers who followed, and otherwise one of the other five numbers. No choice of the receiver is shown until the end, when one round drawn from the seed sets the points, as in Erat and Gneezy's second study; with feedback after every message, lying rose with experience (Gneezy, Rockenbach and Serra-Garcia, 2013). One check question with feedback comes first, and there is no time limit. No screen names honesty or lying. The study decides what the points are worth, and the payment line says whether they are exchanged for money.
The main score is the share of false messages over the seven rounds where option B pays at least one player more than option A. The rates for the selfish and the white-lie rounds, a 0 or 1 for each round's message and how often the participant expected the receiver to follow are also reported, with the true messages sent where option B paid the participant more while expecting the receiver to choose another number, which Sutter (2009) counted as a way to mislead with the truth. Each payoff pair is one decision, so a pair's rate is read across a group, as the sources report it: in Gneezy's study, with real receivers and one decision per sender, 17% to 52% of senders lied, depending on the payoffs. A lie to a disclosed computer program may weigh less than a lie to a person, so compare people who saw the same instructions; studies usually explain the purpose afterwards. The scores describe messages in a game played for points, never a person's honesty, and they do not diagnose.
Scores
- False Message Rate%
- Share of messages that named a number other than the die's, over the 7 rounds where option B pays at least one player more than option A.
- Read across a group; no better direction
- False Rate, Selfish Rounds%
- Share of false messages over the 3 rounds where a followed false message helps the participant at the receiver's cost (Gneezy, 2005).
- Informational
- False Rate, White Lie Rounds%
- Share of false messages over the 3 rounds where a followed false message helps the receiver alone or both players (Erat and Gneezy, 2012).
- Informational
- False Message, Small Gain and Small Cost0/1
- 1 when the message was false in the round where option B gives the participant 5 more points and the receiver 5 fewer; 0 when it was true.
- Read across a group
- False Message, Small Gain and Large Cost0/1
- 1 when the message was false in the round where option B gives the participant 5 more points and the receiver 30 fewer; 0 when it was true.
- Read across a group
- False Message, Large Gain and Large Cost0/1
- 1 when the message was false in the round where option B gives the participant 30 more points and the receiver 30 fewer; 0 when it was true.
- Read across a group
- False Message, Gain With No Cost0/1
- 1 when the message was false in the round where option B gives the participant 30 more points and the receiver the same; 0 when it was true.
- Read across a group
- False Message, Both Gain, Small for the Participant0/1
- 1 when the message was false in the round where option B gives the participant 5 more points and the receiver 30 more; 0 when it was true.
- Read across a group
- False Message, Both Gain0/1
- 1 when the message was false in the round where option B gives each player 30 more points; 0 when it was true.
- Read across a group
- False Message, Small Cost for the Receiver's Gain0/1
- 1 when the message was false in the round where option B gives the participant 5 fewer points and the receiver 30 more; 0 when it was true.
- Read across a group
- False Message, Both Lose0/1
- 1 when the message was false in the round where option B gives each player 5 fewer points; 0 when it was true.
- 0 expected; 1 is flagged for review
- Expected Follow Rate%
- Share of rounds in which the participant expected the receiver to choose the number in the message.
- Informational
- True Messages Expected to Be Ignoredcount
- Rounds where option B paid the participant more, the message was true and the participant expected the receiver to choose another number (Sutter, 2009).
- Informational
- Points Wonpoints
- The participant's points in the round drawn to count.
- Informational
- Check Correctflag
- 1 when the check question on which option sets the points was answered correctly, 0 when not.
- 1 supports interpretation
- Median Message Timems
- Median time from a round's screen to the message, over messages at or after 100 ms.
- Informational
- Messages Under 100 mscount
- Messages sent less than 100 ms after the round's screen appeared.
- 0 expected
- Roundscount
- Rounds played.
- Informational
Use and evidence
68 papers from 2021 to 2025 name this task in their title or abstract, down 4% compared with 2016 to 2020. That ranks it 230 of 395 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 16 |
| 2024 | 15 |
| 2023 | 16 |
| 2022 | 11 |
| 2021 | 10 |
| 2020 | 14 |
| 2019 | 16 |
| 2018 | 20 |
| 2017 | 11 |
| 2016 | 10 |
| 2015 | 10 |
| 2014 | 19 |
| 2013 | 7 |
| 2012 | 7 |
| 2011 | 7 |
| 2010 | 3 |
| 2009 | 4 |
| 2008 | 5 |
| 2007 | 3 |
| 2006 | 3 |
| 2005 | 3 |
| 2004 | 3 |
| 2003 | 1 |
| 2002 | 2 |
| 2001 | 1 |
| 2000 | 0 |
Use grade D
30 to 99 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 0
- Papers on reliability or validity
- 6
- Registered clinical studies using it as an outcome
- 0
Note on the counts: includes game-theory papers that run no task.
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 0.14.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Each round
- A die shows a number from 1 to 6 that only the participant sees. The participant sends one of six messages, from "The die shows 1" to "The die shows 6", then says whether they expect the receiver to choose the number in the message.
- Receiver
- A computer program that follows set rules. The standard instructions say so; a neutral wording leaves that line out. It chooses the number in the message with probability 0.8, otherwise one of the other five numbers, and the participant never sees its choice before the end.
- Points
- Option A, paid when the receiver chooses the number rolled, gives each player 50 points in every round. Option B, paid otherwise, raises or lowers each player's points by 5 or 30: 8 rounds, one for each payoff pair, in an order drawn from the session's seed.
- Payoff pairs
- A false message, if followed, helps the participant at the receiver's cost in 3 rounds, helps the receiver alone or both players in 3, helps only the participant in 1 and costs both in 1.
- Counted round
- One round, drawn from the session's seed, sets the points shown at the end. Points are hypothetical by default; a setting changes the instruction line to say the study pays.
- Check
- One question with feedback before the first round: the die shows 2 and the receiver chooses 5.
- Scores
- The share of false messages over the 7 rounds where option B pays at least one player more, the rates for the selfish and the white-lie rounds, each round's message, and how often the participant expected the receiver to choose the number sent.
Gneezy (2005) gave each sender one payoff pair and a real receiver. Here each participant sends eight messages, one per payoff pair, in the die form of Erat and Gneezy (2012), so one person's score for a pair is 0 or 1 and a pair's rate is read across a group. No screen names honesty or lying. The receiver is a computer program and the standard instructions say so; a lie to a program may weigh less than a lie to a person, so compare people who saw the same instructions. Studies usually explain the purpose afterwards. The scores describe messages in a game played for points, never a person's honesty, 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: the scored run's messages, one per round. No message is right or wrong, so accuracy is not judged against chance, and sending the same message throughout is accepted. A run short of its planned rounds, a wrong check answer and a false message in the round that costs both players are flagged for review. Messages under 200 ms on at least half the rounds invalidate.
Why these rules
There is no correct option to compare with chance, so accuracy is not judged against chance. A message is a choice in a game played for points, so the rule judges only how the messages were given, never what they say about the participant. A false message in the round where it helps neither player often means the rule was misread, so it is flagged and never excludes. 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 |
|---|---|---|
| Receiver wording | A computer program (standard) | A computer program (standard), Neutral: the other player |
| Payment line | Hypothetical points (standard) | Hypothetical points (standard), The study pays for points |
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 Sender-Receiver Game in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Gneezy U, 2005; Erat S, Gneezy U, 2012), with the standard settings (receiver wording: A computer program; payment line: Hypothetical points). The main score was false message rate (share of messages that named a number other than the die's, over the 7 rounds where option B pays at least one player more than option A).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Gneezy, U. (2005). Deception: The role of consequences. American Economic Review, 95(1), 384-394. https://doi.org/10.1257/0002828053828662The sender-receiver game: fewer false messages when they cost the receiver more, and more when they pay the sender more.
- Erat, S., & Gneezy, U. (2012). White lies. Management Science, 58(4), 723-733. https://doi.org/10.1287/mnsc.1110.1449The die form, white lies that help the receiver, and one payoff pair drawn to count.
- Sutter, M. (2009). Deception through telling the truth?! Experimental evidence from individuals and teams. Economic Journal, 119(534), 47-60. https://doi.org/10.1111/j.1468-0297.2008.02205.xAsking which choice the sender expects, and true messages sent expecting them to be ignored.
- Gneezy, U., Rockenbach, B., & Serra-Garcia, M. (2013). Measuring lying aversion. Journal of Economic Behavior & Organization, 93, 293-300. https://doi.org/10.1016/j.jebo.2013.03.025Lying measured for each person over repeated messages; with feedback after each, lying rose.
Source papers
- Gneezy (2005): with real receivers and one decision per sender, 36% of senders lied when the lie gained them $1 and cost the receiver $1, 17% when it cost the receiver $10, and 52% when it gained them $10 and cost the receiver $10 (75 senders per pair). Erat and Gneezy (2012): 33% of senders told a white lie that cost them $1 and gained the receiver $10, and 65% one that gained both $10.Paradigm source, Adults
Questions
How long does the Sender-Receiver Game take?
About 5 to 7 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?
False Message Rate, False Rate, Selfish Rounds, False Rate, White Lie Rounds, False Message, Small Gain and Small Cost, False Message, Small Gain and Large Cost, False Message, Large Gain and Large Cost, False Message, Gain With No Cost, False Message, Both Gain, Small for the Participant, False Message, Both Gain, False Message, Small Cost for the Receiver's Gain, False Message, Both Lose, Expected Follow Rate, True Messages Expected to Be Ignored, Points Won, Check Correct, Median Message Time, Messages Under 100 ms and Rounds. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Sender-Receiver Game used in research?
68 papers from 2021 to 2025 name it in their title or abstract, down 4% compared with 2016 to 2020. That ranks it 230 of 395 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.14.0, counted 2 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Sender-Receiver Game have?
In the published research, 0 papers report its test-retest reliability and 0 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?
Receiver wording (A computer program (standard), Neutral: the other player) and payment line (Hypothetical points (standard), The study pays for points). 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. 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