The Crawford-Sobel Model is a signaling-game framework in Game Theory where a sender with private information chooses a message and a receiver responds strategically. It shows when signals reveal type and when they stay only partly informative.
The Crawford-Sobel Model is a Game Theory model for a signaling game with private information, where one player, the sender, knows something the other player, the receiver, does not. The sender can send a message, and the receiver uses that message to choose an action.
What makes the model useful is that it does not assume messages are automatically truthful. The sender’s signal is only informative if it changes the receiver’s beliefs in a believable way. If a message can be sent by many different types with little cost, the receiver may not learn much from it.
In the Crawford-Sobel setting, the sender and receiver are both strategic. The sender wants the receiver to choose an outcome that favors the sender, but the receiver wants to make the best choice based on the sender’s actual type. Because of that tension, the equilibrium may involve partial revelation instead of full honesty or total deception.
A big idea in this model is that signals can separate types, but only sometimes. When signaling is credible, different types choose different messages and the receiver can distinguish among them. When signaling is weak or cheap, messages may bunch together, so the receiver learns only a rough category rather than the exact private information.
This is why Crawford-Sobel shows up in 8.2 Signaling Games and Information Revelation. It gives you a clean way to think about when communication works, when it gets distorted, and why a message can be strategically useful even when it is not perfectly honest.
A simple way to picture it is this: a job applicant, a salesperson, or any informed player may want to send a message that nudges the other side toward a favorable decision. The receiver is not just decoding words, they are inferring type. The whole model is about that back-and-forth between signaling and belief updating.
The Crawford-Sobel Model is one of the clearest ways to study how information moves in Game Theory when one side has an advantage. It turns a vague idea like “sending a signal” into a strategic problem: what should the informed player say, and how much should the uninformed player believe?
This matters because a lot of real game theory problems are not about perfect information. They are about hidden quality, hidden effort, hidden intentions, or hidden preferences. The model helps you explain why some messages separate players into types while others get ignored or only partially trusted.
It also gives you a sharper way to talk about equilibrium. A signaling game is not solved just by listing possible messages. You have to check whether the sender would really choose that signal and whether the receiver would really respond to it once beliefs are updated. That connection between beliefs, incentives, and outcomes is the heart of the model.
You can use it to compare believable signals with weak ones, or to explain why a sender might prefer a message that is informative but not fully revealing. That distinction shows up all over strategic communication, especially when both players are trying to maximize their own payoff instead of simply sharing facts.
Keep studying Game Theory Unit 8
Visual cheatsheet
view gallerySignaling
Crawford-Sobel is a signaling model, so it sits inside the bigger idea of signaling itself. The sender chooses a message to influence the receiver’s action, but the message only works if the receiver thinks it is informative. Crawford-Sobel shows that signaling is strategic, not just communicative.
Types
The model depends on hidden types, meaning different private states of the sender that affect payoffs. The receiver does not observe the type directly, so the message is really a clue about which type the sender may have. A lot of the analysis is about whether the message separates those types or blurs them together.
Bayesian Nash Equilibrium
Crawford-Sobel equilibria are found by combining best responses with beliefs about types. The receiver updates those beliefs and then chooses an action, while the sender chooses a message anticipating that response. If you are solving a signaling-game problem, equilibrium is where the whole belief-action pattern holds together.
cheap talk games
Crawford-Sobel is often discussed next to cheap talk games because both involve communication without automatic truthfulness. The difference is that Crawford-Sobel focuses on how messages can still carry some information even when they are not directly costly. That makes it a useful model for partial revelation.
A quiz or problem-set question on the Crawford-Sobel Model usually asks you to read a signaling situation and decide what the sender can credibly communicate. You may need to identify the sender’s type, explain why the receiver updates beliefs, and describe whether the equilibrium is separating, pooling, or partially revealing.
When you solve these problems, look for the incentive gap between types. If different types want different outcomes, the message may separate them. If the messages are too easy to imitate, the receiver will discount them, and you should explain why the signal does not fully reveal private information.
A short written response may ask you to compare this model with a costlier or more informative signaling setup, so you should be ready to say what the receiver knows, what the sender knows, and how beliefs shape the final action.
The Crawford-Sobel Model is a signaling-game framework where a sender with private information sends a message to a receiver.
Its main insight is that signals are only useful when the receiver finds them credible enough to update beliefs.
The model often produces partial revelation, not perfect honesty or complete misinformation.
Equilibrium depends on what each type of sender wants and how the receiver responds to the message.
It is a core tool for analyzing strategic communication in Game Theory.
It is a model of signaling with private information, where one player sends a message and the other player responds based on beliefs about the sender’s type. The point of the model is to show that communication can be strategic, so messages may reveal only part of the truth.
They are closely related, but Crawford-Sobel is usually used to show how strategic communication can still carry information even when messages are not automatically truthful. In cheap talk games, messages are costless and credibility can be even weaker, so the receiver may learn less or nothing at all.
A type is the sender’s private information, such as quality, preference, or hidden state. The receiver cannot observe it directly, so the sender’s message is a signal that may reveal, hide, or blur that type depending on incentives.
You identify who knows the private information, what messages are available, and how the receiver updates beliefs after each message. Then you check whether each sender type would actually choose that message, which tells you whether the equilibrium is separating, pooling, or partially revealing.