The false belief test is a psychological task designed to measure whether a person, usually a young child, can understand that someone else can hold a belief that doesn’t match reality. It’s widely considered the gold standard for assessing “theory of mind,” the ability to recognize that other people have their own thoughts, beliefs, and intentions that may differ from your own and from what’s actually true. Most children begin passing the test around age 4, marking a major cognitive milestone.
How the Classic Test Works
The original version, created in the early 1980s, is often called the “Maxi task” or the “Sally-Anne task.” The setup is simple: a child watches a short story play out, usually with puppets or dolls. In a typical version, a character named Sally places her marble in a basket, then leaves the room. While she’s gone, another character moves the marble to a different location, say a box. Sally comes back, and the child is asked: “Where will Sally look for her marble?”
The correct answer is the basket, because that’s where Sally believes the marble still is. She didn’t see it get moved. But to get this right, the child has to separate what they know (the marble is in the box) from what Sally knows (she left it in the basket). Children who haven’t yet developed this ability will point to the box, answering based on reality rather than on Sally’s outdated belief.
The Unexpected Contents Version
A second popular format is the “Smarties task,” named after the candy brand. A researcher shows a child a familiar container, like an egg carton, and asks what they think is inside. The child naturally says eggs. The researcher opens it to reveal something unexpected, like pencils, then closes it again. The child is then asked two questions: “Before I opened it, what did you think was inside?” and “If your friend came in right now and saw this closed container, what would they think is inside?”
Children who pass will say their friend would think it contains eggs, recognizing that someone who hasn’t peeked inside would be fooled by the container just like they were. Children who fail tend to say their friend would know about the pencils, as if everyone automatically shares the same knowledge they now have.
Why Age 4 Is a Turning Point
Most two- and three-year-olds fail false belief tasks consistently. They reliably point to where the object actually is rather than where the character thinks it is. Around age 4, something shifts. Children begin answering correctly, demonstrating they can hold two competing representations in mind: what’s true and what someone else mistakenly believes to be true.
This shift correlates with improvements in other perspective-taking tasks, suggesting it reflects a broader cognitive reorganization rather than just learning one specific trick. Passing the false belief test requires working memory (holding the story details in mind), impulse control (resisting the urge to answer based on reality), and the ability to understand verbal instructions, which is part of why younger children struggle with it.
Cross-cultural research shows some variability in timing. When children are tested using traditional verbal tasks, the age of passing ranges from 4 to 7 years across different societies. But newer nonverbal methods tell a more nuanced story.
What Eyes Reveal Before Words Can
One of the most interesting developments in this field is the gap between what children say and where they look. In eye-tracking versions of the task, researchers don’t ask children to answer a question. Instead, they watch where the child’s gaze goes when the character returns to search for the object.
Almost all 3-year-olds look toward the location where the character believes the object to be, even though only about 30% of them give the correct verbal answer when asked directly. By age 4, the verbal answers catch up. This suggests that some grasp of false belief may be present earlier than the classic test reveals, but children lack the cognitive resources to express it in words or override their impulse to report reality.
Studies with children in traditional, non-Western societies found that even 1- to 4-year-olds showed sensitivity to others’ false beliefs on these nonverbal tasks, performing comparably to Western children. This points toward false belief understanding being a universal feature of early human development rather than something that depends heavily on culture or language.
Second-Order False Beliefs
Once children master the basic version, researchers can test for a more complex layer of reasoning. Second-order false belief tasks ask children to think about what one person believes another person thinks. For example: Murat sees Ayla move a chocolate from a drawer to a toy box, but Ayla doesn’t notice that Murat watched her do it. The question becomes, “Where does Ayla think Murat will look for the chocolate?”
To answer correctly (the drawer), a child must reason recursively: Ayla believes that Murat believes the chocolate is still in the drawer, because she doesn’t know he saw her move it. Children typically begin passing these tasks between ages 5 and 7, one to three years after they master first-order false beliefs.
What Happens in the Brain
In adults, false belief reasoning activates a specific network of brain areas. The right temporoparietal junction, a region near the back of the brain where the temporal and parietal lobes meet, is consistently active during belief reasoning of all kinds. A region in the middle of the frontal lobe plays an additional role specifically during false belief tasks, helping the brain process someone’s mental state as separate from observable reality. This “decoupling” process is what allows you to hold a belief in mind that you know to be wrong, purely because someone else holds it.
The Connection to Autism
False belief tasks have become a routine part of clinical evaluations for autism spectrum disorder. Difficulty understanding other people’s perspectives is one of the core features of autism, and false belief performance often reflects this. In one study of 68 children with autism (average age 7.5 years), the group did not perform above chance on a computer-based false belief task, while age-matched typically developing children performed significantly better.
That said, failing the test doesn’t automatically mean a child’s theory of mind is impaired. The classic verbal version demands memory, language comprehension, and impulse control on top of the mental state reasoning it’s designed to measure. A child might fail because of one of those other demands rather than because they can’t represent someone else’s belief. This is why researchers developed nonverbal, implicit versions of the task and why clinicians consider false belief results alongside other assessments rather than in isolation.
Can Animals Pass the Test?
Researchers have tried various versions of the false belief test with great apes and monkeys, with consistently negative results. Chimpanzees, orangutans, and rhesus macaques all fail when tested on whether they can account for another’s false belief about an object’s location. Rhesus macaques can distinguish between a person who knows where food is hidden and one who doesn’t, showing they understand the difference between knowledge and ignorance. But they don’t take the next step of predicting how someone will act based on a mistaken belief. This gap appears to be a meaningful cognitive boundary: even 15-month-old human infants show sensitivity to false beliefs on eye-tracking tasks, while no non-human primate tested so far has demonstrated the same ability.