Preconceptions are the mental models, assumptions, and expectations you carry into every new experience, and they shape what you notice, what you believe, and how you interpret evidence far more than most people realize. Cognitive scientists describe them as structured knowledge frameworks, or schemas, that guide attention, fill in gaps, and generate predictions before you even become conscious of doing so. What makes preconceptions so powerful is that they are not simply wrong ideas waiting to be corrected. They often arise from real patterns in everyday life, which is exactly why they resist change even when better information comes along.
Where Preconceptions Come From
Your brain does not wait for formal instruction to start building models of how the world works. Research in developmental psychology shows that infants as young as five months already expect physical objects, including people, to be solid and unable to pass through each other.1PubMed. Five-month-old infants know humans are solid, like inanimate objects These early expectations about how objects behave form a kind of core knowledge. When the world violates those expectations, infants pay more attention and learn more readily, suggesting that prediction and surprise have been central to learning since before language develops.2PubMed Central. Violations of Core Knowledge Shape Early Learning
As children grow, these basic expectations about physics, biology, and social behavior become increasingly elaborate. Preschoolers, for instance, tend to assume that everything has a purpose. Research on teleological thinking found that young children attribute functions to all kinds of things, including clouds, tigers, and rocks, not just human-made objects like clocks.3PubMed. The scope of teleological thinking in preschool children Adults generally learn to restrict purpose-based explanations to artifacts and biological adaptations, but the underlying tendency never fully disappears. It lingers as a default way of reasoning, especially under time pressure or when dealing with unfamiliar topics.
Psychologists describe the mature versions of these mental models as schemas: organized knowledge structures that guide how you interpret new information, what inferences you draw, and where your attention lands. A schema is not a flat list of facts. It has internal structure, with slots that get filled in by context as you encounter a specific situation.4Psychology of Learning and Motivation. The Impact of a Schema on Comprehension and Memory When you walk into a restaurant, your “restaurant schema” tells you to expect a host, a menu, a bill. You do not reason through these from scratch every time. The same process operates in less obvious settings: when you meet someone, assess a news headline, or hear a scientific claim, schemas are already filtering what you take in.
Why Science Classrooms Struggle with Them
One of the clearest demonstrations of how stubborn preconceptions can be is in science education. Students arrive in physics, biology, and chemistry courses with intuitive ideas that feel correct because they are based on everyday sensory experience. A ball rolling across a table slows and stops, so it seems natural to believe that motion requires a continuous force. That idea is wrong in the Newtonian sense, but it matches what you see with your own eyes every day. Physics educators consistently find that these everyday intuitions resist change even after direct instruction.5arXiv. Addressing misconceptions in university physics: A review and experiences from quantum physics educators
Biology faces the same challenge. Research has linked many common biological misconceptions to a small set of deeply rooted intuitive patterns: teleological thinking (assuming things exist for a purpose), essentialist thinking (assuming species have a fixed inner nature), and anthropocentric thinking (projecting human qualities onto other organisms). These cognitive tendencies appear across both biology majors and non-majors, suggesting that years of science education do not simply erase them.6PubMed Central. Relations between intuitive biological thinking and biological misconceptions in biology majors and nonmajors A student who understands natural selection at exam time may still instinctively say that giraffes “grew” long necks “in order to” reach high leaves. The teleological framing is easy and feels right; the accurate mechanistic explanation requires actively overriding it.
Researchers who study conceptual change have debated whether these preconceptions function as coherent personal theories or as loose collections of ideas that get activated in different contexts. The two main perspectives, sometimes called “knowledge-as-theory” and “knowledge-as-elements,” disagree on how structured the wrong ideas are, but both agree on the practical problem: preconceptions do not vanish simply because someone presents the correct answer.7Eurasia Journal of Mathematics, Science and Technology Education. An Overview of Conceptual Change Theories
How Confirmation Bias Locks Them In
Once a preconception takes hold, a suite of cognitive tendencies works to protect it. The most well-known is confirmation bias, and recent modeling work has reframed it in an interesting way. Rather than treating confirmation bias as a flaw or a motivational quirk, one computational model shows it emerging naturally from the way the brain handles the overwhelming complexity of updating beliefs. In this model, an individual simultaneously updates their belief about a claim and their trust in the source of the information. Tracking every dependency between those beliefs would require unrealistic memory capacity, so the brain takes a shortcut by treating them as independent. That shortcut alone produces many recognizable patterns: preferring evidence that fits existing beliefs, dismissing evidence that does not, seeking out agreeable sources, and even becoming more polarized after seeing mixed evidence.8Cognition. Confirmation bias emerges from an approximation to Bayesian reasoning
Laboratory experiments confirm that the sampling of evidence itself is biased. When people make an initial judgment, they subsequently spend more time looking at information that supports their choice.9PubMed Central. Humans actively sample evidence to support prior beliefs And the bias is not limited to attention. Research on perception found that after making a choice, people process new evidence asymmetrically: information consistent with the original decision gets amplified, while inconsistent information gets dampened.10Current Biology. Choice-Induced Biases in Perceptual and Numerical Decision Making In other words, the act of committing to a belief changes how subsequent evidence literally registers. This creates a feedback loop: the preconception shapes what you see, and what you see reinforces the preconception.
The Brain as a Prediction Machine
Neuroscience provides a framework for understanding why preconceptions are so deeply woven into perception. A growing body of work treats the brain as fundamentally a prediction engine. Rather than passively receiving sensory input and then making sense of it, the brain constantly generates top-down predictions about what it expects to encounter. Incoming sensory signals then get compared against those predictions, and the brain focuses processing on the “prediction errors,” the mismatches between expectation and reality.11Frontiers. What is Bottom-Up and What is Top-Down in Predictive Coding? Under this view, what you experience as perception is not raw reality but a blend of incoming data and your brain’s best prior guess.
This architecture has clear advantages. Predictions speed up processing, fill in gaps when sensory data is noisy or incomplete, and allow you to react quickly. But they also mean that your prior expectations actively shape what you perceive. If your brain predicts you will see a weapon, ambiguous visual input is more likely to be interpreted as a weapon. If you expect someone to be incompetent, neutral behavior is more likely to be read as confirming that expectation. The prediction machinery does not distinguish between useful priors and harmful stereotypes; it applies the same computational logic to both.
Dopamine signaling plays a role in this system. Dopamine neurons encode prediction errors, firing more when outcomes are better than expected and less when outcomes disappoint. The striatum, amygdala, and frontal cortex also participate in this error-coding process.12PubMed Central. Dopamine reward prediction error coding This means that surprising, expectation-violating information gets a neurochemical flag, which is part of how the brain learns. But it also means that when a preconception correctly predicts an outcome, the absence of a prediction error feels satisfying and reinforces the existing model, even if the prediction was self-fulfilling.
An Evolutionary Perspective on Jumping to Conclusions
Some of the preconceptions that cause the most trouble in modern life may have been genuine survival advantages for most of human evolutionary history. Research on threat detection suggests that the nervous system evolved to anticipate danger before it arrives, simulating possible encounters with threats and selecting defensive responses in advance.13PubMed Central. The ecology of human fear: survival optimization and the nervous system A system that leans toward false positives (seeing danger that is not there) is safer than one that occasionally misses a real threat. Risk assessment in animals involves rapid analysis of the type, distance, and location of a threat, combined with features of the environment like available escape routes.14PubMed. Risk assessment as an evolved threat detection and analysis process
These defensive systems develop on a schedule that tracks the kinds of dangers a child is likely to encounter. Fears of particular threats emerge at developmental stages when those threats would have been historically relevant: strangers become frightening around the time an infant begins to crawl away from caregivers, heights become alarming once a child can walk, and social fears ramp up during adolescence.15PubMed. Threat-detection in child development: an evolutionary perspective The preconceptions baked into these systems, that unfamiliar things may be dangerous, that certain animals deserve more vigilance, that social rejection is threatening, served real protective functions. The problem is that the same quick-and-dirty heuristics now operate in environments full of diverse strangers, abstract statistical risks, and media designed to trigger alarm, contexts where snap judgments based on surface features routinely lead you astray.
Social Preconceptions and Implicit Bias
When preconceptions involve social categories like race, gender, age, or appearance, they become particularly consequential. Decades of research using measures of implicit bias have found that people’s automatic associations between social groups and positive or negative attributes predict discriminatory behavior in real-world settings.16PubMed Central. Implicit-Bias Remedies: Treating Discriminatory Bias as a Public-Health Problem These associations operate quickly and often outside conscious awareness, which is what makes them so difficult to override through willpower alone.
One specific mechanism that helps social preconceptions spread is the halo effect: when a positive impression in one domain (say, physical attractiveness) bleeds into judgments of unrelated traits (intelligence, trustworthiness, competence). Research has shown that the halo effect and a related generalization effect work together. People rated as attractive are judged more positively on unrelated traits, and those positive evaluations can even transfer to unfamiliar faces that merely resemble the attractive person.17Acta Psychologica Sinica. Beauty is in the eye of the beholder: The halo effect and generalization effect in the facial attractiveness evaluation In practical terms, this means that a first impression, which is itself shaped by preconceptions about what “good” looks like, can cascade into a web of subsequent judgments that all feel individually reasonable but share a single biased root.
Political Preconceptions and Motivated Reasoning
Political beliefs offer a vivid demonstration of preconceptions in action, because the stakes feel personal and the evidence is often ambiguous enough to interpret in multiple directions. Experimental work has shown that when political values are made salient, people on both the left and the right shift their factual beliefs to align with the average beliefs of their political group.18European Economic Review. Motivated political reasoning: On the emergence of belief-value constellations This is not people lying about what they believe to look good. The shift appears to be genuine: values shape what people think is true, not just what they say is true.
Brain imaging research paints a similar picture. When processing political messages, people endorsed claims more when they came from an in-group source and when the message content was personally desirable, and less when the source was out-group or the content unwelcome. Reward-related brain regions including the ventral striatum and medial prefrontal cortex showed increased activity that predicted the subsequent over-endorsement of desirable messages.19PubMed Central. Tracking politically motivated reasoning in the brain: the role of mentalizing, value-encoding, and error detection networks In other words, your brain appears to treat belief-confirming political information partly like a reward, which helps explain why consuming agreeable political content can feel good in a way that goes beyond merely being informed.
The influence of preconceptions on credibility judgments is worth highlighting. You might assume that people evaluate the quality of a source first and then decide what to believe based on that assessment. But research suggests that credibility judgments are themselves shaped by directional motivated reasoning, meaning you are more likely to find a source credible if it tells you what you already want to hear.20Journal of Language and Social Psychology. The Roles of Identity- and Belief-Driven Motivated Reasoning and Source Credibility in Persuasion on Climate Change Policy This creates a trap: you dismiss inconvenient evidence because the source seems unreliable, but you judged the source unreliable partly because the evidence was inconvenient.
The Anchoring Effect and Everyday Decisions
Not all preconceptions come from deep-seated beliefs or cultural conditioning. Some are astonishingly cheap to install. The anchoring effect is one of the most reliable findings in psychology: when you encounter a number before making a judgment, that number pulls your estimate toward it, even when it is obviously arbitrary. Ask someone whether the population of Turkey is more or less than ten million, then ask them to guess the actual figure, and their answer will be lower than if you had first asked about 100 million. The number serves as a cognitive anchor that the brain adjusts away from, but not enough.21PubMed Central. Individual Differences in Anchoring Effect: Evidence for the Role of Insufficient Adjustment
Anchoring matters because it shows how easily and invisibly preconceptions can be introduced. The “preconception” in this case is just a recently encountered number, yet it measurably warps your subsequent judgment. In negotiations, pricing, legal sentencing, and medical decision-making, whoever sets the initial anchor wields disproportionate influence. This is not a flaw limited to careless thinkers. It affects experts as much as novices, although awareness of the bias can slightly reduce its grip.
When Experts Get Trapped
If you assumed that education and professional training would immunize people against preconceptions, the evidence is discouraging. In clinical medicine, cognitive biases including availability bias, confirmation bias, and anchoring are increasingly recognized as major contributors to diagnostic error.22PubMed Central. Bias in Clinical Practice A physician who recently saw a case of meningitis may over-diagnose it in the next patient with a headache (availability bias). A specialist who forms an early hypothesis may selectively seek confirming test results while overlooking contradictory signs (confirmation bias). These are the same cognitive mechanisms that produce preconceptions in everyday reasoning, just operating in a high-stakes environment where the consequences are more visible.
Geology provides an illustrative example from the history of science. Research on how competing paradigms in stratigraphy developed and persisted found that geologists’ preconceptions shaped what they observed in the field, that the working environment contributed to consensus around particular theories, and that theoretical arguments were sometimes accepted by the community without the usual standards of documentation and primary data.23Earth-Science Reviews. Sequence stratigraphy as a scientific enterprise: the evolution and persistence of conflicting paradigms The study also found that citation networks and clusters of affiliated researchers could either advance or block the spread of geological knowledge. Experts, in other words, are not simply individual thinkers who might be biased. They operate in social structures that can amplify the preconceptions they already hold.
Overconfidence and the Limits of Self-Assessment
One reason preconceptions are so persistent is that people tend to be overconfident about their own reasoning, especially when they lack relevant knowledge. Research on the Dunning-Kruger effect found that low performers greatly overestimate their ability while top performers make more accurate self-assessments. The original explanation centered on a lack of metacognitive insight: the same skills needed to perform well are needed to recognize poor performance. But closer analysis found that the main driver of the overestimation was simply the low level of task performance itself, with metacognitive ability playing a weaker role than originally thought.24PubMed Central. Wise up: Clarifying the role of metacognition in the Dunning-Kruger effect For preconceptions, the implication is clear: the people most likely to hold inaccurate beliefs are also the least equipped to recognize that those beliefs are wrong, not necessarily because they lack insight, but because they lack the knowledge that would make the error visible.
What Actually Works to Change Preconceptions
Given how deeply rooted preconceptions are and how many cognitive mechanisms protect them, the question of how to change them matters enormously. One of the most extensively studied tools in education is the refutation text, a passage that explicitly states the common misconception, directly refutes it, and then explains the correct idea. A meta-analysis covering thousands of participants found that refutation texts produce a moderate positive effect on learning compared to other instructional approaches, and the effect held up across a wide variety of topics and contexts.25PubMed Central. Refutation Text Facilitates Learning: a Meta-Analysis of Between-Subjects Experiments
The reason refutation texts work better than simply presenting the correct information is that they activate the misconception and then create a cognitive conflict. Simply hearing the right answer does not engage the wrong answer that is already sitting in memory. If you never confront the existing schema, the brain has no reason to revise it. Refutation texts force the brain to acknowledge the mismatch between the old belief and the new evidence, which is what the conceptual change literature has long argued is the key to durable learning. This does not mean that a single paragraph can erase years of intuitive thinking, but it consistently outperforms the strategy of ignoring the misconception and hoping the truth will win on its own.
When Machines Inherit Human Preconceptions
Artificial intelligence adds a new dimension to the problem. AI systems trained on human-generated data can internalize the preconceptions present in that data, reproducing and sometimes amplifying biases that their designers did not intend. In medical imaging, for example, algorithms trained on datasets that underrepresent certain populations or overrepresent particular diagnostic patterns can perform worse for the groups they were not trained on. Research on AI bias in radiology has noted that while AI has the potential to reduce human cognitive biases in image interpretation, extensive work has documented AI’s tendency to absorb biases from its training data instead.26PubMed Central. Bias in artificial intelligence for medical imaging: fundamentals, detection, avoidance, mitigation, challenges, ethics, and prospects
This creates a feedback loop analogous to the one that operates in human cognition. A biased dataset trains a biased model, the model’s outputs get treated as ground truth, and future datasets get shaped by those outputs. The preconception, originally human, becomes encoded in a system that processes millions of cases without the metacognitive ability to question its own assumptions. The corrective strategies being developed for AI bias, such as diversifying training data, auditing outputs for disparate performance, and building in explicit fairness constraints, mirror the strategies that work for human preconceptions: making the bias visible, creating a mismatch between expectation and reality, and redesigning the environment so the default path is less biased rather than relying on individual vigilance.