Hypothesis vs. Theory: The Real Difference

Short answer: a hypothesis is a single, tentative, testable prediction about a specific observation. A theory is a well-substantiated, broad explanatory framework, built up from many tested hypotheses, facts, and laws, that has survived repeated attempts to falsify it. A theory is not an unproven guess waiting to be promoted to “fact” — in science it is closer to the opposite: it is what a hypothesis becomes only after extensive, independent testing has failed to disprove it.

Hypothesis vs. theory at a glance

	Hypothesis	Theory
Scope	A single, specific, testable prediction about one relationship or phenomenon	A broad explanatory framework that accounts for many related observations, facts, and hypotheses
Evidence base	Not yet tested, or tested only a small number of times	Repeatedly tested, corroborated by independent lines of evidence over time
Function	Generates a specific, falsifiable prediction that an experiment or observation can check	Explains why a pattern exists and predicts new, previously unobserved phenomena
Status if disproven	Rejected or revised; a normal, expected outcome of testing	Revised to accommodate the new evidence, or — rarely, and only after sustained failure to account for contrary evidence — abandoned
Example	“If gravitational attraction depends on mass, a heavier and lighter object dropped in a vacuum will still hit the ground at the same time.”	Einstein’s general theory of relativity: gravity is the curvature of spacetime caused by mass and energy

Why “it’s just a theory” gets the science backwards

In everyday English, “theory” is often used to mean a hunch, a guess, or an unproven idea — as in “I have a theory about why the meeting ran long.” That everyday sense is close to what scientists mean by hypothesis, not by theory. The two words are not synonyms in a scientific context, and they are not two points on the same ladder where a hypothesis “graduates” into a theory and a theory eventually “graduates” into a law.

The National Academy of Sciences, in the glossary to its evolution-education guidance, defines the three terms this way:

Hypothesis: “A tentative statement about the natural world leading to deductions that can be tested.” If the deductions hold up, the hypothesis is provisionally supported; if they don’t, it is rejected or revised.
Theory: “A well-substantiated explanation of some aspect of the natural world that can incorporate facts, laws, inferences, and tested hypotheses.”
Law: “A descriptive generalization about how some aspect of the natural world behaves under stated circumstances.”

Under this framework, calling something “just a theory” as a way of dismissing it misreads the term: a scientific theory (germ theory, atomic theory, plate tectonics, evolution by natural selection, general relativity) is the most rigorously supported kind of scientific explanation there is, not the least. Organizations that teach the philosophy and public understanding of science, including the National Center for Science Education, make the same point for the same reason: the informal and technical senses of “theory” are frequently conflated in public debate, and the conflation is usually doing rhetorical work rather than describing how scientists actually use the word.

Theory and law are not a hierarchy either

A related misconception is that theories “become” laws once they are proven. They don’t, because theories and laws answer different questions. A law is typically a compact, often mathematical, description of a regularly observed pattern — what happens, and under what conditions. A theory explains why that pattern exists. Newton’s law of universal gravitation is a mathematical statement of how two masses attract each other (F = Gm₁m₂/r²); it does not say why mass causes attraction at all. General relativity is the theory that explains it. Both remain in active use today — the law is still the right tool for most everyday calculations, and the theory is what you need when gravity gets extreme (near a black hole, or for GPS satellite timing corrections). Neither one demoted or absorbed the other, and plenty of well-established theories, including most of the theoretical framework in evolutionary biology, have no corresponding “law” at all, because not every domain of science produces the kind of simple universal regularity a law describes.

From observation to theory: worked examples

The table below traces the same progression — observation, hypothesis, and the resulting theory (with a law alongside it where one exists) — across three unrelated fields, to make the pattern concrete rather than abstract.

Stage	Gravity	Infectious disease	Evolution
Observation	Unsupported objects consistently accelerate toward the ground	Illness clusters around contact with sick individuals or contaminated water; sterilized instruments and hand-washing measurably reduce infection rates	The fossil record shows a succession of forms over time; living species share nested patterns of anatomical and genetic similarity
Hypothesis	“Objects of different mass, dropped from the same height in the absence of air resistance, will fall at the same rate”	“A specific microorganism, not ‘bad air’ or spontaneous generation, causes a specific disease” — the tentative claim Robert Koch’s postulates were designed to test	“Populations change over generations because heritable traits that improve survival and reproduction become more common” — Darwin’s working hypothesis, tested against breeding, biogeography, and (later) genetics
Law (where one exists)	Newton’s law of universal gravitation — the mathematical relationship between mass, distance, and gravitational force	No widely-used formal “law”; the field is organized around theory rather than a compact mathematical law	The Hardy–Weinberg principle — a mathematical description of allele-frequency equilibrium in a non-evolving population, used as the baseline against which real evolutionary change is measured
Theory	General relativity — gravity as the curvature of spacetime by mass and energy	Germ theory of disease — the substantiated explanatory framework that specific pathogens cause specific diseases, now underpinning infection control, antibiotics, and vaccination	The theory of evolution by natural selection — the explanatory framework, now supported by genetics, paleontology, and biogeography, for how species change over time

How a hypothesis earns its way toward a theory

No single study elevates a hypothesis into a theory. What actually happens is cumulative: a hypothesis is operationalized into a testable, falsifiable prediction, tested, and — if it survives — tested again by independent researchers using different methods, samples, or instruments. A theory emerges once a large, converging body of such tested hypotheses, observations, and (where they exist) laws all point to the same explanatory account, and once that account has demonstrated it can also predict new observations that were not part of the original evidence, not merely explain the ones already known. This is also why a single disconfirming result rarely overturns an established theory outright: researchers first check whether the anomaly reflects a flaw in the new hypothesis, the measurement, or the experimental design, precisely because the theory itself is backed by so much independently-corroborated evidence. It takes a sustained, replicated pattern of disconfirmation — not one result — to force revision of a genuine scientific theory.

The requirement that a claim be capable of being contradicted by an observation, formalized by philosopher of science Karl Popper as the falsifiability criterion, applies at both levels: a hypothesis has to specify, in advance, an observation that would prove it wrong, and a theory has to remain, in principle, vulnerable to being overturned by future evidence — a claim that could never be wrong under any conceivable observation is not doing scientific work, however plausible it sounds.

Two words that look related but aren’t the same kind of “theory”

Two other CASRAI dictionary terms use “theory” in a related but distinct sense, and it’s worth not conflating them with the natural-science usage above:

Theory of change — a program-design and evaluation tool that maps how a specific intervention is expected to produce a specific outcome. It’s a planning artifact, not an empirically-tested explanatory framework in the natural-science sense.
Grounded theory — a qualitative research methodology in which theory is built inductively from coded data, rather than tested against a pre-specified hypothesis. The “theory” it produces is a substantive or formal explanatory account grounded in a specific dataset, evaluated by different standards (theoretical saturation, fit, workability) than a natural-science theory is.

Frequently asked questions

Can a hypothesis be proven true?

Not in the strict sense. A hypothesis can be supported, corroborated, or fail to be falsified by a given test, but science generally treats explanations as provisionally accepted rather than proven with certainty — later evidence or more precise instruments can always, in principle, require revision.

Is “theory” the same as “fact”?

No, and they’re not competitors either. A fact is a directly confirmed, replicable observation (for example, that species change over time is a fact, established from the fossil and genetic record). A theory is the explanatory framework that accounts for a body of related facts (for example, evolution by natural selection is the theory that explains why and how that change happens).

Do all scientific theories eventually become laws?

No. Theories and laws describe different things — a law is a descriptive pattern, usually mathematical; a theory is an explanation for that pattern — so one doesn’t convert into the other. Many well-established theories, particularly in biology, have no corresponding formal “law” at all.

What’s the difference between a hypothesis and a prediction?

A hypothesis is the underlying explanatory claim; a prediction is the specific, measurable outcome you’d expect to observe if the hypothesis were true. See how to write a testable research hypothesis for how the two are typically distinguished and stated in a research proposal.

How is the null hypothesis different from this discussion?

The null and alternative hypotheses are a statistical testing framework (a way of structuring a specific study’s inferential test), not a stage in the hypothesis-to-theory progression described here. See null hypothesis explained and null vs. alternative hypothesis for that distinction.

Hypothesis vs. Theory: What the Difference Actually Is in Science

Ask about Hypothesis vs. Theory: What the Difference Actually Is in Science