What is Science? Definition, Explanation and Examples

Quick test before we start: is astrology a science? Almost everyone says no straight away, then struggles to say why in one sentence.

Whether something counts as science has little to do with the subject and everything to do with how the work is done. Science is the systematic attempt to gain reliable knowledge about the world, test it and pass it on, using methods that are stated openly, evidence anyone can check and results that others are free to overturn. In British usage the word is narrower than it looks, which is where most of the confusion starts. By the end you will be able to judge whether a claim is actually evidenced or simply sounds like it.

📌 Key points at a glance

• Science gains knowledge systematically, openly and testably.
• In British English, science usually means the natural sciences.
• Natural sciences, humanities and social sciences are the three broad groups.
• The Concordat to Support Research Integrity governs UK research conduct.
• Pseudoscience avoids testing, science invites it.

What is science?

Science is the body of systematically gained, ordered and testable knowledge a society holds, together with the activity that produces it: research carried out using accepted methods, published so that other people can check the results.

The word carries two meanings at once, which is exactly why a single neat definition is hard to pin down. Science is the outcome, meaning the stock of reliable knowledge about the world. Science is also the process by which that knowledge comes about. Both halves rest on the same condition: the method is stated, so the result can be examined by somebody who was not there.

What British English means by “science”

In everyday British usage, science normally means the natural sciences: physics, chemistry, biology and the fields built on them. Universities reinforce this split, with faculties of science on one side and arts and humanities on the other, and school pupils choose between them long before they reach a degree. History, English literature and philosophy sit outside that boundary even though they are every bit as rigorous.

German, French and Italian draw the line somewhere else. The German word Wissenschaft covers history and philosophy without hesitation, and so does the Italian scienza. When you read a translated paper that claims something about “science”, it is worth checking which of the two meanings the original used, because the broader one changes what the sentence is actually saying. The safest umbrella term in English is research or scholarship, not science.

How do you tell whether something is scientific?

A claim is scientific when it is clear how the claim came about, and when other people could follow that route themselves and overturn the result. The subject matter does not decide this, the procedure does.

Telling procedure and subject apart will save you in every essay you write. A piece on tarot reading can be perfectly scientific if it examines why people believe in it and sets out its method. A piece on quantum physics is not scientific if it simply strings together assertions with nothing behind them. Five features carry the line:

Openness to being wrong is the most important of the five and the one most often missing. A piece of research says what it cannot show. In methodology this ambition goes under the headings objectivity, reliability and validity, and dissertations are marked against them too.

Being scientific is not about being right. It is about showing how you got there.

What disciplines exist in science?

Academic knowledge is usually divided into three broad groups: the natural sciences, the humanities and the social sciences. The division is a convenience rather than a law of nature, and the edges are genuinely blurred.

What separates the groups is the object of study and the methods used on it. The natural and social sciences lean heavily on observation and measurement, which makes them empirical. The humanities interpret more often than they measure, and their evidence is textual rather than numerical.

• Natural sciences: physics, chemistry, biology, astronomy, earth sciences.
• Humanities: history, philosophy, literature, linguistics, theology.
• Social sciences: sociology, psychology, politics, economics.

Several subjects refuse to sit in any one group. Business information systems straddles technology and commerce, medicine straddles the natural and the social, and mathematics is often left out of the empirical sciences altogether because it proves its statements rather than measuring them. Mathematics, logic and computer science are sometimes grouped separately as the formal sciences.

💡 Tip

If you need to place your own subject, look at the compulsory research methods modules in your programme handbook rather than the name of your faculty. The methods you are required to learn say far more about where the subject sits.

Science and pseudoscience: where the line runs

Pseudoscience borrows the vocabulary and the outward shape of science while avoiding the one step that makes science work: a serious chance of being proved wrong. Pseudoscience looks for confirmation where science goes hunting for counter-examples.

The philosopher of science Karl Popper, who spent much of his career at the London School of Economics, gave the best known test for this: falsifiability. A statement counts as scientific only if you can say what observation would refute it. “A change is coming in your life” cannot be refuted and therefore says nothing. “This drug lowers blood pressure by an average of 10 mmHg” can be refuted, and that makes it a scientific claim.

In practice you spot the difference through recurring patterns:

✓ The method is set out, even where it is inconvenient
✓ Contrary findings are named and discussed
✕ Criticism is treated as an attack or a conspiracy
✕ Evidence consists only of individual cases and testimonials
→ The test question: what result would prove this wrong?

People often say that science is the opposite of belief. That is not quite right. Research rests on assumptions too, starting with the assumption that the world behaves in regular patterns. The difference lies not in whether assumptions are made, but in whether they are declared and opened up to challenge.

What research integrity asks of you

Research integrity is the set of rules that govern how research is carried out and written up: document honestly, attribute sources, keep your data and correct mistakes as soon as you notice them. Students are held to the same principles as experienced researchers, only on a smaller scale.

In the United Kingdom the governing document is the Concordat to Support Research Integrity, and a refreshed edition was published on 4 April 2025, with organisations expected to align with it by April 2026. Alongside it sits the UK Research Integrity Office Code of Practice for Research, first issued in 2009 and updated in July 2025 to cover artificial intelligence (UKRIO, Concordat, as at September 2026). Your own university turns these commitments into regulations, and those regulations are what you are marked against.

The four steps that make a piece of work scientific

Four steps decide whether an essay or a dissertation counts as proper research. They follow on from one another, and skipping one tends to surface at the viva rather than before it. None of the four demands knowledge beyond a first-year methods module.

1. Set the question. Work without a clear question has no criterion for what belongs in it. The difference between a thesis, a hypothesis and a research question is not paperwork, it decides the structure.
2. Choose a method and justify it. Literature review, experiment or your own data collection: the choice belongs in the text, together with the reason it suits the question.
3. Collect and document data properly. Anyone gathering their own data records the period, the sample and the procedure. The empirical research process sets out these phases in detail.
4. Reference everything you borrow. Every claim taken from elsewhere gets its source. The practical rules are in the guide to correct referencing.

Documentation, not statistics, is where most dissertations come unstuck. Anyone running a survey needs the exact wording of the questions, the number of respondents and the fieldwork period long after the fieldwork is over. With empirio.ai, an online survey tool built in Europe, you create the survey and keep all of that in one place alongside the responses, which makes the methods chapter considerably easier to write.

⚠️ Careful

Your university regulations always take precedence. They may go further than the Concordat, particularly on referencing style, the use of AI tools and whether raw data must be deposited. Research involving people usually needs ethics approval before you collect anything. Ask your supervisor before you start, not afterwards. This article is not legal advice.

Three myths about science

A handful of ideas about science circulate widely and cause trouble in almost every discussion. They sound reasonable, they are wrong, and they turn up with striking regularity in the opening paragraphs of student essays.

The three below are the ones you will meet most often. Knowing them makes for noticeably steadier arguments, inside the university and outside it.

Myth 1: science delivers final truths

Scientific findings are always provisional. They hold until better data or better methods correct them, and that willingness to be corrected is the strength of the system rather than a weakness in it. A recommendation changing after new studies appear is therefore not a sign of unreliability. It is the process doing exactly what it is supposed to do.

Myth 2: if it is in a study, it is proven

A single study rarely proves anything. It supplies one finding under particular conditions, with a particular sample and a particular analysis. Only when several independent investigations point the same way does this harden into reliable knowledge. Before citing anything, look at the sample size, the fieldwork period and who paid for the work.

Myth 3: the humanities are not real research

The charge against the humanities confuses method with rigour. A reading of a literary text is scholarly when it sets out the textual basis, argues for its interpretation and names the competing readings. It measures nothing, yet it works systematically, with evidence and open to challenge. Measurement was never one of the conditions for rigour, it is simply one way of producing evidence.

Conclusion

Science is not a collection of subjects, it is a way of working: systematic, openly documented and permanently open to challenge. Check those three properties and you can judge a source in a couple of minutes, whether it is a meta-analysis or a post on social media. For your own work it comes down to one instruction: write down what you did, and say what you cannot show.

Where to go next

• Want to see how research works in practice? Fundamentals of empirical research
• Writing your first assignment? Academic work: definition and typical process
• Looking for the right structure? Dissertation structure explained

Need your own data for a dissertation?

With empirio.ai you can create a survey in minutes that keeps the questions, the fieldwork period and the responses together, so your methods chapter holds up later.

Create a survey for free