Empirical: What It Means, With Examples

“This needs to be empirical, not just a lit review.” Feedback like that lands on a lot of thesis and capstone proposals, and the word usually gets looked up right afterward. Reasonable enough: empirical shows up constantly in coursework and gets defined almost never.

Empirical means “based on observation or experience”. A claim or a study is empirical when it rests on systematically collected observations of the real world, meaning data from surveys, observation, experiments or content analysis, rather than on reasoning, logic or the existing literature alone. By the end of this article, one question will be enough to classify your own project.

📌 The key points

• Empirical means gathered from observation, not from reasoning.
• Empiricism is a philosophical position, empirical a property.
• The word traces back to Greek peîra, meaning “trial”.
• Qualitative interviews count as empirical too.
• Empirically supported does not mean proven.

What does “empirical” mean?

Empirical describes knowledge that comes from observing the world rather than from thinking about it. Merriam-Webster gives three senses that matter here: originating in or based on observation or experience, relying on experience or observation alone, and capable of being verified or disproved by observation or experiment.

The third sense is the useful one for research, and the quiet word in the first is “originating”. Having experienced something does not yet give you an empirical finding. What turns experience into evidence is the procedure around it: a defined set of cases, a consistent way of collecting, a documented analysis. Three conversations in the dining hall are an impression. Two hundred completed questionnaires are a result.

The routes to that kind of data are few. In empirical social research there are four: participant or non-participant observation, the experiment, the written or spoken survey, and content analysis. Anything gathered that way and then analyzed systematically counts as empirical material, and the steps together form the process laid out in the guide to empirical research.

One US-specific step comes before any of it. At most American colleges and universities, a study that involves human participants goes to an Institutional Review Board for review before data collection starts. Whether your particular project needs that review, and in what form, is decided by your own institution, not by your advisor and not by this article.

Where the word “empirical” comes from

Empirical goes back to Greek émpeiros, “experienced, skilled”, which contains peîra, the word for “trial” and “experience”. Latin empiricus carried it forward, and Merriam-Webster dates the first known English use of empirical to 1576.

What happened in between is the good part. In the ancient world, empirics were members of a school of physicians who treated patients using what had been observed to work, rather than what theory predicted. That put them in direct opposition to Galen, whose system dominated European medicine into the seventeenth century, and being the losing side cost them their reputation. Merriam-Webster notes that calling someone an empiric was at times close to calling them a quack.

The word kept a critical edge for a long while and can still carry one, in the sense of relying on observation alone without regard for system or theory. In research writing, though, the positive sense has taken over: evidence grounded in observation and open to being verified or disproved. The insult became the standard.

Empirical was once the charge that you had experience and no theory. Today it is the compliment that you have more than theory.

Empirical, empiricism and empiric: three words, one root

Empirical, empiricism and empiric share a root and mean three different things: a property, a philosophical position and, in older usage, a person. Everyday speech blurs them together, and in a thesis that shows.

Empiricism is the narrowest of the three. It holds that sense experience, rather than innate ideas, is the source of knowledge, and its usual counterpart is rationalism, which puts reason in that place. John Locke, George Berkeley and David Hume are the names that appear on every syllabus for it.

The practical takeaway is a relief. Doing empirical research is a method, not a creed. Nobody expects you to defend empiricism as a philosophy because you fielded a questionnaire.

How to tell whether a claim is empirical

A claim is empirical when you can say which observation would overturn it. If that question has no answer, what you have is a definition, a value judgment or a piece of logic, but not an empirical claim.

The test takes no specialist knowledge. Take a sentence, picture a world in which it is false, and ask how you would notice. If something concrete comes to mind, the sentence can be checked against reality.

The third row is the instructive one. “Tuition is too high” cannot be measured, while “73 percent of students say tuition is too high” certainly can. The judgment itself sits outside empirical research, and how widely it is shared is an ordinary empirical question. A great many research questions fail at exactly this point, and the same rewrite rescues them.

💡 Tip

Put “how many”, “how often” or “to what extent” in front of your research question as a test. If the sentence still makes sense, the question can be answered empirically. If it sounds absurd, you need either different data or a different question.

Empirical or theoretical: what it means for your thesis

Your thesis is empirical once you collect your own data or analyze existing raw data yourself. It stays theoretical as long as you review the state of research, clarify concepts or set positions against each other without collecting anything.

The literature review is where the argument usually starts. Pulling together twenty studies and reporting their results means working with empirical material, not working empirically. It becomes empirical only once the studies are selected against fixed criteria, coded and compared systematically, as in a systematic review. The same holds for statistics: quoting a figure is desk research, while analyzing a public-use dataset yourself is empirical work. The Census Bureau publishes its design and methodology for the American Community Survey precisely so that both uses can be checked.

Whether numbers come out at the end makes no difference to the label. Six transcribed interviews are as empirical as a survey with 500 responses, because both rest on data that was collected. What remains contested in the social sciences is what may count as experience at all: only what can be measured and counted, or also what resists being pinned to a number. What follows from that for your choice of method is covered in the article on qualitative and quantitative research methods.

Three mistakes that show up in the methods section

Three misunderstandings appear around the word empirical again and again, and each rests on an equation that looks plausible at first. In a methods section they matter, because that is where you justify what you did.

One question finds all three when you proofread: am I talking about where my data came from, or about something else that merely sounds similar?

Empirical is not the same as quantitative

Where the data came from and what form it takes are two separate matters. Eight coded interviews come from a collection process and are therefore empirical, even though the analysis contains not a single percentage. A literature review, in turn, does not become empirical by arranging other people’s figures into a tidy table. Treating the two as one leads you to justify the wrong decision.

Supported and proven are different words

An empirical finding is provisional: it holds until new observations speak against it. Academic writing is cautious for that reason, using phrases such as “the data support the assumption” or “the results suggest”. That caution is craft, not hedging. Why an inference from single cases to a general rule is never compelling is explained in the article on induction and deduction.

Empirical does not yet mean representative

A study can be impeccably empirical and still apply to almost nobody. Collecting 300 responses through your own Instagram story produces real data, but a sample that reflects your own contacts. Empirical says that data was collected, representative says who the numbers speak for. Where the line runs is explained in the article on the representative survey.

Conclusion

Empirical is not a badge of quality but a statement about where the data came from. What matters in your own work is therefore not whether it is empirical enough, but whether you can say in two sentences where your claims come from, how you collected them and who they apply to. Everything else, from sampling to analysis, builds on those two sentences!

Where to go next

• Want the whole process? Empirical research: definition, methods, guide
• Planning an empirical thesis? Writing an empirical paper
• Looking for the individual stages? The empirical research process
• Wondering how large your sample needs to be? The sample

The theory is ready, the data are missing?

With empirio.ai, an online survey tool from Germany, you can create a survey free of charge, collect the responses and export them for analysis.

Try the free survey tool