Fundamentals of Empirical Research | empirio

Ask three students what “empirical” actually means and you get three answers: running your own survey, working with numbers, or simply the opposite of theory. All three fall short.

Empirical research answers a research question systematically, with data about observable reality rather than with literature and argument alone. Whether that data consists of numbers or of transcripts, and whether you gather it yourself or take it from an archive, does not decide whether your work is empirical. By the end of this article you will know which route suits your question and how to defend it in your methods chapter.

📌 The key points at a glance

• Empirical research answers a research question with data instead of with argument.
• Empirical is not the same as quantitative. Interviews are just as empirical.
• Ready-made data counts too: the UK Data Service releases open data without registration.
• Before your first question come a lawful basis and, often, ethical approval.
• A literature-based dissertation is not an escape route but a method in its own right.

What is empirical research? Definition and boundaries

Empirical research is a way of working in which a research question is answered through the systematic collection and analysis of data about observable reality. The opposite pole is not literature, but pure argument.

The word goes back to the Greek empeiria, meaning experience. Therein lies the risk of confusion, because anyone who asks around among friends has experience too. What separates the empirical approach from everyday observation is not the experience itself but the way it is handled.

The difference comes down to four points. Everyday knowledge accumulates by accident, empirical research follows a plan set out in advance. Everyday knowledge keeps whatever fits the picture, empirical research also records the awkward cases. Everyday knowledge manages with vague terms, empirical research pins every term down until it can be measured. And everyday knowledge stays private, while empirical research is written up so that someone else can repeat it.

One further point matters just as much: not every piece of academic work is empirical, and that is no shortcoming. A mathematical proof, a legal interpretation and a philosophical clarification of terms all count as academic work without a single figure collected. Empirical enquiry is a tool for a certain kind of question, not a badge of scholarly quality.

Two misconceptions you will find almost everywhere

Two claims about empirical research are so widespread that hardly anyone checks them. Both sound plausible, and both push students towards a more laborious design than their question actually calls for.

We will take them one at a time, because they do damage in different places. The first misconception costs you suitable methods, the second costs you suitable data.

The misconception that empirical means numbers

Empirical and quantitative are not the same thing. Six semi-structured interviews with nurses collect data about observable reality, analyse it systematically and are therefore fully empirical, even though no percentage appears at the end. The quantitative versus qualitative axis describes the kind of data, not the question of whether data is collected at all.

The misconception has a visible side effect. Because many guides declare objectivity, reliability and validity to be preconditions for any empirical work, qualitative dissertations get judged against a standard they were never built for. Those three criteria come from quantitative test theory. Qualitative research has catalogues of its own, and the wrong one makes sound work look methodologically weak.

The misconception that empirical research is the opposite of a literature review

Setting empirical work against literature-based work is a simplification that undermines itself on closer inspection. A meta-analysis and a systematic review draw exclusively on existing material and are still empirical: they gather, code and analyse findings from published studies. The same guides that assert the contrast list meta-analysis as an empirical method two paragraphs later.

The clean line runs not between your own data and other people’s literature but between two activities: working through a body of material systematically, or ordering and interpreting the state of research. The first is empirical, even without fieldwork of your own. The second is theoretical work, even if hundreds of studies get read.

💡 Tip

Test your planned project against one question: at the end, will there be a body of material that you worked through according to rules set in advance? If the answer is yes, your work is empirical. Whether that material consists of questionnaires, transcripts or 40 coded studies makes no difference.

Empirical research methods: qualitative, quantitative and mixed

The choice of method follows from the research question, not from personal taste. A question about frequencies, distributions or relationships calls for numbers. A question about meanings, motives or processes calls for text. Anyone who picks the method first and bends the question afterwards notices at the analysis stage.

In practice that means phrasing your question precisely enough that the kind of data follows from it. “How satisfied are second-year students with academic support?” calls for a standardised survey. “What makes students say a supervision meeting helped them?” calls for interviews. Both questions concern the same topic and lead to completely different research methods. Where the vocabulary itself is still fuzzy, the article What is a poll? separates poll, survey and questionnaire.

Quantitative methods: measuring and comparing

Quantitative research defines a characteristic so that it can be captured the same way across many cases. The instrument used in quantitative research is fixed beforehand and is not altered during fieldwork, because comparability would break. Standardised surveys, structured observation and experiments are the usual formats. The gain is comparability, the price is depth: whatever the questionnaire did not anticipate never shows up in the data.

Qualitative methods: understanding and reconstructing

Qualitative research deliberately keeps the procedure open so that things nobody expected can surface in the material. Qualitative research covers semi-structured interviews, focus groups, participant observation and qualitative content analysis. The topic guide sets themes, not wording and not sequence. Case numbers are small, analysis is laborious, and the findings hold in the first instance for the cases studied, not for a population.

Mixed methods: both in one project

Mixed methods combines both approaches within one research design instead of splitting them across two projects. The most common sequence in a dissertation starts with a handful of interviews, simply to understand the field, and then tests the resulting hunches with a standardised questionnaire. The reverse works just as well: survey first, then interviews to explain the odd results. What matters is that the link is justified, or two half studies read like one overloaded project.

Where your data comes from: collect it or reuse it

Almost every guide assumes that empirical work begins with fieldwork of your own. Running a survey yourself is the best known route, but not the only one, and for many questions not the best one. The alternative is secondary analysis: you work with data other people collected and put your own question to it.

The difference is substantial. Your own fieldwork typically costs several weeks for instrument, data collection and cleaning, and often ends with a sample drawn from people you already know. An existing dataset hands you thousands of cases on day one, with a documented sampling method. What you give up is freedom over the questions: you can only analyse what was asked.

Four routes to your data

Four routes are open to you in practice, and they do not rule one another out. Plenty of strong dissertations pair a small piece of original fieldwork with a reference dataset that shows how the sample compares with the wider population.

1. Your own fieldwork. You build an instrument and run the data collection yourself. Full control over the questions, full workload.
2. Social science research data. The UK Data Service, led by the UK Data Archive at the University of Essex and funded by UKRI through the ESRC, holds three tiers: open data available “without registration”, safeguarded data that “can be downloaded by registering”, and controlled data released only through SecureLab.
3. Official microdata. Survey microdata from the Office for National Statistics reaches students through that same service, and registration runs on your institutional login, since “you can register using the username and password provided by your institution”.
4. Published findings. For a systematic review or a meta-analysis, the results of other people’s studies are themselves the material.

One limit belongs in the picture. The UK Data Service states that controlled data are “not suitable for use by inexperienced researchers, such as undergraduates”, so an undergraduate dissertation realistically works with the open and safeguarded tiers, both described on the UK Data Service access conditions page (as at September 2026).

Voluntary or compulsory: two collections from one office

Two flagship collections run by the Office for National Statistics follow opposite rules, which makes them a useful pair to think with. The Annual Population Survey is voluntary, and the ONS page on household and individual surveys puts the achieved sample at “approximately 80,000 households (or 175,000 respondents)” and states plainly that “No-one has to take part if they do not want to”.

The census works the other way round. Taking part is compulsory under the Census Act 1920, and a person who fails to comply is “liable on summary conviction to a fine not exceeding level 3 on the standard scale” under section 8(1). Section 8(1A) carves out an exception in England and Wales, where questions on religion, sexual orientation and gender identity carry no penalty. The last census there was held on 21 March 2021.

Should you go for fieldwork of your own, the online survey is the most common route. Why it works and who it misses is covered in the article on the online survey as an empirical research method, and the comparison with postal, telephone and face-to-face modes sits in the article on the online survey with its advantages and limits.

How an empirical study runs: six decisions

Methods textbooks divide an empirical study into anything from five to nine phases. The number matters less than the insight behind it: at six points you take a decision that fixes everything that follows and can barely be corrected later.

The order of the six decisions is not a recommendation but a chain of dependencies. Anyone who picks the analysis technique only after fieldwork sometimes finds that the scales do not permit the intended calculation. Which is why step six belongs, in your head, to step three.

1. Set the research question. Your research question has to be narrow enough that a data collection can be derived from it. It also fixes the aim: describing, explaining or exploring.
2. Form hypotheses where they belong. Testing studies need hypotheses, exploratory ones do not. An exploratory project with forced hypotheses is a common beginner error.
3. Choose the research design. Your research design fixes method, timing and case selection. Whether you collect once or repeatedly is settled here too.
4. Make concepts measurable. “Student satisfaction” is not a variable, it is a word. Only the translation into concrete items and response scales makes it collectable.
5. Determine the sample. Method and size of your sample decide what you may claim at the end and what you may not.
6. Fix the analysis technique. Which of the analysis methods is open to you depends on your data and level of measurement, not on your preference.

The full sequence with examples and a timetable sits in the article on the empirical research process.

Data protection and ethics: what has to be settled first

As soon as you ask people questions, you are processing personal data, and you are doing it before the first answer is analysed. Almost every guide to empirical research skips the point, even while sending students straight into interviews and surveys. Nothing about it can be repaired afterwards, because fieldwork without a lawful basis does not become lawful when the data gets anonymised later.

The rule of thumb that carries this chapter is short. If your data is genuinely anonymous, data protection law does not apply to it at all; if it is not, you need a lawful basis before you collect anything. In the United Kingdom the relevant law is the UK GDPR together with the Data Protection Act 2018, and the regulator is the Information Commissioner’s Office.

Lawful basis: anonymity first, consent second

Research purposes are recognised in UK law, but recognition is not a free pass. Schedule 1, Part 1, paragraph 4 of the Data Protection Act 2018 covers processing “necessary for archiving purposes, scientific or historical research purposes or statistical purposes” where that processing “is in the public interest”. For a student project the practical order stays the same: strip out anything identifying, and where you cannot, secure informed consent and record what you told participants.

Ethical approval: it depends on your institution

No single national rule governs ethical approval for student projects, so the answer sits with your own institution. Some are strict. The UCL Institute of Education states that “All student research is required to gain ethical approval before starting”, and for undergraduate and taught postgraduate work two people review rather than a committee: “The supervisor and an extra member of the academic staff should review” (as at September 2026).

Other institutions draw the line at risk rather than at status. The framework for research ethics published by the Economic and Social Research Council within UKRI notes that “student research may be managed at school or department level” and warns that “It cannot be assumed that all students’ projects involve minimal risk.” Ask your supervisor early, since approval that takes three weeks will not fit into the fortnight before your deadline.

⚠️ Watch out

A survey run for your dissertation is not a private survey. Where a university, a charity or a company stands behind the fieldwork, the survey tool acts as a processor, and the UK GDPR requires a written data processing agreement with it. Settle that before you choose the tool, not after the responses arrive.

Responsibility for the individual case sits with the data protection officer at your institution, and for ethical questions with your supervisor. This article is not legal advice.

When a dissertation does not have to be empirical

Many students start from the decision to do something empirical and then look for a question to match. That order is the wrong way round and produces the familiar project with 43 responses from the student’s own cohort.

The honest test runs in the opposite direction. Some questions are better served by existing literature or existing data than by fieldwork of your own. And some regulations require an empirical component outright. Both belong settled before you register a topic.

A literature-based dissertation is not a fallback for students who cannot find participants. Working through the literature is a method with quality standards of its own: a search strategy someone else could repeat, a justified selection of sources, and interpretation instead of listing. Done well, it beats a thinly populated survey.

Naming tracks the type of degree, not the difficulty. The Quality Assurance Agency for Higher Education ties the dissertation to taught degrees and the thesis to research degrees, where a research master’s “usually requires a dissertation or thesis” and “The thesis is commonly defended in an oral examination”. A thesis in the United Kingdom is therefore not automatically a doctorate.

Your own regulations have the final word. If they demand an empirical component, the article on the empirical bachelor’s and master’s dissertation shows how that part fits the chapter structure.

How to tell that an empirical study holds up

Whether an empirical study holds up is decided not by sample size but by whether the procedure fits the question and is disclosed. Quality criteria exist for exactly that, and the most common mistake is reaching for the wrong catalogue.

Quantitative work is judged on objectivity, reliability and validity. Objectivity means the result does not depend on who collects and analyses. Reliability means a repetition under the same conditions produces the same result. Validity means the instrument measures what it claims to measure. That third definition often gets garbled into a question about whether the results answer the research question, which is something else entirely.

Qualitative work has catalogues of its own, because repeatability is no sensible yardstick there. At the centre sit transparency of procedure, the fit between method and subject matter, and an honest account of your own role in the field. Which catalogue suits your project is set out in the article on the quality criteria of qualitative and quantitative research.

A study does not become good by having many cases. A study becomes good when someone else can follow what you did.

One standard reference covers both approaches in a single volume: Bryman’s Social Research Methods, now in its seventh edition under Tom Clark, Liam Foster, Luke Sloan and Charlotte Brookfield and listed by Oxford University Press (as at September 2026). Check which edition your reading list names, because earlier ones carried Alan Bryman as sole author.

Common mistakes in a first empirical project

Four mistakes turn up again and again in dissertations. What they share is that they arise early and surface late, usually once the data is collected and nothing can be salvaged.

Knowing about them lets you steer around them at the planning stage. Steering costs a few days and can save an entire round of fieldwork.

The method comes before the question

Deciding to run a survey and then wondering what to ask reverses the dependency. The result is a questionnaire that touches on plenty and answers nothing. The question comes first, and the method follows from it. Where the method is fixed for practical reasons, perhaps because your supervisor requires it, the question adapts before the questionnaire gets built.

The sample is your own social circle

A survey shared through your course group chat and Instagram reaches people who resemble one another. No great harm is done as long as the methods chapter says so and the findings are worded with care. Trouble starts when the conclusion talks about students in general. Whether a study is representative depends on the sampling method, not on the number of responses.

The level of measurement gets checked only at the analysis stage

Whether you may calculate a mean, run a correlation or only count frequencies depends on the level of measurement of your variables. Which level of measurement a question has is settled when the questionnaire is built, not when the analysis starts. Realising in your statistics software that every item is nominal leaves you unable to run the planned calculation.

The methods chapter describes instead of justifying

Writing that an online survey was carried out is a description. What markers expect is a justification: why this method suits this question, what the alternative would have been and why it fell away. Three sentences usually do it, and they separate a documented project from an argued one. Almost no guide mentions the point, and almost every viva asks about it.

Conclusion

Empirical research is less a technique than a stance: you decide in advance what will settle your question, and you stick to it even when the answer turns out awkward. Whether numbers or transcripts come out of it, and whether you collect the material yourself or draw on an existing dataset, is a question of fit rather than rank. Check first whether your question needs data at all.

Where to go next

• You want to know how the empirical part fits the chapter structure? Writing an empirical dissertation
• You are unsure whether your study is descriptive or explanatory? Exploratory, descriptive, explanatory, causal
• You have the data and need the right analysis? Descriptive and inferential statistics
• You are building your questionnaire right now? Creating a questionnaire
• You have the methods chapter but the tone is off? Academic writing

Ready to collect your own data?

With empirio.ai, an online survey tool from Germany, you build your survey on the free plan, host the data in the EU and get the data processing agreement thrown in.

Create a survey for free