Two students, same department, same submission date. One analyses a ready-made dataset from a national archive, the other spends four weeks collecting answers. Both are writing an empirical dissertation.
An empirical dissertation answers its research question with data: you derive a testable assumption from theory, then either collect your own material or analyse an existing dataset, and you decide on the basis of that data whether the assumption holds. The work involved sits earlier than most people plan for.
📌 Key takeaways
- Empirical means the answer comes from data, not literature.
- The research process runs through seven phases (Raithel 2008).
- Methods, results and discussion stay strictly separate.
- The UK Data Service supplies survey data for dissertations.
- Ethics approval and fieldwork need weeks of lead time.
Create a survey for free
With empirio.ai you can create a modern online survey in minutes — with hosting in the EU.
- AI-built survey
- Adjust by drag & drop
- Real-time analysis
What is an empirical dissertation?
An empirical dissertation tests its research question against reality instead of answering it from the literature alone. The claim rests on observations, measurements or responses that are collected systematically and analysed in a way someone else could follow.
The word “empirical” comes from the ancient Greek “empeiria”, meaning experience. What it describes is not a hunch but a procedure with a fixed order. Empirical research therefore never starts with the data. It starts with a theory and the question that follows from it.
Collecting your own data is not a requirement. Analysing an existing dataset counts as empirical work too, as long as you document how you selected and analysed it. In the UK that route runs through the UK Data Service, funded by UKRI through the ESRC and hosted by the universities of Essex and Manchester among others. It holds the country’s largest collection of social and economic research data and even runs an annual dissertation award for undergraduate work based on archived data. Official statistics come from the Office for National Statistics.
Set against this is the theoretical dissertation, which draws its findings from existing knowledge. Which type suits your topic depends on the research question, not on how much work each one looks like. Our article on empirical or theoretical dissertations works through that decision.
When empirical work pays off and when it does not
An empirical dissertation pays off when your research question calls for a statement about reality that nobody has published in that form. Questions about frequencies, about the relationship between two characteristics, or about the effect of an intervention all point that way.
The appeal lies in the finding being yours. You produce figures or statements that did not exist before, and your conclusion rests on first-hand material. What you pay for that advantage is a set of dependencies you can only partly control.
What empirical work gives you
- You produce a finding nobody has published in that form.
- The practical angle opens doors to employers and institutions.
- The pure literature review stays leaner than in a theoretical dissertation.
- You practise methods that many graduate jobs expect.
What you take on in return
- Data collection eats weeks that must be fixed in the timetable.
- Your response rate depends on people you do not control.
- Analysis and statistics require learning a piece of software.
- Errors in the questionnaire are hard to repair once fieldwork starts.
If the drawbacks weigh heavily in your case, that argues for a narrower scope rather than against empirical work. One carefully run survey with a single clear hypothesis beats any over-ambitious study that stalls during fieldwork.
How does an empirical dissertation work, step by step?
An empirical dissertation runs through seven phases, from defining the problem to interpreting the results. Jürgen Raithel describes this sequence in “Quantitative Forschung” as the empirical research process (Raithel 2008, pp. 25–32).
The order is not a formality but a chain, in which each phase supplies the material for the next one. A skipped phase catches up with you during analysis at the latest, and by then fieldwork is over.
- Problem and research question. You narrow the topic and phrase a question that can actually be answered.
- Theory and hypotheses. From the state of research you derive testable hypotheses.
- Design. You fix your measures, instrument, study design and sample, and you plan the pilot test.
- Collection. Fieldwork begins and responses come in.
- Preparation. You clean the dataset and code the open answers.
- Analysis. You run the planned procedures and test the hypotheses.
- Interpretation. You relate the findings back to theory and research question.
Phase three decides the rest. Skip the pilot test and you often discover only during analysis that a question was ambiguous, at which point the responses cannot be rescued. Our article on the empirical research process explains each phase in turn.
Structure of an empirical dissertation
The structure of an empirical dissertation follows nearly the same logic everywhere: introduction, theory, research question, method, results, discussion. What separates it from a literature review is the standalone methods chapter and the strict split between finding and interpretation.
A model outline from the learning psychology unit at the Freie Universität Berlin shows how that looks in practice (as of September 2021). It splits the methods chapter into five subsections and treats aims and research questions as a chapter of their own alongside the theory.
| Chapter | What belongs in it | Common mistake |
|---|---|---|
| Introduction | relevance, aim, structure of the dissertation | giving away the results |
| Theory | concepts and state of research | summarising everything instead of selecting |
| Research question | aims and hypotheses | question stays too broad |
| Method | design, sample, instrument, procedure, analysis | procedure not described repeatably |
| Results | findings without interpretation, tables and figures | interpreting already here |
| Discussion | context, limitations, outlook | limitations of the study missing |
The results chapter reports, the discussion interprets. Blur that line and you give away the exact place where markers recognise methodological care. Formalities such as abstract, contents pages and the declaration of originality are set by your department handbook, alongside the general structure of an academic paper.
How strictly that six-chapter model is applied differs across UK departments. Psychology and education usually keep close to it, while business schools often fold the research question into the introduction, and some courses add a viva in which you defend your methods out loud. Your module handbook and supervisor decide, so read both before you fix the outline.
Create a survey for free
With empirio.ai you can create a modern online survey in minutes — with hosting in the EU.
- AI-built survey
- Adjust by drag & drop
- Real-time analysis
Quantitative or qualitative: which method fits your question?
You work quantitatively when your question can be answered in numbers, so with frequencies, distributions or relationships. You work qualitatively when you want to understand how people interpret, experience or justify something. The question decides, not your preference.
Quantitative studies use standardised instruments across many cases: questionnaires, counts, experiments. What comes out is either a description of what you collected or an inference from the sample to a larger group, depending on how the sample was drawn.
Qualitative studies reverse the ratio: few cases, covered in depth. Semi-structured interviews, focus groups and content analysis suit topics where little theory exists yet or where the relationships refuse to fit into a fixed answer grid.
Four research questions make the difference tangible:
- Quantitative: How does evening screen time relate to how long students sleep?
- Quantitative: How do A-level results differ between regions with high and low education spending?
- Qualitative: How do early-career employees justify turning down home working?
- Qualitative: Which content do parents regard as problematic in their children’s media use?
The two can be combined, for example a survey first and interviews on the striking cases afterwards. A mixed-methods design means two complete rounds of data collection, which is usually too much for an undergraduate dissertation. Our article on qualitative and quantitative research methods compares the individual procedures.
Timetable and lead time: what to allow for
The timetable for an empirical dissertation is built backwards from the submission date. Four dates belong in the calendar: pilot test, start of fieldwork, end of fieldwork and start of analysis. Most plans fall apart between finished questionnaire and first analysis.
The reason is a simple asymmetry. Writing time can be compressed, response time cannot. Fieldwork is the only stage whose length other people decide, and if you are still waiting for responses three weeks before submission, your analysis has no buffer left.
Five items that cost time before data collection
Nearly every broken timetable fails on the same preparatory steps. All five sit before fieldwork, get overlooked in planning and cannot be made up afterwards.
- Ethics approval from your research ethics committee, often weeks ahead
- Gatekeeper permission from a school, trust or employer for access to the field
- Informed consent and data protection under UK GDPR and the Data Protection Act 2018
- A pilot test with five to ten people and the rewrite that follows
- Building a distribution list and finding participants
Your plan B for a weak response rate
Too few responses is the classic emergency. Keep three ways out open: one extra week of fieldwork, a second distribution channel, or a switch to a qualitative design with a handful of interviews. Agree that route with your supervisor while you still do not need it.
Tip
Put the start and end of fieldwork in your calendar before you create the questionnaire. A date that exists only in your head slips quietly, and every slip comes out of your analysis time.
Common mistakes in an empirical dissertation
Most difficulties in an empirical dissertation arise before any writing happens, during planning. Four patterns come up so often that every supervisor has seen them before.
All four share one thing: a decision taken early and questioned late. Knowing the patterns means noticing in time which track your own timetable is heading down.
The questionnaire comes before the hypothesis
Writing questions feels productive, so almost everyone starts there. As long as the hypothesis is missing, though, there is no criterion for which question is needed. The result is thirty items, half of which contribute nothing. Write the hypothesis first and derive every item from it.
Fieldwork without an analysis plan
Send out the link without knowing how the answers will be analysed and you collect data that cannot answer your research question. A five-point rating scale allows different calculations than a yes-no item. Decide before fieldwork which procedure you will run and which level of measurement it requires.
The sample from your own circle
Thirty responses from your own year group are quickly gathered and support no generalisation. A convenience sample is a legitimate route in, but it belongs openly in the methods chapter and in the limitations. Hidden, it becomes a line of attack. Named, it shows methodological awareness.
Interpretation in the results chapter
The results chapter reports findings, explanations come later. Explain there why a difference occurs and you strip the discussion of its substance and make the work look careless. Show first what the data say, interpret afterwards.
Warning
A statistically significant result does not prove your hypothesis. Significance only says this: if there were no relationship in reality, a finding this strong or stronger would be rare. Write cautiously, because the data support the hypothesis, they do not prove it.
Further reading
Four titles go further than this article can. Two of them walk through a study step by step, one serves as a reference work for methods and analysis, and one helps with the writing itself.
- Raithel, Jürgen (2008): Quantitative Forschung. Ein Praxiskurs. 2nd edition, Wiesbaden: VS Verlag für Sozialwissenschaften.
- Beller, Sieghard (2016): Empirisch forschen lernen. Konzepte, Methoden, Fallbeispiele, Tipps. 3rd edition, Göttingen: Hogrefe.
- Döring, Nicola (2023): Forschungsmethoden und Evaluation in den Sozial- und Humanwissenschaften. 6th edition, Berlin and Heidelberg: Springer.
- Hunziker, Alexander W. (2025): Spass am wissenschaftlichen Arbeiten. Zurich: Verlag SKV.
Conclusion
An empirical dissertation is not a bigger project than a theoretical one, the work is simply distributed differently. It piles up at the front, in design and fieldwork, and whoever works carefully there has far less to repair during the writing. Decide early, count backwards and state the limits of your study openly.
What makes a dissertation empirical is not the amount of data, but the route to it.
Where to go next
- Ready to set up the survey for your dissertation? Survey for a bachelor or master thesis
- Wondering why an online survey counts as a method? The online survey as an empirical research method
- Data collected and analysis ahead? Descriptive and inferential statistics
Do you need data for your empirical dissertation?
With empirio.ai, an online survey tool from Germany, you create the questionnaire for your dissertation free of charge and analyse the responses with AI. Your data stays stored in the EU.