Research methods for a bachelor’s or master’s thesis: examples and how to choose

Choose research methods for a bachelor’s or master’s thesis according to your question, the evidence you can access and the conclusions you want to draw. A questionnaire is one option. Interviews, documents, statistical datasets, a bounded case or academic publications may be more suitable. Decide what evidence would answer your question before choosing how to collect and analyse it.

This guide takes you from a broad topic to a feasible research design. Its worked examples are original, hypothetical teaching scenarios, not findings from completed studies. Your programme’s requirements and the design agreed with your supervisor take priority over a generic online template. References to Polish university rules below concern those specific institutions.

How do you choose a research method for your thesis?

Start with one sentence: “I want to establish…, using…, by comparing or analysing…”. If you cannot identify the evidence, the design is not ready. A topic such as employee motivation does not tell you whether you need people’s accounts, organisational records, observations of behaviour or a comparison of theories.

1. Question: what exactly should the thesis establish? Distinguish describing a phenomenon from examining an association or evaluating an intervention.
2. Evidence: who or what can provide the information? Check access, time coverage, quality and whether you may use it.
3. Analysis: what will you do with the material? Plan the categories, comparisons or calculations before collecting it.
4. Limits: which questions will this design leave unanswered? Narrow the question if the evidence cannot support the broader conclusion.

Do not start with a page count or a software package. Excel, SPSS and an online form are technical tools, not a justification for the research design. If you need to organise these decisions before writing, consider the service “Help with your thesis outline and structure”.

Method, data collection technique and research instrument: what is the difference?

Polish methodology textbooks do not always classify these terms in exactly the same way. Zdzisław Szkutnik discusses both their distinction and differences between classifications. Use a framework appropriate to your discipline, cite it and apply it consistently. Do not combine several textbook lists without checking whether their definitions agree.

For documentary research, a coding sheet performs a comparable organising role: it may record document type, date, content category and the rule for assigning a code. A spreadsheet alone is not an analysis. You need defensible categories and instructions for ambiguous cases.

What research methods are available? Match them to the question

No short list covers every discipline. Social research commonly uses surveys, interviews, observation, documentary analysis and case studies. Experiments, modelling, legal analysis or historical investigation may be central elsewhere. The routes below help select evidence for common social science, public administration and business thesis questions.

• How many, how often, associated with what? Consider quantitative research using suitable data. Define the variables and the comparisons before counting.
• How do people understand a situation or describe an experience? Consider interviews and qualitative analysis. Explain how you will move from accounts to categories.
• What was recorded, published or done during a defined period? Consider documents, secondary data or observation, depending on whether the object is a record, an outcome or behaviour.
• How does a process work in a particular organisation? Consider a case study with explicit boundaries and several appropriate sources of evidence.

Qualitative or quantitative research: which should you choose?

Quantitative research can describe the extent of a phenomenon or examine relationships between variables. Qualitative research can explore meanings, experiences and processes in depth. Neither the belief that numbers are automatically more scientific nor the hope that a few interviews will be easier than a survey should decide the design. Both require explicit selection and analysis procedures.

“What proportion of participants find their supervisor’s comments clear?” needs different evidence from “How do students interpret ambiguous supervisory feedback?”. You may combine approaches, but explain the contribution of each and where their results will be integrated. Two unrelated sections in the same document do not automatically form a coherent mixed methods study.

When is a questionnaire useful, and what can it not answer?

A questionnaire is useful when you need comparable responses to a defined set of questions, for example about reported experiences or service ratings. Use clear concepts, suitable response options and a recruitment route to the relevant people. Map each research question to an indicator, an item and a planned analysis. Remove items that do not help answer the research question.

“I use the app because it saves time” is a participant’s account, not a measurement of minutes saved. A one-off questionnaire on motivation and performance does not by itself establish that motivation caused better performance. Nor does a large number of responses to an open link repair an unknown sampling process. An attractive chart does not remove these limitations.

If this route fits your objective, the service “Questionnaire design for your thesis” may help. Describing a form does not replace a research design, and a design does not replace responses that were actually collected.

Interviews and observation: what evidence do they provide?

An interview can help you understand how participants describe a decision, difficulty or change. Specify recruitment criteria, an interview guide, recording arrangements and an analysis procedure. A useful prompt invites a concrete experience: “Please describe the last occasion when…”. “Do you agree that the system is ineffective?” supplies the judgement and makes the answer harder to interpret.

Observation concerns what can be observed under defined conditions, not everything a person thinks. In a service encounter you might record stages, waiting time or predefined behaviours. Do not infer an employee’s motives from a gesture alone. Define the observation unit, recording rules and the possible effect of the observer. Before research with people, establish the required permissions, participant information and data protection arrangements.

Document analysis and secondary research instead of a new survey

Secondary data already exist, usually because they were created for another purpose. They may include official statistics, reports, institutional records or published materials. Your contribution lies in the question, a defensible selection and your analysis. Copying a report’s conclusions is not an independent investigation. The Polish guide edited by Marta Makowska covers desk research, content analysis and statistical analysis of existing data.

Example: what can Polish public information records establish?

Suppose you compare how selected municipalities announce public consultations. Define the period, document types and search procedure in their Public Information Bulletin, known as BIP. Your coding sheet might record the announcement channel, deadline, availability of a comments form and publication of outcomes. This can assess the documented communication procedure. It cannot automatically establish that residents read the notices or felt heard.

Keep source URLs, retrieval dates and exclusion criteria. Failure to find a document in the location you checked does not always mean it never existed. Distinguish a finding about the material’s availability from a finding about what the institution did.

Example: comparing municipalities using Statistics Poland data

In Statistics Poland’s Local Data Bank, check the variable definition, measurement unit, territorial level, years and data attributes. The API documentation treats these as distinct resources. Distinguish a count from a per-capita indicator and a year-end stock from a value covering the whole period. Preserve the downloaded version because the database can be updated.

If spending and population both change in two municipalities, the comparison alone does not demonstrate that spending caused migration. Other factors may be missing, and area-level data do not describe each person’s motives. Use the evidence for a descriptive conclusion or a carefully framed question for further research rather than adding a causal explanation.

A thesis case study is more than a description of one company

A case study needs an explicit case and boundaries. It might examine implementation of a particular procedure in one organisation during a defined period. A company history and description of its products do not yet answer a research question. You need evidence about the process and rules for comparing that evidence.

Teaching example: you examine a change in complaints handling. A procedure states what should happen; interviews reveal how employees understand the change; anonymised case records may show what actually happened. If sources disagree, analyse the disagreement. Do not select only the evidence that supports your initial view. Findings about this case are not automatically findings about all companies.

Can a thesis use existing evidence without a questionnaire?

Without a questionnaire, often yes. Without independent analysis, that is a different question. Documentary analysis, calculations using secondary data and critical synthesis of publications can involve your own analytical work without collecting new responses. Do not equate original analysis exclusively with a questionnaire, but do not call every compilation of text a research study either.

For example, the University of Łódź’s Faculty of International and Political Studies guidelines from 2022 distinguish theoretical, empirical, combined and review-based theses. The University of Life Sciences in Lublin’s Faculty of Production Engineering guidance effective from 2023 also describes a review-based master’s thesis, requiring a documented critical synthesis and independent conclusions. These are examples of local rules, not permission applicable to every university.

Before choosing a literature-based thesis, check your current programme rules and show your supervisor an analysis plan. Specify the sources, comparison criteria and unresolved question. For the detailed process, use the guide to a thesis literature review.

How do you justify your method? A worked design example

Justify why this evidence answers the question and what an alternative approach would miss. Avoid “a questionnaire was selected because it was easy”. Accessibility matters for feasibility, but it does not establish fitness for the objective. For help aligning these elements, consider “Help with your thesis methodology chapter”.

How many methods and participants do you need?

There is no method count that guarantees quality. One appropriate, carefully conducted procedure can be more informative than several superficial ones. Add another evidence source when it solves a specific problem, such as comparing reported practice with records or exploring a conflicting result. Every extension also adds time and an analysis obligation.

Sample size depends on the objective, recruitment or sampling procedure, planned analysis and variability. For interviews, the scope of the question and diversity of experiences also matter. Do not treat “100 questionnaires” or “five interviews” as universal requirements. Use the guide to sampling and sample size for a more detailed discussion.

What belongs in the methodology chapter?

A reader should be able to follow your procedure: objective and questions, material, selection criteria, data collection, instrument, analysis and limitations. Define methods where doing so clarifies your design. Several pages of textbook definitions cannot replace an account of where the data came from and what you did with them.

For objectives, research questions and hypotheses, use the guide to thesis methodology. This article focuses on choosing an evidence strategy; that guide develops the relationship between your questions and research assumptions.

Common mistakes when choosing thesis research methods

• Choosing the method before the question. Reverse the order: problem, evidence, analysis.
• Evidence does not support the conclusion. Area statistics do not reveal residents’ motives; opinions are not a direct measure of effectiveness.
• No selection rules. Record why these participants, documents, periods or publications belong in the study.
• Leaving analysis until everything is collected. Check feasibility on a small part of the material and document any changes to the procedure.
• Labels without procedures. “Analysis and synthesis” does not explain which sources you compare, against what criteria, or how conclusions are produced.

Checklist before agreeing your research design

1. Does the question specify what you intend to establish?
2. Does the evidence contain the necessary information, and can you access it in time?
3. Are participant or source selection criteria explicit and justified?
4. Does the instrument serve the questions rather than collect unrelated data?
5. Can you explain how the material will become findings and conclusions?
6. Have you distinguished reported accounts, observations, associations and causal claims?
7. Have you established permissions, anonymisation and secure storage where required?
8. Can you state the limitations and confirm alignment with current programme requirements?

If an answer is “I do not know”, record the unresolved decision before writing more. This is usually more useful than adding pages of theory. If the study is already complete, report its real limitations and available evidence. Do not rewrite the history of the research to make it resemble a textbook design.

Polish bibliography and sources for further work

Sources were checked on 28 October 2026. Worked examples and the checklist are editorial teaching material. University documents illustrate differences in local requirements; they do not establish your programme’s rules. For Makowska’s edited book, the bibliographic details and scope were verified through the publisher; no unverified quotations or page-specific claims are attributed to it.

• Szkutnik, Z. (2005). Metodyka pisania pracy dyplomowej. Skrypt dla studentów. Poznań: Wydawnictwo Poznańskie, chapter 3, especially pp. 32-34. Full Polish text in the Silesian Digital Library.
• Makowska, M. (ed.). (2013). Analiza danych zastanych. Przewodnik dla studentów. Warszawa: Wydawnictwo Naukowe Scholar. ISBN 978-83-7383-620-4. Publisher’s description and scope.
• University of Łódź, Faculty of International and Political Studies (2022). Wytyczne w zakresie dyplomowania, section 2.1.2. Thesis types: an example of faculty-specific rules.
• University of Life Sciences in Lublin, Faculty of Production Engineering (2023). Wskazówki redakcyjne dla autorów prac dyplomowych magisterskich, part 1. Research-based and review-based thesis requirements.
• Statistics Poland. Local Data Bank API, documentation of resources, measurement units, variables and data attributes. Official BDL documentation.