You have a topic, you have research questions, and now your supervisor is asking the question that makes many final year students pause: what is your research design? If you are not sure what to say, you are not alone. Most students pick a design the wrong way round. They choose a favourite method first, a questionnaire because their friend used one, or interviews because they seem easier, and then they try to force their research question to fit. This guide does it the other way round. You will start from your research question, match it to the design that can actually answer it, and check that the design is possible within your time, budget and ethics clearance. Follow the steps in order, and by the end you will know exactly which design to name in Chapter Three, and why.
Writing a high-quality Project Thesis requires more than just data; it requires precision. To ensure your Research Methodology meets international standards, we recommend using advanced AI Plagiarism Checkers and Grammar Correction Software before submission.
Many academic institutions now use automated AI Detection Tools. If you are preparing your Statement of Purpose (SOP) or Manuscript, investing in professional Academic Proofreading Services can help you maintain a high Similarity Index score while ensuring your research remains authentic and impactful.
First, a quick definition, because the term itself causes half the confusion. A research design is the overall plan for your study. It says what evidence you will collect, from whom, under what conditions, and how that evidence will answer your research question. It is not the same as your method. A questionnaire is a method, a tool for collecting data. A survey design is the plan that says you will give that questionnaire to a defined group of people at one point in time in order to describe something or test a relationship. The design is the logic. The method is the tool that serves it. If you can explain your logic, the tools follow naturally. For the bigger picture of how design fits into methodology, read our guide on research methodology, what it is.
Everything starts here, and this step is the one students rush. Take your main research question and ask what kind of answer it needs. There are only a few kinds, and each one points to a different design.
Write your question out, then underline the verb. Describe, relate, test, explore, explain. That verb is your first clue. If you have not yet settled your question, fix that before you choose a design. Our guide on how to write research questions and hypotheses for your project will help you get the wording right.
Before you pick a specific design, choose the approach, because it decides your data, your instruments and your analysis.
You work with numbers: scores, counts, percentages, test results. You measure variables and test relationships or differences statistically. Choose it when your question asks how much, how many, or whether two things are related. Typical tools are questionnaires, structured observation and secondary data. Typical analysis is descriptive statistics, correlation or tests of difference.
You work with words, meanings and experiences. You explore why people behave the way they do and what an experience means to them. Choose it when your question asks how or why, or when little is known about the topic. Typical tools are interviews, focus groups, observation and case study. Typical analysis is thematic analysis. If that is your direction, our guide on how to analyse qualitative data in your final year project shows the steps.
You combine both in one study, usually a survey to measure something, then interviews to explain the results. Choose it only when one approach alone cannot answer your question. Mixed methods are powerful, but they take more time, more planning and more skill, because you must analyse two kinds of data and integrate them. Do not choose mixed methods just to impress your supervisor. Choose it because your question genuinely needs both numbers and explanations. Our comparison of qualitative vs quantitative research for final year projects will help you weigh the two sides.
A common mistake is to say your study is mixed methods when it is not. If you used a questionnaire with a few open ended questions at the end, that is still a quantitative study with qualitative items, not a mixed methods design. A real mixed methods study collects and analyses both kinds of data as separate strands and brings them together in the findings.
Here are the six designs Nigerian undergraduates use most, with what each one can and cannot claim. Read the "cannot claim" line carefully, because that is where examiners catch students out.
Best for: describing what exists, how common something is, or what people think, feel or do. You collect data from a sample at one point in time, usually with a questionnaire, and you report frequencies, percentages and means.
Example: "A survey of 300 undergraduates to describe their awareness of mental health services on campus."
Cannot claim: causation, or why something happens. A survey can tell you that 70 percent of students are stressed. It cannot tell you why they are stressed, or whether one thing caused the stress. For that, you need a different design, or follow-up interviews.
Best for: testing whether two or more variables are related, and how strongly. You measure both variables and calculate a correlation coefficient.
Example: "The relationship between hours spent on social media and GPA among undergraduates."
Cannot claim: causation. This is the rule every student must memorise: correlation is not causation. If heavy social media users have lower GPAs, it may be that social media hurts grades, or that struggling students spend more time online, or that a third factor, such as poor time management, causes both. A correlational design can only report the association. Say so plainly in your project, and your examiner will trust you more, not less.
Best for: testing whether one thing causes a change in another. You introduce a treatment, such as a new teaching method, to one group, keep another group as a control, and compare the results. A true experiment assigns people to groups randomly. A quasi-experiment uses existing groups, such as two intact classes, because random assignment is not possible in a real school.
Example: "Two SS2 classes are taught Mathematics, one with group discussion and one by lecture, and their end of term scores are compared."
Cannot claim: a clean causal result if the groups were not equal to start with. If the discussion class was already stronger in Mathematics, you cannot be sure the method caused the difference. Control for that by comparing pre-test scores, or by using the students' previous term scores as a control variable.
Best for: a deep investigation of one person, group, organisation or event. You collect several kinds of evidence about that single case: documents, interviews, observation, records.
Example: "A case study of how one small business in Onitsha survived the fuel subsidy removal."
Cannot claim: that the findings apply to every similar case. One business is not all businesses. What a case study offers instead is depth and detail, and a rich account that other researchers can learn from. Say that honestly, and do not generalise beyond your case.
Best for: topics where little is known, or where you need to understand experiences and meanings. You interview or observe a small number of people and build themes from what they say.
Example: "How final year nursing students experience their first clinical placement."
Cannot claim: that the findings represent a whole population. Qualitative work trades breadth for depth. Its strength is understanding, not counting, and your Chapter Four should present themes and quotes, not percentages dressed up as statistics.
Best for: questions that need both numbers and explanations. The commonest pattern for undergraduates is explanatory: a survey first, then interviews with a few respondents to explain the survey results.
Example: "A survey of 250 students measures the extent of exam malpractice, then interviews with 12 students explore why they say they do it."
Cannot claim: anything extra for free. Mixed methods cost more time and more pages. You must report both strands properly and show how they connect. A survey plus three short interviews tacked on at the end is not a mixed methods study.
This matching logic is the heart of choosing a design. If you want a second opinion on the same approach, this guide on choosing the right research design walks through the same steps, and this guide on matching your question to a design explains the reasoning in more depth.
A design that fits your question perfectly is useless if you cannot carry it out. Run these four checks before you commit.
If your ideal design fails one of these checks, do not quietly change your question to fit a weaker design and pretend nothing happened. Either scale the design honestly, for example by using two intact classes instead of random assignment, and say so, or narrow the question to something the design you can afford can actually answer.
Once the design is chosen, Chapter Three must state it clearly. Your methodology chapter, which most Nigerian universities call Chapter Three, should contain these lines, in this order.
Our guide on how to write a strong Chapter Three on methodology shows where each of these pieces sits in the chapter. And for the full range of design types with worked examples, this guide on writing a research design with examples is a useful second reference.
Here are three complete chains, from question to design, so you can see the reasoning in action.
Example 1: social media and academic performance. The question is whether social media use is related to GPA. That asks about a relationship, so the design is correlational, using a quantitative approach. The instrument is a questionnaire measuring daily social media hours and a record of GPA. The analysis is a correlation coefficient. What it cannot claim is that social media causes low grades, and the project must say so.
Example 2: a new teaching method. The question is whether group discussion improves Mathematics scores. That asks about cause, so the design is quasi-experimental, because the researcher must use two existing classes rather than randomly assigning students. One class is taught with group discussion, the other by lecture, and both sit the same test. The analysis compares the mean scores and controls for previous term scores. What it cannot claim is a perfectly clean causal result, because the classes may have differed before the study began.
Example 3: why students drop out. The question is why final year students abandon their projects. Little is known, and the answer needs personal accounts, so the design is exploratory qualitative, using semi-structured interviews with 15 students who dropped out. The analysis is thematic. What it cannot claim is that the findings describe all students, but it can explain the reasons in depth, and a future survey could test how common each reason is. That suggestion, for further study, is exactly what a good Chapter Five recommendation looks like.
Answer these honestly, and you are ready.
Choosing a research design is not about finding the most impressive label. It is about finding the plan that can produce the evidence your question needs, within the time and access you actually have. Start from the question, match it to a design, check that the design is possible, and write the choice down in Chapter Three with its justification and its limits. Do that, and the design section stops being the part of your methodology you dread, and becomes the part that makes the rest of the project hold together. When your data comes back, our guide on how to interpret your research findings will help you make sense of what the design actually produced.