Types of Research Design: Examples and How to Choose for Your Study

What is a research design?

Research design is the overall strategy linking a research question to the collection, measurement and interpretation of evidence. Choose it by asking what you need to learn, what evidence is available and what conclusions that evidence can support. A questionnaire is a data-collection tool, not a complete design. Software is also not a design.

There is no single universal list of design types. Some labels describe the purpose, some describe timing, and others describe whether a researcher intervenes. A study can therefore be both descriptive and cross-sectional, for example. Explain the labels you use rather than treating them as mutually exclusive boxes.

Descriptive and correlational designs

Descriptive research characterises a population, situation or pattern. An illustrative question is: what library services do students report using this semester? Correlational research examines relationships between measured variables, such as whether reported study time is associated with examination scores. An association alone does not establish that one variable caused the other: other factors, measurement issues or reverse direction may explain it.

Sampling and measurement remain central. A large convenience survey does not automatically represent a whole university, and a poorly defined measure cannot be rescued simply by collecting more responses.

Experimental and quasi-experimental designs

An experiment involves an intervention and comparison arrangements; random allocation, where appropriate and feasible, helps address differences between groups. A quasi-experimental design evaluates an intervention without full random allocation and needs careful attention to alternative explanations. For example, comparing classes before and after a teaching change may reflect other changes over time, not only the intervention.

Do not describe an observational survey as an experiment merely because it tests a hypothesis. Any intervention involving people must meet the applicable ethical and institutional requirements. Design and approval come before recruitment or implementation.

Qualitative, case-study and mixed-methods approaches

Qualitative designs can investigate meaning, experience and processes through interviews, observations or texts. A case study investigates a bounded case in its context and may use more than one source of evidence. Explain why the case and sources are appropriate; studying one organisation does not automatically make every project a well-designed case study.

Mixed-methods research combines qualitative and quantitative work in a planned way. Its value depends on how the strands inform each other, not simply on including an interview and a survey. State where integration occurs: in sampling, instrument development, analysis or interpretation.

Cross-sectional vs longitudinal: a separate choice

Cross-sectional research examines evidence from one time point or defined period; longitudinal work follows change across time. Repeated measurement may help examine trajectories, but brings practical issues such as participant loss and consistent measurement. Timing does not by itself determine whether a study is qualitative, quantitative or causal.

Before choosing, write one sentence describing the intended conclusion. If it says that an intervention caused a change, check whether the comparison, allocation and timing support that claim. If access only permits a single survey, narrow the question and conclusions instead of overstating what the data can show.

A design-selection checklist for your own proposal

Discuss the design with your supervisor before collecting evidence. Keep a short record of the alternatives you considered and why the chosen strategy is feasible. Academic Compass can offer research guidance and edit your own explanation, but cannot conduct or write assessed research for you.

• Match each question to the evidence needed to answer it.
• Check access, recruitment, permissions, ethics and time constraints.
• Specify how evidence will be analysed and interpreted.
• Identify bias, alternative explanations and limits on generalisation.
• Confirm that the planned claim is no stronger than the design supports.