Explainer · Plain-language

What Is a Research Design? Types & How to Choose | CASRAI

A research design is the overall plan specifying how a study will be conducted — defining the research questions, selecting methods, identifying data sources, and establishing validity strategies. It is the logical blueprint that connects the research question to the evidence collected.

CASRAI plain-language explainers — clear answers to recurring research-administration questions

The step most authors miss

Doing CRediT right? Don’t stop at the statement.

A CRediT statement credits you inside one paper. The recognition CRediT was built for happens when those roles are tied to you, persistently. Sign in with your ORCID — free — and claim your CRediT contributions on casrai.org, the home of the standard. They become a verified, portable part of your identity, not a line that disappears into one PDF.

Free: claim your contributions, then export a journal-ready CRediT statement, schema.org structured data, JATS XML, CSV or BibTeX — and preview your public profile. A membership publishes that profile publicly and verifies the journals you serve.

What a research design is and why it matters

John Creswell’s Research Design (5th ed., 2022) defines research design as the plan and procedures spanning decisions about worldview, strategies of inquiry, and methods. A research design is not a synonym for method (the tool) or for methodology (the philosophical justification of the tool). It is the connector: it specifies the logic by which the research question will be addressed, including how variables will be operationalised, what comparison will be made, when and from whom data will be collected, how threats to validity will be handled, and how findings will be interpreted. A poorly specified design produces data that cannot answer the research question no matter how meticulously the data are collected.

Experimental, quasi-experimental and non-experimental designs

The three broadest design categories differ in the degree of control the researcher exerts. Experimental designs (including randomised controlled trials) involve random assignment of participants to conditions, direct manipulation of an independent variable, and a control condition, enabling causal inference. Quasi-experimental designs involve manipulation of an independent variable but no random assignment — using naturally occurring comparison groups, interrupted time series, regression-discontinuity designs, or difference-in-differences approaches. Non-experimental (observational) designs do not manipulate any variable; they include cross-sectional surveys (data collected at one time point), longitudinal studies and cohort designs (following groups over time), case-control studies (comparing cases with matched controls retrospectively), and ecological or correlational studies.

Descriptive, explanatory and exploratory purposes

Research designs can also be categorised by their purpose. Descriptive designs document what exists — the prevalence of a phenomenon, the characteristics of a population, or the state of a system. Explanatory designs test causal hypotheses — why does X cause Y? Exploratory designs investigate an area about which little is known, often to generate hypotheses for later confirmatory work. A single study can combine purposes: a longitudinal cohort study might be primarily descriptive (tracking disease incidence) while also testing explanatory hypotheses about risk factors. The research question drives the purpose; the purpose drives the appropriate design.

Internal vs external validity, fixed vs flexible designs

A fundamental tension in research design is between internal validity — the degree to which causal conclusions are justified — and external validity (generalisability) — the degree to which findings apply to other populations, settings, or times. Experimental designs maximise internal validity through control, often at the cost of external validity (artificial lab conditions). Observational designs in natural settings have higher external validity but weaker causal inference. Colin Robson (Real World Research, 2011) distinguishes fixed designs (specified before data collection, quantitative, testing theory) from flexible designs (emergent, qualitative, developing theory) — a distinction that maps roughly onto hypothetico-deductive and inductive research logics.

Key facts

At a glance

Definition: Overall plan connecting research question to methods and evidence
Not the same as: Method (the tool) or methodology (philosophical justification)
Three categories: Experimental, quasi-experimental, non-experimental (observational)
Purposes: Descriptive, explanatory, exploratory
Key tension: Internal validity (causal inference) vs external validity (generalisability)
Robson (2011): Fixed designs (quantitative) vs flexible designs (qualitative)
Principle: Research question drives design — not the reverse

Common misconceptions

What people often get wrong

Often heard: Research design and research methodology are the same thing.

Actually: No — methodology is the philosophical framework justifying method choices (e.g., positivism justifying experimental methods); research design is the specific plan for a study specifying what will be done and how. Methodology explains why a design is defensible; the design specifies how the study will be conducted.

Often heard: Experimental designs are always superior to observational ones.

Actually: No — the best design is the one that most appropriately addresses the research question. When random assignment is impossible or unethical, a well-designed observational study with appropriate controls is more informative than a poorly executed experiment. External validity concerns also limit the superiority of experimental designs in some contexts.

Often heard: A qualitative study doesn't have a research design.

Actually: No — qualitative studies have research designs; they are just more likely to use flexible designs (Robson) that evolve as understanding develops. Design choices in qualitative research include selection of cases, data-collection methods, analysis approach, and strategies for trustworthiness — all of which constitute a design.

Going deeper