03 · What You Need to Know
Research Purpose and Research Design Answer Different Questions
Purpose tells us what the research is trying to accomplish
Terms such as exploratory, descriptive, and explanatory commonly identify the goal of an investigation.
An exploratory study seeks to develop understanding where important aspects of a phenomenon remain insufficiently known. A descriptive study seeks to characterize what exists or occurs. An explanatory study seeks to account for how or why a phenomenon occurs. Evaluative inquiry assesses a program, policy, intervention, service, practice, or other evaluand to support learning, judgment, improvement, or decisions.
These distinctions are valuable because they clarify the intellectual task embedded in the research question.
They do not necessarily tell us how that task will be accomplished.
Design tells us how the study is structured to produce an answer
A research design concerns the structure and logic through which evidence is generated and interpreted.
Depending on the research question and tradition, design features may include whether observations occur once or repeatedly, whether an intervention is manipulated, whether participants are randomly assigned, whether comparison groups are used, how cases are selected, what the unit of analysis is, how temporal order is established, and how the overall structure addresses important threats to the intended conclusion.
Consider the statement:
“This study uses a descriptive research design.”
What do you know?
You know that the study probably aims to characterize some phenomenon. But you may still not know whether it is cross-sectional or longitudinal, whether it uses survey data or observations, how participants are sampled, whether the description concerns a population or a bounded case, what the unit of analysis is, or how measurements are made.
The label communicates purpose, but it may underspecify design.
A purpose can be achieved through several designs
One reason to distinguish purpose from design is that the same purpose can be pursued through different study structures.
Suppose the purpose is descriptive: researchers want to characterize how university students use generative AI for academic work.
| Possible structure |
What it could describe |
Important design feature |
| Cross-sectional survey |
Reported patterns of AI use during a defined period |
Observations collected from a sample at one period or point in time |
| Longitudinal cohort |
How AI-use patterns change across semesters |
Repeated observations of the same or defined cohort over time |
| Qualitative case study |
Detailed practices and contexts of AI use within a bounded educational setting |
In-depth investigation of a bounded case using contextually relevant evidence |
| Observational study using platform records |
Recorded patterns of actual platform use |
Behavior characterized through naturally occurring records rather than self-report alone |
All could contribute to description, but they are not methodologically interchangeable.
The same principle applies to exploration and explanation. Exploratory work may use qualitative inquiry, pilot surveys, case studies, document analysis, or exploratory quantitative analyses. Explanatory work may use experiments, quasi-experiments, longitudinal observational designs, comparative cases, qualitative process analysis, or mixed methods, depending on the kind of explanation sought.
A design can serve more than one purpose
The relationship also works in the other direction.
A longitudinal cohort study might primarily describe how an outcome changes over time. Another longitudinal cohort study might examine temporal relationships in pursuit of an explanatory question. A case study may be exploratory, descriptive, explanatory, evaluative, or some combination depending on its questions and analytic purpose.
Calling something a “case study” therefore does not tell you its purpose. Calling something “descriptive” does not fully tell you its design.
Research purpose
What the study is trying to accomplish: for example, exploration, description, explanation, or evaluation.
Research design
How the investigation is structured so that the resulting evidence can answer the research question and support the intended claims.
The two should align, but alignment does not make them synonymous.
Why do so many sources call them research designs?
Because research design is used at different levels of abstraction.
Some authors use the term broadly to mean the overall plan or strategy for conducting a study. Under that broad usage, describing a project as exploratory or descriptive may legitimately be treated as specifying its design orientation.
Other methodological treatments distinguish research goals from study structures more explicitly. In these frameworks, exploration, description, and explanation characterize the goals of inquiry, while terms such as experimental, quasi-experimental, cross-sectional, longitudinal, cohort, case-control, case study, ethnographic, or other discipline-specific labels communicate different aspects of design.
Neither vocabulary can simply be imposed on every discipline.
The problem arises only when terminology creates false precision. If “descriptive research design” is the only information provided about how a study is organized, readers may know the intended purpose without knowing enough to evaluate how the evidence was generated.
“Descriptive” is especially easy to mistake for a complete design
The phrase descriptive research design is widespread, particularly in student research. It can be useful shorthand, but it often becomes a stopping point.
Imagine two theses that both state:
“The study employed a descriptive research design.”
One uses a probability sample of teachers and a one-time questionnaire to estimate the prevalence of particular practices. The other conducts repeated classroom observations across an academic year to document how those practices change.
Both are descriptive in purpose. Their design structures are substantially different.
If the researcher stops at the shared label, information that matters for interpretation disappears.
“Explanatory design” can hide even more consequential differences
The problem becomes particularly important when a study claims explanation.
Suppose two studies both call themselves explanatory. One randomly assigns an intervention. The other measures naturally occurring variables once and uses regression to examine their associations.
Those studies do not provide equivalent evidence for causal explanation merely because both use the same purpose label.
The design determines what alternative explanations can be addressed, whether temporal order can be established, and how strongly a causal interpretation can be defended.
Watch Out
Calling a study “explanatory” does not itself justify causal language. The strength of an explanatory claim depends on the evidence and design used to support that particular explanation, not on the label attached to the study.
Evaluation makes the distinction particularly visible
Evaluation provides another useful example because an evaluation can use many different designs.
The CDC's current Program Evaluation Framework explicitly separates evaluation purpose and questions from the selection of an overarching evaluation design. It notes that design selection depends on the evaluation purpose, questions, context, resources, and constraints, and discusses experimental, quasi-experimental, and observational design options.
An evaluation may therefore be experimental when estimating an intervention effect, quasi-experimental when random assignment is unavailable, or observational when investigating implementation, fidelity, efficiency, or other noncausal questions.
Calling something “evaluative” tells us why the inquiry is being undertaken. It does not by itself specify the structure through which the evaluation questions will be answered.
Purpose is also different from approach
A second distinction helps prevent another common mix-up.
Exploratory and descriptive describe purposes. Quantitative, qualitative, and mixed methods are broader approaches to evidence and inquiry. Terms such as cross-sectional, longitudinal, experimental, and various tradition-specific labels describe other dimensions of study design.
These dimensions can be combined.
You might conduct a descriptive quantitative cross-sectional study. You might conduct an exploratory qualitative case study. You might conduct an explanatory mixed methods study with a quasi-experimental quantitative component and qualitative process inquiry.
Each adjective answers a somewhat different methodological question.
This is why the choice among quantitative, qualitative, and mixed methods approaches should not be treated as identical to choosing the purpose or complete design of the study.
You do not need to pile every possible label into one sentence
Once researchers recognize that a study can be characterized along several dimensions, another temptation appears: methodological label stacking.
A sentence begins to resemble this:
“The study employed a quantitative, nonexperimental, cross-sectional, descriptive-correlational survey research design.”
Some of those terms may communicate useful information. Others may be redundant, ambiguous, discipline-specific, or better explained in ordinary prose.
The goal is not to create the longest possible methodological title. The goal is to communicate the design accurately enough for readers to understand how the evidence was generated and what claims it can support.
The appropriate level of specificity when naming a research design depends on disciplinary convention and on whether each term adds meaningful information.
A research purpose should influence the design
Although purpose and design are distinct, they should not be disconnected.
If your purpose is descriptive, the design must provide evidence capable of producing the description you claim. If you want population prevalence, sampling and measurement become central. If you want change over time, repeated observations may be necessary.
If your purpose is explanatory, the design must address the particular explanatory claim. Causal explanation may require a credible counterfactual or another defensible identification strategy. Mechanistic explanation may require evidence capable of revealing the relevant process.
If your purpose is evaluative, the design should follow the evaluation questions, intended uses, context, and standards relevant to the evaluation.
This is the broader logic behind matching a research design to the research question.
Terminological flexibility does not mean methodological vagueness
You do not need to correct every author who uses the phrase descriptive research design. Nor should you assume that a thesis is methodologically flawed merely because its terminology differs from your preferred framework.
Instead, ask what the label communicates and what information remains missing.
If a discipline conventionally recognizes a particular term as a design, use it when appropriate. Then provide the structural details needed to understand the study.
Methodological precision should clarify research, not turn vocabulary into a border-control exercise.