When Does One Study Become a Research Program?

What If the Research Problem Is Bigger Than Any One Study?

Some research questions do not end neatly when the first study is completed. One finding raises a question about mechanism. That question suggests an intervention. The intervention needs testing in another population. Results then raise questions about implementation, long-term effects, or differences across contexts.

At what point are you no longer expanding one study but building a research program?

The boundary is not defined by a particular number of questions, papers, years, or collaborators. A research program emerges when a larger scientific problem is pursued systematically through multiple connected studies, each addressing a defensible part of the broader agenda.

A Research Program Connects Multiple Studies Around a Larger Problem

A Research Program Is Larger Than a Study but More Focused Than a Topic

The term research program, also written research programme, is used in several ways across academia. Universities and funders may use it administratively for coordinated portfolios of projects. Researchers may use it more informally to describe a sustained line of inquiry. In philosophy of science, Imre Lakatos used the more specific concept of a scientific research programme to describe sequences of theories organized around a shared theoretical core.

Those meanings should not be treated as identical. For practical research planning, a useful working definition is simpler: a research program is a connected body of studies organized around a larger scientific problem or agenda that cannot be answered adequately by one investigation.

A Topic Is Not Yet a Research Program

"Generative AI in education" is a topic. "Student mental health" is a topic. "Climate adaptation" is a topic. Publishing several studies that happen to contain the same topic does not automatically make them a research program.

A program needs greater intellectual organization. There should be some broader problem, explanatory aim, theoretical proposition, practical challenge, or cumulative set of questions connecting the individual studies.

For example, a researcher might pursue the broader problem of how university students develop the capacity to use generative AI critically and responsibly. That agenda could generate studies of AI literacy measurement, relationships between literacy and evaluation behavior, educational interventions, transfer across academic tasks, and longer-term outcomes.

The individual projects differ, but their relationship is intelligible.

The Research Questions No Longer Fit One Design

One of the clearest signs that a research program is emerging is methodological divergence.

Your first question may require a cross-sectional survey. The next concerns change and requires longitudinal evidence. A third asks whether an intervention works and therefore requires an experimental or quasi-experimental design. A fourth concerns participants' experiences and may require qualitative inquiry.

You could theoretically attempt to place all of these components into an enormous multimethod project. Sometimes that is justified. Often it is cleaner to recognize that different questions require different studies.

This is the same logic behind deciding when research questions should become separate studies. A research program provides the larger structure that keeps those studies intellectually connected.

Later Questions Depend on What Earlier Studies Find

Another strong indicator is sequential dependence. You cannot design the later question properly until you know what the earlier study reveals.

Suppose Study 1 identifies several factors associated with students' critical use of generative AI. Study 2 investigates which of those factors appear to explain differences in behavior. Those findings inform the design of an intervention evaluated in Study 3. Study 4 then examines implementation in authentic courses.

The sequence matters. Attempting to specify every question in Study 1 would require researchers to anticipate findings they do not yet have.

Research programs allow inquiry to remain cumulative and adaptive. Each study can reduce uncertainty before the next major methodological commitment is made.

Different Studies Can Test Different Parts of the Same Explanation

A research program need not follow a simple linear sequence. Several studies may examine different implications of a broader explanation.

One project might test whether a relationship appears in one population. Another could examine whether it replicates elsewhere. A third might investigate a proposed mechanism. Another might examine boundary conditions or alternative explanations.

This logic resembles the philosophical idea that scientific theories are rarely evaluated through one isolated observation. Lakatos's account of scientific research programmes, although more specific and theoretical than the practical usage here, likewise emphasized evaluating connected sequences of theoretical and empirical developments rather than treating theories as isolated tests.

For ordinary project planning, you do not need to adopt Lakatos's philosophical framework to recognize the useful principle: stronger scientific understanding often emerges across related studies rather than from expecting one study to settle an entire problem.

Replication Can Be Part of the Program, Not a Detour From It

A research program should not be understood as a staircase in which every new study must introduce another variable or increasingly elaborate model. Replication can be a substantive step.

If an early study identifies an important relationship, repeating the investigation in another sample, setting, measurement context, or research team may test whether the finding is robust. A program concerned only with producing novel extensions can accumulate complexity faster than confidence.

Cumulative research sometimes advances by asking, "Does this still hold?" rather than always asking, "What else can we add?"

One Dataset Can Support Several Projects Within the Program

A research program does not require new data collection for every study. Large cohorts, longitudinal datasets, registries, institutional databases, or recurring surveys may support multiple investigations.

Several projects can share the same dataset while addressing different research questions. What makes them part of a program is not dataset reuse itself but their contribution to the larger scientific agenda.

Likewise, the same participants may contribute to different studies within a longitudinal or multi-project program, subject to appropriate consent, ethics, governance, and methodological requirements.

A Research Program Does Not Require One Researcher to Do Everything

As research expands, expertise requirements often expand with it. Later studies may require statisticians, qualitative researchers, domain specialists, implementation scientists, software developers, laboratory expertise, community partners, or researchers from other disciplines.

A research program can therefore become collaborative without losing coherence. In fact, the larger problem may be better served when different investigators lead studies suited to their expertise.

The intellectual connection among projects matters more than whether every paper has the same first author.

Programmatic Research Can Protect Individual Studies From Scope Creep

Researchers sometimes overload a study because they fear that excluding a question means losing it. A programmatic perspective changes that calculation.

The current study does not need to answer every question if there is a credible path for subsequent work. A mechanism can wait until the phenomenon is established. An intervention can wait until the target is better understood. An implementation study can wait until there is something sufficiently developed to implement.

This perspective can help researchers limit the number of secondary questions in a single study without reducing their broader scientific ambition.

Not Every Sequence of Papers Is a Research Program

Publishing repeatedly on the same dataset, population, or fashionable topic does not by itself establish programmatic research.

Ask what changes in knowledge from one study to the next. Does a later study address uncertainty created by earlier findings? Does it test an explanation, alternative, intervention, replication, boundary condition, or consequence? Can you explain why these projects belong together without relying only on a shared keyword?

A Research Program Should Be Able to Change

A program is not a commitment to prove an idea repeatedly. Findings should be able to alter the direction of subsequent work.

If early evidence contradicts the proposed explanation, later studies should investigate alternatives, revise assumptions, or perhaps abandon an unproductive line of inquiry. A program that treats every finding as support for its original position risks becoming advocacy rather than cumulative research.

This is another reason to distinguish a research program from a predetermined publication plan. A publication plan says what researchers intend to produce. A research program should remain responsive to what the evidence actually shows.

Programs Can Branch Without Losing Their Core

As knowledge develops, one line of inquiry may generate several branches. An educational technology program might begin with student learning, then develop connected streams concerning teacher implementation, assessment, equity, and institutional adoption.

At some point, those branches may themselves become distinct programs. There is no formal threshold. The practical question is whether a common scientific problem still provides enough intellectual structure to make the studies meaningfully cumulative.

A Dissertation Can Sometimes Contain the Beginning of a Research Program

Doctoral research occasionally contains several linked studies rather than one oversized investigation. For example, a dissertation might develop a measure, conduct an observational study, and evaluate an intervention.

Whether that structure is appropriate depends on the research problem and institutional requirements. When several linked investigations are necessary, it may be useful to consider whether a dissertation should contain several linked studies rather than one oversized study.

The dissertation may complete a coherent doctoral project while also establishing questions that continue afterward as a broader research program.

From One AI Literacy Study to a Program of Research

The first study did not become larger each time a new question appeared. Instead, the broader problem generated a sequence of designs, each suited to a particular uncertainty. Together, those studies form a more credible research program than an initial project attempting to measure associations, mechanisms, intervention effects, implementation, and long-term transfer simultaneously.

Common Misunderstandings About Research Programs

Think in Studies When the Problem Requires Different Kinds of Evidence

If your research plan keeps growing, do not immediately ask how to fit everything into one protocol. Ask whether the larger problem would be investigated more convincingly through a sequence of studies.

A useful exercise is to draw the logic of the program as questions rather than papers. Start with the larger problem. Identify what must be established first, what uncertainties remain after that, and what kind of evidence would resolve each uncertainty. The resulting sequence may look quite different from a list of manuscripts you hope to publish.

That difference matters. A research program is fundamentally about accumulating knowledge, even though publications are one of the ways that knowledge is communicated.

Before Calling the Work a Research Program, Check the Connections

Questions About Building a Research Program

Some Questions Belong Together Across Studies, Not Inside One Study

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.