When Should an Interdisciplinary Question Be Divided Into Several Linked Studies?

When Is One Interdisciplinary Study Trying to Do the Work of Several?

You narrow the research question. You remove interesting but unnecessary variables. You identify the disciplinary contributions that genuinely matter. Yet the remaining question still seems to require different populations, methods, datasets, levels of analysis, or stages of investigation.

At that point, the problem may not be poor scoping.

The research question may contain several distinct empirical tasks that depend on one another. Trying to compress them into one study can produce a design in which every component receives insufficient attention. Separating them completely, however, can destroy the interdisciplinary relationship that motivated the project.

The alternative is to treat the problem as a coordinated program of linked studies.

Divide the Question When Its Necessary Parts Need Distinct Empirical Designs

One Research Problem Can Legitimately Require More Than One Study

A research question and a study are not the same thing

Researchers sometimes assume that every important research question should correspond neatly to one empirical study. That is often convenient, but it is not a requirement of scientific reasoning.

A larger question may require several pieces of evidence. One study might establish whether a phenomenon occurs. Another may investigate the mechanism producing it. A third may test whether that mechanism changes under particular conditions.

Interdisciplinary research makes this especially likely because different components of the question can involve different levels of analysis, forms of evidence, methodological traditions, or disciplinary expertise.

The National Academies characterizes interdisciplinary research by the integration of knowledge from multiple disciplines around questions or problems that extend beyond a single discipline. Integration therefore concerns the intellectual relationship among contributions. It does not require every contribution to be generated within one monolithic study.

First distinguish excessive scope from necessary complexity

Before dividing a project, make sure the apparent complexity is not simply caused by an overstuffed question.

If half of the proposed constructs are merely interesting, remove them. If three outcomes can be reduced to the one that matters, do that first. If several disciplines are present but only two are actually necessary, reduce the disciplinary load.

The first response to an unwieldy project should usually be to narrow the interdisciplinary research question.

Linked studies become appropriate when the question remains complex after unnecessary scope has been removed because several empirical tasks are genuinely required.

Different subquestions may require genuinely different methods

One common reason to divide a project is methodological incompatibility.

Suppose an interdisciplinary project asks how characteristics of an AI tutoring system influence students' trust and subsequent learning behavior.

Before examining learner responses, researchers may need to characterize the system's outputs systematically. That could require computational or content-based analysis. Understanding how students interpret those outputs may require experiments, surveys, interviews, observations, or combinations appropriate to the specific question.

Trying to treat all of these as one undifferentiated “study” can obscure what each method is actually designed to establish.

Separate linked studies can allow each empirical question to use a method appropriate to its evidentiary task while preserving the connection among findings.

Different levels of analysis may justify separate studies

Interdisciplinary questions often connect phenomena operating at different levels.

Imagine a research program examining how university AI policies affect students' use of generative AI. One component concerns institutional policy. Another concerns classroom implementation. Another concerns individual student behavior.

These levels may involve different units of analysis, samples, data sources, and theoretical mechanisms.

One study could analyze institutional policies. Another could examine how instructors translate those policies into assessment practices. A third could investigate student behavior under different assessment conditions.

The studies remain linked if the project explicitly investigates the pathway among those levels. Without that integration, they are simply three studies sharing a broad topic.

Sequential dependence is a strong reason to use linked studies

Sometimes one part of the research must logically precede another.

You may need to identify a phenomenon before testing its consequences. Develop or validate a measure before using it in a larger explanatory study. Characterize technological behavior before exposing participants to it. Explore an unfamiliar mechanism qualitatively before designing a quantitative test. Establish an association before testing an intervention intended to change it.

This sequence is only an illustration. Not every multi-study project should follow it. The design should follow the logical dependencies within the particular research question.

Linked studies need an explicit architecture

Several studies become a coherent research program only when their relationship is specified.

For each study, ask:

• Which part of the overarching question does this study answer?
• Why can that part not be answered adequately by another study in the program?
• What does this study produce that another study needs?
• How will findings be integrated across the studies?

If you cannot answer these questions, the project may be a collection rather than a program.

There are several ways studies can be linked

These categories can overlap. What matters is that the connection is planned rather than invented retrospectively after several studies happen to produce related findings.

Do not divide the project along disciplinary borders automatically

An intuitive arrangement is to assign one study to each discipline: the computer-science study, the psychology study, the education study, and so forth.

That structure can reproduce the very disciplinary separation interdisciplinarity is supposed to address.

A stronger division usually follows empirical or explanatory tasks rather than departmental boundaries. A study may itself draw on more than one discipline because the phenomenon it investigates already lies at their interface.

Some integration should occur before all studies are complete

Waiting until the final discussion section to ask how the studies relate can be risky.

Integration can shape problem framing, terminology, construct definitions, sampling, measures, data structures, hypotheses, and the sequence of studies. The National Academies emphasizes designing research plans around integration rather than merely assembling overlapping disciplinary proposals.

In genuinely linked research, the findings of an earlier study may alter a later study. Even parallel studies can use intentionally aligned constructs or measures so their evidence can eventually be related.

The point is not that every study must use identical terminology or methods. It is that the larger research program should make their eventual integration possible by design.

Linked studies can protect against premature causal stories

A broad interdisciplinary question can tempt researchers to leap from one association to an elaborate explanation.

Suppose Study 1 finds that a characteristic of AI-generated feedback is associated with student reliance. That does not automatically establish the psychological mechanism responsible. A second study can investigate that mechanism explicitly rather than embedding an untested explanation into the interpretation of the first.

Dividing evidentiary tasks can therefore strengthen inference. Each study is responsible for the claim its design can actually support.

Do not create multiple studies merely to make the project look substantial

The opposite problem also occurs. A coherent study can be fragmented artificially into several small studies, sometimes producing unnecessary duplication and weak individual contributions.

A separate study should have an empirical or logical reason to exist. If two questions use the same sample, design, theoretical mechanism, and analysis and are naturally answered together, splitting them may add organizational complexity without improving rigor.

The relevant question is not “Can this become another study?” With enough enthusiasm, almost anything can. Ask whether separation improves the quality of the evidence or the logic of the research program.

The full program must still be feasible

Dividing one impossible study into five impossible studies is not progress.

Evaluate the entire program's demands: participant recruitment, access, data collection, specialist expertise, equipment, ethics approval, analysis, coordination, funding, and time.

A multi-study design can improve intellectual organization while increasing logistical complexity. For student researchers in particular, a beautifully interconnected five-study program may simply exceed the time available for a thesis or dissertation.

In that situation, identify the smallest sequence capable of answering the core question. The remaining studies can become a future research agenda rather than mandatory chapters.

Turning One Overloaded AI Question Into Three Connected Studies

The question contains at least three evidentiary problems: identifying relevant properties of AI responses, understanding how students interpret those properties, and testing whether the proposed mechanism affects reliance behavior.

The studies are not independent. The first informs which features deserve investigation in the second. The second informs the mechanism tested in the third. The final interpretation integrates technological characteristics with learner processes.

Trying to conduct all three tasks simultaneously could require the researchers to specify the important AI characteristics and psychological mechanism before the earlier evidence needed to justify those choices exists.

The linked design therefore does more than reduce workload. It reflects the logical order of the knowledge needed to answer the larger interdisciplinary question.

Common Mistakes When Dividing Interdisciplinary Research Into Studies

Divide the Evidence Where the Research Logic Requires It

When considering several linked studies, identify the dependencies within the larger question before deciding how many studies you need.

Also consider who needs to participate in designing each link. If later studies depend on concepts, methods, or interpretations from other fields, relevant expertise should usually be involved before those dependencies are fixed. This is one reason not to postpone involving collaborators from the relevant disciplines until each study has already been designed.

Before Turning One Interdisciplinary Question Into Several Studies, Check This

Questions About Using Several Linked Studies

Use Several Studies When the Evidence Has Several Necessary Jobs

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.