03 · What You Need to Know
Collaboration Should Begin Before the Study Becomes Difficult to Change
You do not need a complete team to recognize an interdisciplinary problem
Researchers routinely encounter questions that extend beyond their own expertise. Reading outside your field, attending seminars, reviewing unfamiliar literatures, or observing a real-world problem can reveal that another disciplinary perspective may be necessary.
You do not need permission from another discipline to notice that connection.
It is reasonable to develop a preliminary account of the problem, search the relevant literatures, identify possible disciplinary contributions, and formulate tentative questions before recruiting collaborators. Indeed, some degree of preparation helps you explain why collaboration might be worthwhile.
The difficulty begins when preliminary exploration quietly becomes disciplinary decision-making.
Distinguish a provisional idea from a finalized interdisciplinary design
Before collaborators join, you might reasonably identify:
- the broad problem you want to investigate;
- why your current disciplinary approach may be insufficient;
- which additional forms of expertise appear relevant;
- the literature that led you to suspect an interdisciplinary connection;
- tentative questions or hypotheses that collaborators can challenge.
What deserves greater caution is finalizing another discipline's constructs, theories, measures, methodological procedures, or interpretation rules without sufficient expertise in that field.
Productive preparation
You develop enough understanding to identify the problem, explain the potential connection, and have an informed conversation with prospective collaborators.
Premature design
You make consequential theoretical or methodological decisions for another disciplinary component and then recruit someone primarily to implement or endorse them.
Ask whether the collaborator's expertise can still change the study
This is perhaps the most useful practical test.
If you invite a collaborator from another discipline, what can they still change?
If they can challenge the research question, redefine constructs, recommend another theoretical approach, reject an inappropriate measure, alter sampling, redesign a method, or change how disciplinary components are integrated, they are participating in the intellectual development of the project.
If all major decisions are already fixed and their role is simply to run an analysis, provide access to participants, review terminology, or lend their disciplinary credentials to the team, the collaboration is qualitatively different.
That arrangement is not necessarily wrong. Researchers can legitimately provide specialized services or bounded expertise. But it should not be confused with meaningful interdisciplinary co-development when the project claims that several disciplines jointly shape the research.
Interdisciplinary research depends on integration, not representation
The National Library of Medicine's MeSH description characterizes interdisciplinary research as research combining mastery in distinct fields that apply and exchange tools, concepts, ideas, data, methods, or results around a common project. Team-science literature similarly emphasizes collaboration across expertise rather than simply assembling people with different departmental affiliations.
This distinction matters because a person from another discipline does not automatically make the study interdisciplinary. Nor does an interdisciplinary study necessarily require one collaborator for every disciplinary label appearing in the literature review.
What matters is whether the project requires substantive knowledge that the team can competently bring into the design and integrate.
Before recruiting widely, establish whether the research question actually requires multiple disciplines. Otherwise, you may build a team around disciplinary variety that the question itself does not need.
Why late involvement can create theoretical problems
Suppose an education researcher decides to integrate a psychological construct into a study. The researcher selects a popular scale, adds the construct to a conceptual model, formulates hypotheses, and completes most of the proposal. A psychologist is invited afterward.
The psychologist may immediately identify problems. Perhaps the scale measures a narrower construct than the proposal assumes. Perhaps the theory has boundary conditions the education literature rarely discusses. Perhaps another construct would better explain the phenomenon. Perhaps the proposed causal ordering contradicts the theory's original logic.
At that point, the project faces an awkward choice: redesign substantial portions of the study or retain decisions that relevant expertise would have challenged earlier.
Early collaboration does not guarantee agreement. It makes disagreement cheaper.
Methodological expertise can be equally consequential
The same problem occurs with methods.
If an interdisciplinary question depends on qualitative inquiry, computational modelling, physiological measurement, advanced causal inference, network analysis, laboratory procedures, or another specialized method, involving the relevant expertise after data have been collected can be much too late.
Analytical choices are often constrained by design decisions made earlier: sampling, measurement, instrumentation, data structure, timing, missing-data planning, experimental conditions, and documentation.
A collaborator cannot always repair those decisions retrospectively, regardless of how good the analysis is.
Watch Out
Do not treat methodological collaborators as emergency services called after data collection. If their method is essential to answering the research question, their expertise may be needed while the data-generating process is still being designed.
Early collaboration can expose disciplinary assumptions you did not know you were making
One of the less visible benefits of early interdisciplinary discussion is that collaborators can identify assumptions that seem perfectly ordinary within your own field.
You may assume that a construct should be measured at the individual level while another discipline treats it as relational. You may assume that a particular outcome is the natural endpoint while another field considers it an intermediate process. You may treat a technology as an intervention while another discipline conceptualizes it as part of a sociotechnical system.
These disagreements can materially change the research question.
When different disciplinary assumptions make a question internally inconsistent, discovering the problem before preregistration, ethics approval, funding submission, or data collection is considerably preferable.
But “involve everyone from day one” is too simple
There is also a practical limit to early collaboration.
A vague idea such as “I want to combine education, AI, psychology, sociology, ethics, and maybe economics” is not necessarily ready for a large interdisciplinary team. Potential collaborators need enough specificity to determine whether their expertise is actually relevant.
Prematurely assembling a large team can also broaden the project as every participant identifies another worthwhile dimension. The resulting research question may become difficult to bound.
A productive sequence is often iterative: develop the problem far enough to identify what expertise seems necessary, involve relevant collaborators, allow their expertise to reshape the problem, and then refine the design together.
The collaborator does not need to be a representative of an entire discipline
No individual speaks for “psychology,” “computer science,” “education,” or any other discipline. Fields contain competing theories, methods, specialties, and epistemological traditions.
Recruit for the expertise the question requires rather than departmental labels.
If your study concerns natural language processing, a computer scientist whose work is entirely in computer networks may not provide the expertise you need. If the question concerns psychometric measurement, not every psychologist will be a measurement specialist. If it concerns classroom assessment, an education researcher working exclusively in educational policy may be an imperfect fit.
The unit of recruitment should therefore be relevant expertise, not disciplinary decoration.
Collaboration should influence integration, not just separate work packages
In multidisciplinary projects, researchers from different fields may contribute separate analyses or components. Interdisciplinary work places greater emphasis on bringing those contributions into relationship.
That means collaborators should discuss where integration occurs. Does it happen in the research question? The conceptual framework? Measurement? Data collection? Analysis? Interpretation? Several of these?
If every collaborator produces a separate disciplinary work package and the findings meet only in the discussion section, the project may remain primarily multidisciplinary.
This is why the distinction between a genuinely integrated and merely multidisciplinary research idea should influence how the team is formed.
Authorship and responsibility should be discussed before they become awkward
Interdisciplinary projects can create ambiguity about roles because intellectual contributions may not fit familiar disciplinary patterns. Someone may contribute substantially to conceptualization but little to data collection. Another may develop a specialized method. Another may provide interpretation that changes the conclusions.
Discuss expected responsibilities, decision-making, data access, outputs, authorship practices, and communication early enough that collaboration does not depend on unspoken assumptions.
The exact arrangements will vary by project and institutional context. The general principle is simpler: collaboration works better when expectations are explicit before substantial work accumulates.
