03 · What You Need to Know
Collaboration Should Improve the Research, Not Simply Increase Its Scope
Research collaboration can broaden expertise, expose assumptions, improve methods, and reveal questions that one investigator might not have considered. A collaborator who understands a population, theory, measurement problem, analytical method, or practical setting particularly well may identify an important omission.
That is one of the intellectual benefits of collaboration. It does not follow, however, that every worthwhile idea should be incorporated into the same study.
Ask Whether the Question Strengthens the Existing Research Problem
Begin with the same criterion you would apply if you had proposed the question yourself: what does it contribute?
Suppose your study asks whether university students' AI literacy is associated with their ability to critically evaluate AI-generated academic information. A collaborator suggests examining whether that relationship differs according to prior experience with generative AI.
The additional question could fit naturally. It qualifies the primary relationship, uses the same population, and may be investigated using compatible evidence.
Now suppose another collaborator wants to examine whether those students support institutional bans on generative AI in examinations. The topic remains generative AI, but the question may require a different conceptual rationale involving academic integrity, assessment policy, fairness, and institutional regulation.
Both questions may be good. Only one necessarily belongs in the existing investigation.
Do Not Confuse Expertise With Scope
A collaborator may possess expertise that makes a new question feasible. For example, adding a qualitative researcher may make interviews possible, while adding a statistician may enable a sophisticated longitudinal analysis.
That new capacity can be scientifically useful, but capability does not establish necessity. The study should not expand merely because the research team can now perform another analysis.
Useful expertise
A collaborator contributes knowledge or methods that improve how the study answers its existing or naturally extending questions.
Scope expansion
A collaborator introduces another substantial research purpose primarily because their expertise or interests make it possible to pursue.
The distinction can be subtle. Collaboration often changes research questions for the better. The test is whether the revised question structure becomes more coherent or merely more ambitious.
Determine What the New Question Requires
Every additional research question carries methodological consequences. Even a seemingly modest addition may require new variables, instruments, participant groups, time points, sample-size considerations, qualitative data, statistical models, or ethical procedures.
Before agreeing to include the question, design the analysis as though it mattered. Ask what evidence would actually be necessary to answer it convincingly.
| Question to ask |
Adding it may make sense when... |
Separation deserves consideration when... |
| How does it relate to the primary problem? |
It explains, extends, qualifies, or contextualizes the main investigation. |
It introduces another substantial scientific purpose. |
| What evidence does it require? |
The existing design can answer it appropriately with proportionate additions. |
It requires major new measures, samples, methods, or follow-up. |
| What happens to feasibility? |
The project remains manageable and adequately resourced. |
The addition compromises recruitment, measurement, analysis, time, or budget. |
| How will the findings be interpreted? |
The answer contributes naturally to the same scientific argument. |
It requires a largely independent discussion and set of implications. |
| Why is it being added? |
The scientific case remains persuasive regardless of who proposed it. |
The principal justification is that a collaborator wants it included. |
Protect the Primary Question
Where a study distinguishes primary and secondary questions, the hierarchy should have practical meaning. The primary question commonly drives major design decisions, while secondary questions support or extend the primary objective.
Clinical-trial guidance provides a particularly explicit version of this principle. ICH E9 recommends that secondary variables be predefined, that their roles in interpreting the study be explained, and that their number be limited to questions relevant to the trial's objectives. The exact conventions differ across research traditions, but the broader lesson travels reasonably well: additional questions should have defined roles rather than accumulating without constraint.
If accommodating a collaborator's question reduces statistical power, qualitative depth, participant retention, measurement quality, analytical attention, or interpretive clarity for the primary question, the cost may be too high.
This is the point at which a secondary question begins distracting from the primary question.
Do Not Add a Question Merely to Create a Publication Opportunity
A collaborator may reasonably see publication potential in data being collected. Large projects often produce multiple legitimate papers, and distinct research questions can share participants or datasets. The problem is not planning several outputs.
The problem arises when the research design begins accumulating weakly connected questions mainly so that each team member can have a paper.
Publication planning should follow the scientific structure of the project rather than determine it. If a collaborator's question is sufficiently independent to justify a separate paper, that may itself suggest that the question should be developed as a separate project using shared research infrastructure where appropriate.
Shared Participants Do Not Require a Shared Study
Researchers sometimes feel compelled to incorporate a collaborator's question because recruiting another sample later would be inefficient. That concern is understandable, but participant sharing does not necessarily mean intellectual integration.
Several questions may use the same participants while belonging to different studies, provided the relevant methodological, ethical, consent, and reporting requirements are satisfied.
A research team could therefore coordinate data collection efficiently without pretending that every question belongs to one study.
The Same Dataset Can Support the Collaborator's Separate Question
The same principle applies when a collaborator proposes a question that can be answered from data already being collected. The fact that the variables will exist makes the project feasible, but it does not determine where the question belongs.
A sufficiently rich dataset may support several distinct research projects. If the collaborator's question has an independent rationale, it may be cleaner to specify it as a separate analysis or project rather than stretch the current study's purpose.
That decision should be made prospectively when possible, with appropriate attention to consent, ethics approval, data governance, analytical plans, and later transparency about overlapping data.
Collaboration Also Raises Questions About Contributions and Credit
Adding a research question may change who contributes what to the project. These conversations are easier before the work is completed than after the manuscript exists.
The Contributor Roles Taxonomy, or CRediT, provides a standardized vocabulary for describing contributions to research outputs. Its roles include conceptualization, methodology, formal analysis, investigation, data curation, supervision, and several forms of writing, among others. CRediT is intended to improve transparency about contributions; it does not itself define who qualifies for authorship.
If a collaborator proposes and leads a substantial additional question, discuss responsibilities, data access, analysis, writing, and intended outputs early. Do not assume that suggesting a question automatically determines authorship position, nor that authorship expectations should determine whether the question is scientifically included.
Watch Out
Avoid bargaining with research questions: one question for one collaborator, another question for another paper. Contribution and credit should be discussed transparently, but the study's question structure should remain scientifically defensible.
Sometimes the Best Answer Is “Good Question, Different Project”
Rejecting a question from the current study does not mean rejecting the idea or the collaborator. A question may be excellent precisely because it deserves more attention than the current project can give it.
If the new question has its own theoretical foundation, substantial analytical requirements, or independent implications, it may be better to develop it as a separate study.
This can preserve collaboration while protecting both projects from unnecessary scope. Academic diplomacy occasionally benefits from good research design.
04 · A Practical Example
When a Collaborator's Idea Strengthens the Study and When It Expands It
Hypothetical Example
A Collaborative Study of Generative AI Use
A team is planning a study examining factors associated with university students' use of generative AI for academic writing. A collaborator with expertise in AI literacy proposes asking whether AI literacy is associated with students' patterns of use.
Another collaborator, whose research focuses on faculty development, proposes adding a question about how teachers redesign assessment in response to generative AI.
Evaluate the first suggestion AI literacy provides a plausible explanatory factor within the existing student-focused problem and can be measured within the planned design.
Decision on the first suggestion
The team incorporates the question because it strengthens the explanation of the primary phenomenon, not simply because the collaborator proposed it.
Evaluate the second suggestion
Faculty assessment redesign is also important, but it requires another population, different literature, different evidence, and a substantially different analytical approach.
Consider the relationship
The faculty question remains relevant to the team's broader interest in generative AI in higher education, but the student study does not need it to answer its central question.
Decision on the second suggestion
The team develops the faculty question as a related project rather than expanding the student study.
Both collaborators contributed useful ideas. Treating the ideas differently reflects their relationship to the study, not their relative importance or the status of the people proposing them.
