03 · What You Need to Know
Multiple Questions Are Acceptable When They Still Constitute One Coherent Inquiry
A research study does not become incoherent merely because it contains RQ1, RQ2, and RQ3. Larger projects such as theses and dissertations may reasonably contain multiple research questions, provided those questions remain connected to a central research problem. Research-methods guidance similarly acknowledges that one project can potentially answer multiple questions while warning that additional questions should not compromise the primary inquiry. ([onlinelibrary.wiley.com](https://doi.org/10.1002/jac5.70095?utm_source=chatgpt.com))
The difficulty is that every research question creates work. It may require particular participants or cases, variables or phenomena, data, measurements, analyses, theoretical justification, and interpretation.
Adding a question therefore does more than add another numbered line to Chapter 1.
The questions should share a central research problem
Imagine a study investigating university students' use of generative AI for academic writing:
RQ1: How frequently do undergraduate students use generative AI for different academic-writing activities?
RQ2: Is frequency of generative AI use associated with students' academic-writing self-efficacy?
These questions are different, but they can plausibly belong to one coherent study. The first describes AI-use patterns. The second examines whether those patterns are associated with another construct relevant to the same research problem.
Now consider adding:
RQ3: How do university administrators develop institutional cybersecurity policies?
The fact that both topics involve universities and technology does not make them one study. The third question concerns a different phenomenon, unit of inquiry, literature, evidence base, and likely methodology.
Multiple questions in one study
Different but connected questions whose answers jointly illuminate one defined research problem.
Several studies disguised as one
Questions connected mainly by a broad topic but requiring substantially independent intellectual and methodological investigations.
The same distinction appears when considering whether you need a main research question and subquestions. Hierarchy can help organize a complex inquiry, but numbering alone cannot create coherence.
A primary question can provide the study with a center of gravity
Many studies benefit from identifying one primary research question. This is particularly important in confirmatory clinical and intervention research, where the primary question may determine the main outcome, sample-size calculation, design, and principal analysis.
Methodological guidance on research-question formulation recommends focusing on the primary question because attempting to pursue multiple primary questions can complicate study design and statistical power. One review specifically cautions that clinical trials with more than one primary study question may become infeasible because questions can require different designs or larger sample sizes. ([pmc.ncbi.nlm.nih.gov](https://pmc.ncbi.nlm.nih.gov/articles/PMC6691636/?utm_source=chatgpt.com))
This does not establish a universal law that every study in every discipline must contain exactly one primary question. Exploratory qualitative studies, case studies, mixed-methods projects, and some dissertations may organize their inquiries differently.
The broader principle is useful nevertheless: readers should be able to identify what the study is fundamentally about.
Secondary questions should contribute something distinct
If a study has a primary question and secondary questions, the secondary questions should not merely restate the primary question with slightly different wording.
Suppose the primary question asks:
“Does retrieval-practice instruction improve delayed recall compared with rereading among undergraduate students?”
A secondary question might investigate whether the intervention's effect differs according to a prespecified characteristic, whether another relevant outcome changes, or whether an important safety or implementation outcome occurs.
Recent methodological guidance describes additional questions as secondary objectives that should remain related to the primary question and should not compromise it. ([onlinelibrary.wiley.com](https://doi.org/10.1002/jac5.70095?utm_source=chatgpt.com))
The distinction becomes important when deciding whether your study should designate one question as primary.
Multiple questions can represent different levels of one inquiry
Not every set of research questions is a flat list. Sometimes one broad central question needs several more focused questions to make the inquiry manageable.
For example:
Central question: “How do first-year university students navigate institutional expectations concerning generative AI in academic writing?”
Subquestion 1: “How do students interpret institutional and instructor guidance about acceptable AI use?”
Subquestion 2: “What considerations shape students' decisions about when to use generative AI?”
Subquestion 3: “How do students respond when expectations differ across courses?”
The subquestions examine identifiable dimensions of one central phenomenon. Their answers can be integrated into an answer to the overarching question.
This structure is particularly natural in some qualitative studies, although it is not limited to qualitative research.
Several questions can use the same dataset
One dataset can legitimately provide evidence relevant to several research questions.
A survey might describe prevalence, estimate associations, and compare prespecified groups. A longitudinal dataset could support questions about change over time and predictors of subsequent outcomes. An interview dataset might illuminate several related dimensions of one phenomenon.
But “the data are already there” is not sufficient justification for adding a question.
Each question still needs a substantive rationale and an appropriate analytical plan. Otherwise, the study risks becoming a search through available variables for publishable patterns rather than an investigation organized around meaningful questions.
This matters particularly in confirmatory quantitative work, where questions and analyses developed after examining results should not be represented as though they had been prespecified.
Different questions can require different analyses without requiring different studies
Suppose a survey study asks:
RQ1: What proportion of students use generative AI for academic writing?
RQ2: Is frequency of generative AI use associated with academic-writing self-efficacy?
The first may require descriptive estimates. The second may require an appropriate associational analysis. Different analyses do not automatically mean different studies.
The study remains coherent if the population, data-generation strategy, conceptual rationale, and overall design were intentionally developed to answer both questions.
This is one reason a descriptive research question can coexist with other question types without being treated as methodologically inferior.
Different questions can even require different methods
A mixed-methods study may intentionally contain questions requiring different forms of evidence.
For example:
RQ1: “How frequently do students use generative AI for assessed writing?”
RQ2: “How do students explain their decisions about when generative AI use is acceptable?”
The first calls for numerical evidence. The second calls for in-depth evidence about reasoning and experience.
Those questions could belong to one mixed-methods study if there is a defensible reason to integrate the two forms of evidence. The methodological difference is not itself a reason to split the project.
However, simply placing a survey and interviews under the same title does not automatically create an integrated mixed-methods study. Whether two research questions can require different methods depends partly on why those methods are needed and how their answers contribute to the overall inquiry.
Each additional question increases the evidentiary burden
Consider a study with five research questions. One requires a prevalence estimate. Another examines an association. A third compares three subgroups. A fourth investigates mediation. A fifth asks participants to explain their experiences qualitatively.
That may be an excellent study. It may also be five ambitious analyses competing for one underpowered sample and one semester.
Multiple questions affect practical decisions such as sample size, recruitment, measurement burden, interview duration, statistical multiplicity, analytical expertise, and time required for interpretation.
Research-question guidance therefore emphasizes feasibility as a central criterion. A question should be answerable within the available participants, expertise, resources, funding, and time. The same principle applies collectively when several questions are placed in one study. ([u.osu.edu](https://u.osu.edu/qmc/research-qs-and-hypotheses/?utm_source=chatgpt.com))
Statistical multiplicity can become important
In quantitative studies, adding multiple hypotheses or outcome comparisons can increase the probability of obtaining apparently noteworthy results simply because many tests have been conducted.
This does not mean every study with several research questions requires the same statistical correction. The appropriate treatment depends on the design, inferential framework, relationship among hypotheses, distinction between primary and secondary analyses, and whether analyses are confirmatory or exploratory.
What matters here is conceptual: more questions can change the statistical problem. Researchers should not assume that ten hypothesis tests create no additional inferential considerations simply because all ten originated from the same questionnaire.
One question should not be split artificially to inflate the count
Sometimes several numbered questions are really one question broken into fragments:
RQ1: What is the level of academic self-efficacy among students?
RQ2: What is the level of academic engagement among students?
RQ3: Is academic self-efficacy related to academic engagement?
This structure may be appropriate if the descriptive estimates are substantively important in their own right. But if RQ1 and RQ2 exist only because the researcher believes every measured variable requires its own research question before testing RQ3, the fragmentation may add little.
Not every descriptive statistic needs a corresponding research question. Researchers routinely describe their samples and variables as part of analysis without elevating every table entry into an independent objective.
One broad question should not hide several unrelated questions either
The opposite strategy is equally problematic:
“What are the prevalence, predictors, causes, experiences, consequences, and possible interventions associated with student burnout?”
Technically, this is one sentence and perhaps one question mark. Methodologically, it may represent an entire research program.
The number of question marks is therefore a poor measure of study scope. A single research question can be too broad to answer in one study, while several carefully bounded questions may form a coherent and feasible investigation.
The questions should be answerable by the study you actually conduct
Every research question creates an implicit promise to the reader: the methods will generate evidence capable of answering it.
If RQ1 requires student survey data and RQ2 requires institutional financial records you cannot access, the fact that both concern the same topic does not solve the problem. If RQ3 asks about long-term outcomes but the study ends after four weeks, it cannot be answered simply because it appears in the proposal.
A useful test is to map each research question to:
- the evidence needed;
- the participants, cases, or data source;
- the relevant measures or qualitative material;
- the analysis required; and
- the conclusion that analysis could support.
If one question has no credible path through that map, either redesign the study or remove the question.
Multiple questions should usually be distinguishable from exploratory analyses
Not every potentially interesting analysis needs to become a formal research question before the study starts.
Researchers may conduct secondary or exploratory analyses that investigate unexpected patterns, generate hypotheses, or examine questions that emerged after data collection. Such work can be scientifically valuable when it is reported transparently.
The problem arises when exploratory questions discovered after looking at the data are presented retrospectively as if they were the study's original prespecified questions.
For confirmatory research especially, distinguishing prespecified primary and secondary questions from later exploratory analyses helps readers understand the evidentiary status of the findings.
More questions do not automatically make a study more comprehensive
Researchers sometimes add questions because a longer list appears to make the project more substantial.
But comprehensive research is not research that asks everything. It is research that adequately answers what it claims to investigate.
Adding a question about every variable, subgroup, outcome, and possible explanation can dilute attention from the central problem. One methodological discussion of question formulation warns that researchers can be tempted to formulate multiple questions once a problem has been identified and emphasizes focusing resources on the primary research question. ([pmc.ncbi.nlm.nih.gov](https://pmc.ncbi.nlm.nih.gov/articles/PMC6691636/?utm_source=chatgpt.com))
Sometimes removing a research question strengthens the study because the remaining questions can be investigated more thoroughly.
Several worthwhile questions may indicate a research program rather than one project
A productive research topic often generates more questions than one study should answer.
You might first describe a phenomenon, then investigate experiences behind the observed pattern, test a plausible mechanism, evaluate an intervention, and later examine whether the findings generalize to another setting. Those are all legitimate questions without needing to coexist in one protocol.
If your list keeps expanding, the problem may not be that the questions are poor. You may simply have discovered a larger research agenda.
The discipline required is deciding which questions belong in the study you are conducting now.
