03 · What You Need to Know
Research Development Is Often Less Linear Than Textbooks Make It Look
Starting with theory makes sense when theory is the object of inquiry
Some studies originate from a theoretical problem rather than an empirical topic alone. Perhaps previous studies have produced inconsistent evidence for a theoretical proposition. A theory may contain an unresolved relationship, an uncertain mechanism, or a boundary condition that has received little empirical attention. Researchers may also want to determine whether an established explanation applies in a different context.
In such cases, theory logically enters early.
Theory Identify an unresolved theoretical proposition, mechanism, assumption, or boundary.
Research problem Determine what remains uncertain or insufficiently explained.
Research question Formulate a question capable of producing evidence relevant to that theoretical issue.
Study design Select evidence and methods appropriate for answering the question.
This sequence is common in deductive research, where propositions from theory can inform hypotheses that are then examined empirically. Theory is not being attached to an already completed study. It helps generate the inquiry itself.
This is particularly appropriate when the intended contribution involves testing, applying, extending, or building theory.
Starting with the research problem can be equally legitimate
Many researchers do not begin because a theory has caught their attention. They begin because something in the empirical world needs explanation.
An educator notices that students use an AI tool in unexpectedly different ways. A health researcher encounters a persistent implementation problem. An organizational researcher observes that an intervention works in one setting but appears ineffective in another. These problems can generate worthwhile research questions before a particular theory has been selected.
The researcher can then examine the literature to determine which theoretical perspectives illuminate the phenomenon.
Research problem Identify what is unknown, puzzling, consequential, or insufficiently understood.
Preliminary research question Clarify what the study needs to understand or explain.
Theoretical search Evaluate theories that address the phenomenon, mechanisms, concepts, or relationships implicated by the question.
Refinement Revise the question if theoretical engagement reveals a more precise or consequential way to formulate the inquiry.
Starting with the problem has one practical advantage: it reduces the temptation to begin with a favorite theory and then search for something to study with it.
The research question does not need to be permanently fixed before you read theory
A preliminary research question is not a contractual obligation.
Suppose you initially ask, “What factors influence university faculty members' use of generative AI in teaching?” As you review the literature, you encounter theories distinguishing intention, actual behavior, perceived capability, environmental conditions, and other constructs. You may realize that “factors influence use” is conceptually vague.
Theoretical engagement could lead you to ask a more focused question about a particular mechanism or relationship. That is not evidence that the original question was wrong. It is part of developing the study.
Conversely, reading theory may reveal that your original question is broader than the theory's explanatory scope. Rather than distorting the question to make the theory fit, you might reject the theory and continue searching.
This is why choosing a theory without forcing one to fit requires treating theoretical selection as an evaluative process rather than a hunt for terminology that resembles your topic.
Theory can sharpen a research question
A research question determines what you are asking. Theory can help clarify why particular relationships, mechanisms, or interpretations deserve attention.
For example, “Why do students continue using generative AI?” is a legitimate broad question. Different theories might transform how the researcher conceptualizes “why.” One might emphasize individual beliefs and intentions. Another might foreground social practices. Another might focus on institutional conditions, identity, motivation, or learning processes.
Each theoretical perspective directs attention differently.
The question may therefore become more precise after theoretical engagement. This reflects one of the central roles theory can perform in research: it structures what the researcher considers potentially explanatory.
But theory can also distort a research question
The iterative relationship between theory and question has a danger. Once researchers become committed to a theory, they may begin rewriting the research problem until it resembles the theory.
Suppose the original problem concerns how institutional rules affect faculty use of generative AI. The researcher discovers a familiar individual-level technology adoption theory that largely emphasizes personal beliefs. If the study is redesigned to measure only those beliefs because the theory provides convenient constructs and validated instruments, the research question may no longer address the original problem adequately.
Watch Out
Refining a research question because theory reveals a more useful way to understand the problem is legitimate. Changing the question primarily so that a preferred theory appears to fit is different. The research problem should not become collateral damage in the search for a theoretical framework.
Deductive studies often give theory an earlier role
In strongly deductive research, theory commonly precedes the final hypotheses because theoretical propositions provide the reasoning from which empirical expectations are derived. Published guidance on hypothesis development likewise emphasizes grounding hypotheses in coherent theoretical mechanisms rather than assembling disconnected predictions.
The broad sequence may therefore be theory, theoretical reasoning, hypotheses, empirical test.
Even here, however, the actual research process may be less tidy. Researchers may begin with an empirical puzzle, search for explanations, compare theories, refine the research question, and only then derive formal hypotheses. The finished article often presents a cleaner logical sequence than the chronological history of how the project developed.
Inductive inquiry may give theory a later or more flexible role
Some qualitative and exploratory methodologies place greater emphasis on remaining open to meanings, concepts, and relationships that emerge through engagement with data. In such work, committing too early to a highly specified theoretical framework may direct researchers toward what the theory predicts they should find.
This does not make the research intellectually empty or permanently disconnected from theory. Existing concepts may still sensitize researchers to potentially important phenomena, and theoretical engagement may become central during analysis or interpretation.
Empirical examples of theory-driven qualitative research also show that theory can legitimately inform research questions, sampling, coding, and analysis when this is consistent with the study's methodological orientation. The relevant issue is therefore not whether qualitative inquiry is “allowed” to begin with theory, but what theoretical commitment does to the openness and logic of the particular design.
This is why exploratory research can still be theory-driven.
Different theories can lead to different versions of the question
There may be several plausible theories for the same phenomenon. Contemporary methodological guidance emphasizes that different theories foreground different concepts and relationships. Choosing among them therefore affects how researchers understand the phenomenon rather than simply supplying alternative citations.
Imagine a study about student disengagement in online learning. A motivational theory might direct attention toward beliefs and goals. A social perspective might emphasize interaction and belonging. A self-regulation perspective might foreground planning, monitoring, and control of learning.
The empirical phenomenon remains student disengagement, but the theoretical lens changes what becomes analytically salient.
Researchers should therefore ask not only “Does this theory relate to my question?” but also “What version of the problem does this theory encourage me to see, and what might it leave outside the frame?”
Final alignment matters more than chronological order
A defensible study should eventually demonstrate coherence among its research problem, research question, theoretical framework, methodological choices, analysis, and claims.
That coherence does not require pretending the study emerged in a perfectly linear sequence.
You may revise your research question after reading theory. You may reconsider your theory after piloting an instrument. You may discover during qualitative analysis that another theoretical perspective explains emerging patterns more productively. Depending on the methodology, such iteration can be entirely appropriate if it is handled transparently.
The final study should tell a coherent intellectual story, but research development itself is often recursive.
04 · A Practical Example
How a Research Question and Theory Can Develop Together
Hypothetical Example
Why do faculty members continue using generative AI?
A researcher observes that many faculty members experiment with generative AI, but some abandon it while others integrate it into routine teaching. The researcher initially formulates a broad question: “What influences faculty members' continued use of generative AI for teaching?”
Initial question The question establishes the phenomenon of interest but does not yet specify an explanatory mechanism.
Literature and theory The researcher reviews competing theoretical explanations of continued technology use and identifies one that distinguishes beliefs, intentions, and subsequent behavior.
Fit assessment The researcher checks whether the theory addresses continued use rather than merely initial adoption and whether its assumptions make sense for faculty use of generative AI.
Question refinement The broad question is refined into questions concerning the particular theoretically relevant relationships that the study can investigate.
Design alignment Measures and analyses are selected to examine those constructs and relationships while preserving contextual variables needed to address the original problem.
The research question came first chronologically, but it did not remain unchanged. Theory sharpened it. Had the researcher instead begun with a theoretical limitation concerning continued technology use and then selected generative AI as a context for examining that limitation, theory could legitimately have come first.
Both pathways can produce coherent research.
