Stating Your Ontological/Epistemological Position

Every research design rests on two prior commitments: what you believe about the nature of reality (ontology) and what you believe about how that reality can be known (epistemology). CASRAI’s Research Paradigm Explained guide covers what the commonly-taught paradigms are and how they got their names. This guide is the practical companion to that one: a short diagnostic you can run against your own research question, worked through once as an illustrative example, so you can locate your own position on each axis and see exactly how it should show up in the concrete choices a reviewer or examiner will be checking for.

Why this has to be explicit, not assumed

A paradigm statement that names positivism, interpretivism, or pragmatism but isn’t visible anywhere in the actual sampling, instrument, or analysis choices is a common source of reviewer pushback — the mismatch signals the position was written after the fact to justify a method chosen for convenience, not the other way around. Many programs and journals with a strong methodological-reflexivity tradition (education, nursing, social work, information science, most qualitative dissertations) expect an explicit statement early in the methods section; much quantitative and biomedical work leaves the position unstated by convention. Either way, the exercise below is the same: work out where your research question actually puts you, then let that position visibly drive the rest of the methods section.

The two axes, as a spectrum rather than a box

It’s easier to locate an honest position by treating ontology and epistemology as two separate spectrums rather than trying to pick one of four named paradigms outright:

Ontology — realism to relativism. At the realist end, a single reality exists independently of anyone studying it and can, at least in principle, be measured. At the relativist end, reality is socially and experientially constructed, so there can be multiple, equally valid accounts of the same phenomenon depending on who is experiencing it.
Epistemology — objectivism to subjectivism. At the objectivist end, the researcher can study a phenomenon as a detached, external observer without changing it. At the subjectivist end, knowledge is inevitably co-created through the interaction between researcher and participant, and detachment isn’t achievable or even desirable.

Most working researchers land somewhere along both spectrums rather than at a pure endpoint — the point of the diagnostic below is to locate your actual position using your own research question, not to memorize which named paradigm you’re supposed to pick.

A five-question diagnostic

Answer these against your actual research question — not against what you think you’re “supposed” to say:

What does the phenomenon mean? Is there one correct account of it, multiple valid accounts depending on who’s asked, or is the honest answer that you mainly need a workable account good enough to act on?
Your role relative to what you’re studying. Can you observe it without your presence changing it, or is your presence — and the relationship you build with participants — unavoidably part of what produces the data?
What counts as strong evidence for your question. Standardized, replicable measurement; rich, first-person, context-rich accounts; or whatever mix of the two actually answers the question in front of you?
What you want to be able to say at the end. A generalizable estimate that should hold across a defined population; a transferable insight that readers in a similar context can judge for themselves; or a specific, actionable finding for a particular setting?
How the question is already phrased. Is it a “how much / how many / is there a relationship between” question, a “how do people experience or make sense of” question, or a “what should we do about” question?

Answers that consistently point toward “one correct account,” “detached observation,” and “standardized measurement” lean realist/objectivist (the post-positivist end). Answers that consistently point toward “multiple valid accounts,” “co-created through interaction,” and “rich first-person accounts” lean relativist/subjectivist (the constructivist/interpretivist end). A mix that keeps landing on “whatever actually answers the question” is a pragmatist lean, and is a legitimate, defensible answer in its own right — not a failure to commit.

Worked example: running the diagnostic

The example below is an illustrative composite built to show how the diagnostic works end to end — it is not a real study, and no specific researcher, institution, or dataset is being described.

Composite research question: “How do first-generation graduate students experience mentorship within a research-intensive doctoral program?” Running it through the five questions:

What the phenomenon means — multiple valid accounts: mentorship is experienced differently depending on the student’s background, department, and relationship with their advisor.
Researcher’s role — co-created: interviews and the rapport built with participants will shape what they choose to disclose.
Strongest evidence — rich first-person accounts: a standardized mentorship-satisfaction scale would flatten exactly the variation the question is asking about.
What to say at the end — a transferable insight that a graduate program could weigh against its own context, plus concrete, program-relevant recommendations for improving mentorship structures.
How the question is phrased — “how do students experience,” already an interpretivist framing rather than a “how much/how many” one.

Tallying those five answers: four point to the relativist/subjectivist (constructivist) end, and one — the desire for concrete, program-relevant recommendations rather than pure description — has a pragmatist lean in how the findings will be used, without changing the underlying ontology/epistemology. The composite study’s position: a constructivist ontology and a subjectivist epistemology, with a pragmatist orientation toward producing findings the program can actually act on. That’s a coherent, common combination — pragmatism about the use of findings doesn’t require abandoning a constructivist account of what the data actually are.

From position to method choice

The diagnosed position should be traceable into every major methods decision, not just asserted once and left behind. For the composite study above:

Sampling logic: purposive sampling for information-rich cases (e.g., maximum-variation across departments and mentorship structures), not probability sampling aimed at population-level representativeness.
Instrument: a semi-structured interview guide with open-ended prompts the participant can redirect, not a fixed-response survey instrument.
Researcher’s stance: disclosed reflexivity about the researcher’s own position relative to participants (see CASRAI’s Reflexivity and Positionality Statements guide for the template for that specific paragraph — a related but distinct artifact from the ontological/epistemological position statement this guide covers), rather than a blinding or standardization protocol designed to remove the researcher’s influence.
Analysis approach: interpretive, inductive coding that stays close to participants’ own language, not a pre-registered statistical hypothesis test.
Rigor criteria: credibility, transferability, and trustworthiness — the qualitative analogues to reliability and validity — rather than statistical generalizability to a broader population.

A study that names a constructivist position in its opening paragraph and then reports only a structured survey with closed Likert items and inferential statistics has exactly the mismatch reviewers flag — the position statement and the actual method choices have to agree.

Writing the position statement for your methods section

A working template, in four moves:

Name the ontological stance.
Name the epistemological stance.
State briefly why that stance fits the research question.
Name two or three concrete methodological consequences that follow from it.

Applied to the composite example above, one workable version of that paragraph:

“This study adopts a constructivist ontology and a subjectivist epistemology, treating first-generation graduate students’ experiences of mentorship as multiple, context-dependent accounts rather than a single measurable phenomenon. Because the aim is to understand how meaning is made within specific mentoring relationships, the study uses purposive, maximum-variation sampling and semi-structured interviews rather than a standardized survey instrument, and analyzes the resulting data through inductive, interpretive coding rather than statistical hypothesis testing. Trustworthiness is addressed through prolonged engagement with participants and an audit trail of coding decisions, consistent with the qualitative criteria appropriate to this paradigm.”

Notice the paragraph does three things in four sentences: names both stances, ties them to the actual research question, and points forward to specific choices the rest of the methods section will then have to deliver on.

Common mistakes

Naming a paradigm the methods don’t back up. The most common failure mode — see above.
Treating pragmatism as “I don’t have to justify anything.” Pragmatism still requires stating why the specific mix of methods chosen fits the research question; it removes the requirement to pick a single ontology/epistemology, not the requirement to justify the design.
Conflating epistemology with method. Interviews aren’t inherently interpretivist — a structured interview with a fixed coding scheme can sit comfortably in a post-positivist design; what matters is the underlying stance on reality and knowledge, not which data-collection technique is used.
Overclaiming certainty in either direction. A realist position doesn’t require claiming perfect, certain knowledge (that’s what post-positivism corrects for), and a relativist position doesn’t require claiming that no account is ever more defensible than another within its own context.

Frequently asked questions

Do I have to choose exactly one of the four named paradigms, or can I sit between them?

You can sit between them. The five-question diagnostic above is deliberately built around two spectrums rather than four fixed boxes, because most real research questions land somewhere along both axes rather than at a pure endpoint. What matters for your methods section is that the position you state is coherent with the choices you actually made, not that it matches a named paradigm exactly.

Is “realism” the same as “positivism”?

They’re related but not identical. Realism is the ontological claim (a mind-independent reality exists). Positivism bundles that ontological claim with a specific epistemological claim (that reality can be known with certainty through detached observation) and a specific methodological consequence (standardized, quantitative measurement). Post-positivism keeps the realist ontology but softens the epistemological claim to “knowable only imperfectly and probabilistically” — see CASRAI’s Research Paradigm Explained guide for the full breakdown of post-positivism versus classical positivism.

What if my discipline or journal never asks for this statement explicitly?

Then don’t force one in artificially — check your program’s or target journal’s actual expectations rather than assuming a paradigm statement is required. Running the diagnostic is still useful privately even when it won’t appear as its own paragraph, because it’s the same reasoning that should be driving your sampling, instrument, and analysis choices either way.

Can a quantitative study still require you to state a position?

Yes, though it’s less commonly asked for explicitly. Most contemporary quantitative social science and biomedical research operates from an unstated post-positivist position by convention — a single, imperfectly-knowable reality, measured through standardized instruments, with findings treated as probabilistic rather than certain. Some programs and journals do ask quantitative researchers to name that assumption explicitly rather than leave it implicit.

Where does axiology fit into this?

Axiology — the role of the researcher’s own values in the inquiry — is often added as a fourth element alongside ontology, epistemology, and methodology. It’s most directly relevant when a study also involves disclosing the researcher’s own positionality (identity, background, relationship to participants) rather than only their philosophical stance on reality and knowledge; CASRAI’s Reflexivity and Positionality Statements guide covers that specific, related artifact in detail.

Ontology and Epistemology in Research Design: Making Your Position Explicit