How Do Philosophical Assumptions Influence What Counts as Evidence?

Why Can Researchers Disagree About What Counts as Evidence?

A test score, interview transcript, observation, physiological measurement, photograph, administrative record, social media post, historical document, and field note can all become research evidence. Yet they do not provide the same kind of knowledge, and their usefulness cannot be judged independently of the question being asked.

This raises a deeper issue. Why might one study treat standardized measurements as central evidence while another gives substantial weight to participants' accounts? Why might one researcher regard observable patterns as sufficient for a particular claim while another looks for underlying mechanisms, meanings, or contextual explanations?

The difference is not necessarily that one researcher believes in evidence and the other does not. Researchers can hold different assumptions about what can be known, what the phenomenon is like, and how observations relate to knowledge claims. Those philosophical assumptions help shape what counts as relevant and persuasive evidence within a particular inquiry.

Evidence Depends on the Question and the Logic of Inquiry

Evidence Is Always Evidence for a Particular Claim

Data and Evidence Are Related but Not Identical Ideas

Researchers often use data and evidence interchangeably, but distinguishing them can be useful. Data are observations, measurements, records, accounts, texts, images, or other materials generated or collected during inquiry. They function as evidence when they are used to support, challenge, refine, or otherwise inform a claim.

A dataset does not explain its own significance. A researcher must establish why the observations are relevant to the question and what conclusions they warrant.

Consider a dataset containing the number of times students access a learning management system. Those records may provide evidence about login frequency. Whether they provide good evidence of engagement, learning, motivation, persistence, or academic success is another matter. Each additional claim requires conceptual and methodological justification.

Evidence therefore has a relational quality: it is evidence for something.

Epistemology Is Central to What Counts as Evidence

Epistemology concerns what can be known and how knowledge claims can be justified. It therefore has an especially direct relationship with evidence.

If researchers assume that credible knowledge about a phenomenon can be developed through systematic measurement and empirical testing, evidence capable of measurement, comparison, replication, or testing may receive particular emphasis.

If researchers seek to understand how people experience and interpret a phenomenon, participants' accounts, observations, interactions, language, and contextual information may provide central evidence because the meanings themselves are part of what the study seeks to know.

If researchers seek explanations involving structures or mechanisms that may not be directly observable, they may reason from observed events and patterns toward explanations of what could have generated them.

The epistemological issue is not simply which data type appears in the study. It is why that data can provide knowledge about the research question.

Ontology Influences What Researchers Think Evidence Needs to Represent

Ontological assumptions concern what exists and the nature of the phenomenon being investigated. These assumptions can affect evidence because researchers need some account of what their observations are supposed to represent.

Suppose the phenomenon is student engagement. If engagement is conceptualized as a construct with dimensions that can be operationalized through observable indicators, standardized measurements may provide relevant evidence for certain questions.

If the study instead investigates engagement as a context-dependent experience whose meaning differs among students, evidence may need to preserve participants' perspectives and the contexts in which engagement is experienced.

If engagement is conceptualized as emerging through relationships among students, teachers, institutional structures, technologies, and practices, evidence focused only on an individual student's score might illuminate only part of the phenomenon.

Ontology does not mechanically select the evidence, but it influences what researchers believe the evidence needs to capture.

The Same Observation Can Support Some Claims but Not Others

One of the most useful habits in research is asking exactly what a piece of evidence establishes.

Suppose learning analytics show that students who frequently access optional online exercises tend to receive higher examination scores. That observation may support a claim that exercise use and examination performance are associated in the studied data.

It does not, by itself, establish that using the exercises caused the higher scores. Students who use optional exercises frequently may differ from other students in prior achievement, motivation, available study time, or other relevant characteristics.

Nor does the association tell us how students experienced the exercises, why they used them, or what learning processes occurred while they were using them.

The evidence has not changed. The proposed claim has.

Research rigor depends partly on keeping the strength and type of claim proportional to what the design and evidence can support.

Different Questions Can Make Different Evidence Relevant

The table does not prescribe methods. It illustrates a more important principle: evidence becomes appropriate in relation to the knowledge claim the researcher is trying to establish.

Quantitative Evidence Is Not Automatically Stronger Evidence

Numbers can provide extraordinary precision. They can represent magnitude, frequency, variation, uncertainty, relationships, and other properties in ways that qualitative evidence cannot replicate. But numerical form does not make evidence automatically relevant or valid.

A precisely calculated statistic from a poorly measured construct may provide weak evidence for the intended claim. A large sample cannot repair a measure that does not adequately represent the phenomenon. Statistical significance does not establish theoretical importance, practical importance, causality, or measurement validity.

Numbers are powerful when the research question, construct definition, measurement, design, analysis, and inference support one another.

Qualitative Evidence Is Not Merely Anecdotal Evidence

Qualitative research is sometimes dismissed as anecdotal because it may involve words, observations, individual experiences, or relatively small samples. That criticism confuses the form of evidence with the methodological reasoning used to produce and analyze it.

A single casual story offered without systematic inquiry is not equivalent to a qualitative study involving purposeful sampling, sustained observation or interviewing, documented analytical procedures, reflexivity, and a defensible interpretive framework.

Qualitative evidence can be particularly appropriate for questions about meaning, experience, context, interaction, process, and interpretation. Its strengths and limitations should be evaluated according to the claims being made rather than by asking whether the findings could have been expressed numerically.

Multiple Forms of Evidence Do Not Automatically Produce a Better Answer

Combining forms of evidence can be valuable. A researcher might examine student outcomes quantitatively while also investigating how students experienced the intervention. The two forms of evidence can address different dimensions of the research problem.

But more evidence is not automatically better evidence.

Researchers need a rationale for how different forms of evidence relate to the research question and to one another. If qualitative and quantitative findings appear inconsistent, that disagreement is not necessarily a problem to be averaged away. It may reveal differences in constructs, contexts, perspectives, or levels of analysis that deserve investigation.

The methodological challenge is integration, not accumulation.

Evidence Is Theory-Laden to Some Degree

Researchers do not encounter observations without concepts. Decisions about what to measure, what to record, which categories to use, which variables to include, what counts as an event, and what deserves interpretation depend partly on prior conceptual and theoretical commitments.

This does not mean observations are arbitrary. It means evidence is produced and interpreted through conceptual frameworks rather than arriving in research as philosophically untouched facts.

For example, counting “student participation” requires a definition of participation. Does asking a question count? Posting in an online forum? Listening attentively? Completing an activity without speaking? The observation procedure depends on what the researcher has decided the concept includes.

Making such decisions explicit improves research because readers can evaluate whether the evidence adequately represents the concept.

Philosophical Pluralism Does Not Mean Anything Can Count as Evidence

Recognizing multiple philosophical traditions can be misunderstood as saying that researchers may call anything evidence if their philosophy permits it.

That does not follow.

Research traditions establish standards for generating, evaluating, and interpreting evidence. Experimental research has standards concerning design, measurement, confounding, statistical inference, and related matters. Qualitative methodologies develop standards concerning sampling, interpretation, credibility, reflexivity, contextualization, and analytical transparency. Historical, ethnographic, participatory, critical realist, and other forms of inquiry likewise require disciplined reasoning appropriate to their aims.

The Crucial Question Is What the Evidence Entitles You to Claim

Evidence should ultimately be evaluated in relation to inference. What conclusion does the researcher want readers to accept, and why should the available evidence justify it?

This question connects methodology with research philosophy. Philosophy helps articulate assumptions about knowledge and reality; methodology explains how inquiry proceeds under relevant assumptions; methods generate or analyze evidence; and the resulting argument establishes what the researcher believes the evidence supports.

The chain can fail at any point. A valid measurement may be irrelevant to the research question. Relevant observations may be analyzed inappropriately. A sound analysis may still be used to support a conclusion stronger than the design permits.

Evidence is therefore not simply something a study possesses. Its strength depends on the relationship among question, assumptions, design, data, analysis, and claim.

How the Same Research Topic Can Require Different Evidence

No single dataset automatically answers all four inquiries. Examination scores cannot fully reveal how students interpret their learning practices. Students' perceptions cannot by themselves establish that learning outcomes improved. Usage logs can show recorded behavior without necessarily revealing why the behavior occurred.

The philosophical and methodological lesson is not that one form of evidence is superior. It is that evidence must fit the phenomenon, question, and claim.

Common Misunderstandings About Evidence in Research

Choose Evidence by Working Backward From the Claim

Instead of beginning with “What data can I collect?”, begin with the question you want the study to answer.

Then identify what claim would constitute an answer and what evidence could reasonably justify that claim. This approach makes philosophical assumptions practical because it forces you to examine the connection between knowing something and demonstrating that you know it.

As you make these decisions, remember that philosophical assumptions can also influence how data are collected and analyzed. What you observe and how you interpret it are part of the same chain of reasoning.

The practical standard is straightforward: be able to explain why this evidence, generated and analyzed in this way, provides defensible grounds for this particular claim.

Before Deciding What Counts as Evidence

Frequently Asked Questions About Philosophical Assumptions and Evidence

Evidence Makes Sense Only in Relation to What You Are Trying to Know

How to Cite This Guide

This guide is intended to be read, shared, and used in research, teaching, and academic work. If you draw on its ideas, explanations, or other content, please acknowledge the source by citing the guide. Doing so gives appropriate credit and helps your readers locate the original resource.