03 · What You Need to Know
Statistics Are One Way of Analyzing Evidence, Not the Definition of Research
Research is broader than numerical analysis
Research is fundamentally concerned with systematic investigation and defensible knowledge claims. The evidence and analytical procedures required depend on what the researcher wants to understand.
If the question asks how common a phenomenon is, whether two measured variables are associated, how accurately an outcome can be predicted, or how large an intervention effect may be, quantitative evidence and statistical analysis will often be appropriate.
If the question asks how people experience a phenomenon, how they interpret an event, how a social process unfolds, how practices develop within a particular context, or why participants understand the same situation differently, qualitative evidence may provide the more appropriate route to understanding.
Neither approach becomes research merely because it uses a recognized technique. Both must satisfy the broader characteristics expected of systematic research.
Quantitative research works with numerical representation
Quantitative research represents relevant features of a phenomenon numerically so that they can be described, compared, modeled, or analyzed mathematically or statistically.
Researchers might estimate the prevalence of a condition, compare average outcomes between groups, model relationships among variables, estimate an intervention effect, quantify uncertainty, or make predictions from measured characteristics.
Statistics are therefore not decorative calculations added after data collection. They provide tools for reasoning from numerical evidence. Appropriate statistical methods depend on the study design, measurement, sampling process, assumptions, and inferential question.
Qualitative research addresses questions that are not adequately reduced to numerical variables
The U.S. National Library of Medicine's Medical Subject Headings describes qualitative research as research using nonnumeric information to explore individual or group characteristics, with findings not produced through statistical procedures or other quantitative means.
That definition captures an important distinction, although qualitative research encompasses diverse traditions that cannot be reduced to a single procedural formula. Qualitative researchers may investigate experiences, meanings, perceptions, interactions, practices, processes, identities, cultures, institutions, texts, or other phenomena for which context and interpretation are central.
Evidence may include interview transcripts, field notes, observations, documents, photographs, audiovisual materials, online interactions, diaries, artifacts, or other forms of material. The analytical work involves systematic interpretation rather than simply converting everything into numerical variables.
Qualitative research is not simply research with words
A useful introductory distinction is that quantitative research often works primarily with numbers while qualitative research often works primarily with words, images, observations, and other nonnumeric material. But treating that distinction as the complete definition can be misleading.
Qualitative research involves methodological choices about how phenomena are conceptualized, how participants or cases are selected, how evidence is generated, how context is handled, how interpretation proceeds, and how claims are justified.
Researchers may code interview transcripts, compare cases, identify patterns, construct categories, develop themes, examine narratives, analyze discourse, or build theoretical interpretations. Which procedures are appropriate depends on the qualitative methodology being used.
The rigor lies in the logic connecting the question, evidence, analytical process, and interpretation, not simply in the fact that the source material consists of words.
Qualitative research can be a complete standalone study
Qualitative research is sometimes treated as preliminary work that researchers conduct before the "real" quantitative study. That is unnecessarily restrictive.
Qualitative findings can certainly help generate hypotheses, develop measurement instruments, identify relevant constructs, or explain quantitative results. But qualitative research can also answer substantive research questions in its own right.
For example, a researcher might investigate how doctoral students experience supervisory conflict, how teachers make decisions about generative AI in assessment, how patients understand a difficult diagnosis, or how organizational practices change during institutional reform. If the objective is to understand meanings, experiences, processes, and context, qualitative analysis may provide the central evidence rather than merely preparing the way for subsequent statistics.
Authoritative health-sciences guidance likewise recognizes qualitative research as capable of standing alone as well as forming part of mixed-methods research.
Qualitative does not mean anecdotal
A researcher conducting interviews does not simply collect interesting quotations and select a few that support a preferred argument. That would provide little protection against selective interpretation.
Rigorous qualitative research requires a defensible approach to generating, organizing, examining, and interpreting evidence. Depending on the methodology, this may involve purposive sampling, iterative data collection and analysis, systematic coding, constant comparison, attention to discrepant cases, reflexive documentation, triangulation, detailed contextualization, or an audit trail of analytical decisions.
Not every qualitative methodology uses all of these practices, and applying them mechanically does not guarantee quality. The broader principle is that qualitative claims should emerge from a transparent and methodologically coherent engagement with the evidence.
This is one reason research differs from anecdote and expert opinion. A participant's individual account may become evidence within a qualitative study, but the research contribution comes from how such evidence is systematically examined and interpreted in relation to the question.
Rigor does not belong exclusively to statistics
Researchers sometimes use rigorous as though it meant "quantitative." The concepts should be separated.
A quantitative study can be methodologically weak. Its sample may be inappropriate, measurements invalid, assumptions violated, analyses poorly chosen, missing data mishandled, or conclusions exaggerated. Running sophisticated software does not repair those problems.
A qualitative study can likewise be weak if its methodological choices are incoherent, evidence is insufficient for the claims, analytical procedures are opaque, contradictory evidence is ignored, or interpretations outrun the material.
Rigor therefore concerns the quality and defensibility of the research process relative to the methodology being used. The appropriate standards differ, but the need for systematic and rigorous research does not disappear when statistics do.
Qualitative research does not require statistical significance
A qualitative study ordinarily does not need to produce a p-value to validate its findings. Statistical significance answers a particular type of question within particular statistical frameworks. It is not a general-purpose test of whether any research finding is trustworthy.
Qualitative researchers instead justify claims through standards appropriate to their methodological tradition. These may concern credibility, transparency, reflexivity, coherence, contextual adequacy, evidential grounding, transferability, or other criteria.
The terminology varies across qualitative traditions, which is important. Grounded theory, ethnography, phenomenology, narrative inquiry, case study research, and qualitative content analysis do not become rigorous by following one generic checklist. Each has methodological commitments that affect how evidence should be generated and interpreted.
Sample size works differently in qualitative research
A common objection to qualitative research is that its samples are often small. This criticism assumes that every study is trying to estimate population parameters statistically.
Qualitative sampling may instead be designed to obtain information-rich cases, examine variation, understand a process deeply, develop conceptual categories, or investigate a particular context. Participants may be selected purposively because their experiences are especially relevant to the research question.
This does not mean sample size is irrelevant. Researchers still need enough appropriate evidence to support their analytical aims, and the rationale should be explicit. What changes is the logic used to determine adequacy.
A qualitative study with 20 carefully selected participants and intensive interviews is not simply an underpowered quantitative study waiting for another 180 respondents. It is answering a different kind of question.
Counting things does not automatically make a study quantitative
Qualitative researchers may sometimes report frequencies or descriptive counts. For example, a researcher might indicate that a particular issue appeared in most interviews or describe how often certain categories occurred.
The presence of a number does not automatically transform the overall methodology into quantitative research. The relevant issue is how evidence is conceptualized and analyzed and what role quantification plays in the argument.
Conversely, collecting open-ended responses does not automatically make a study qualitative. If researchers mechanically convert responses into predefined numerical categories and analyze only the resulting counts, the analytical logic may be primarily quantitative.
Methodological labels should describe how the study actually produces and interprets evidence rather than the superficial appearance of the raw data.
Qualitative and quantitative research are not methodological enemies
Debates sometimes frame qualitative and quantitative research as competing camps, as though researchers must choose which side they belong to before discovering what their question is.
A more productive principle is methodological fit. Different forms of evidence make different aspects of a phenomenon visible.
| Research Question |
Potentially Appropriate Approach |
Why? |
| How prevalent is academic burnout in a defined student population? |
Quantitative |
The question requires numerical estimation in a defined population. |
| How do students experiencing burnout describe its effect on their academic lives? |
Qualitative |
The question concerns experience, meaning, and context. |
| Does a specified intervention reduce measured burnout compared with a control condition? |
Quantitative experimental design |
The question concerns estimating an intervention effect. |
| Why did participants respond differently to the intervention? |
Qualitative or mixed methods |
Understanding mechanisms and experiences may require contextual evidence beyond outcome measurements. |
| How much did burnout change, and how did students explain that change? |
Mixed methods |
Numerical change and participants' interpretations provide complementary evidence. |
The table illustrates tendencies rather than rigid boundaries. Complex questions may be addressed through several defensible designs, and methodological traditions differ in how they conceptualize the relationship between evidence and knowledge.
Mixed methods deliberately integrates qualitative and quantitative evidence
Mixed methods research uses both qualitative and quantitative approaches within an integrated research design. The NIH Office of Behavioral and Social Sciences Research describes mixed methods research in terms of collecting, analyzing, and integrating quantitative and qualitative data to provide a more comprehensive understanding of a research problem than either approach might provide alone.
The word integrating matters. Conducting a survey and adding several interview quotations does not automatically produce a strong mixed-methods study. Researchers should explain why both forms of evidence are needed, how each component addresses the research problem, and how the findings will be connected.
Quantitative research
Uses numerical representation and quantitative analysis to address questions involving measurement, distributions, differences, relationships, prediction, effects, or other numerical patterns.
Qualitative research
Uses systematic analysis of nonnumeric or interpretive evidence to investigate meanings, experiences, processes, practices, contexts, and other phenomena requiring in-depth understanding.
Mixed methods research
Integrates qualitative and quantitative approaches because their combined contribution provides a more useful understanding of the research problem.
Mixed methods is not automatically better than using one method
Using both qualitative and quantitative methods can be valuable when the research problem genuinely requires both. It can also create unnecessary complexity when one approach would answer the question adequately.
Mixed-methods research demands competence in both components as well as a defensible strategy for integration. It may require more time, expertise, coordination, and analytical work. NIH guidance therefore emphasizes that the research methods should fit the research problem and that mixed methods is particularly appropriate when a quantitative or qualitative approach alone would be inadequate for the desired understanding.
Adding interviews to a quantitative study merely to call it mixed methods does not make the study more rigorous. Neither does adding percentages to qualitative findings.
The question should determine whether statistics are needed
A simple methodological principle resolves much of the confusion: identify what you need to know before deciding what kind of evidence to collect or analyze.
If you need to estimate magnitude, frequency, numerical differences, associations, probabilities, or effects, statistics may be indispensable. If you need to understand how people interpret an experience, how a process unfolds, or how context shapes practice, qualitative analysis may be more informative. If the problem genuinely requires both, integration may justify a mixed-methods design.
This follows the same logic behind the conclusion that research does not universally require an experiment or hypothesis. Methods are selected because they help answer particular questions, not because every research project must display the same methodological equipment.
