Qualitative and quantitative approaches are widely distinguished in research, information analysis, problem-solving, and decision-making. Qualitative analysis is generally associated with language, meaning, characteristics, and interpretation, while quantitative analysis deals with numerical measurements and relationships. Qualitative and quantitative research have also developed as relatively distinct methodological traditions, each with its own techniques and standards of evaluation.
However, the world we seek to understand is not inherently divided into a qualitative world and a quantitative world. Objects and phenomena have characteristics and properties, while also exhibiting measurable differences in size, intensity, frequency, quantity, and other dimensions. Qualitative and quantitative descriptions arise from how we perceive, conceptualize, describe, and compare these aspects of reality.
For example, saying “This material is hard” is a qualitative description, while saying “The material has a Vickers hardness of 250 HV” is quantitative. These statements do not concern entirely different realities. They describe the same property using different forms of representation. Nevertheless, the numerical value has meaning only because we have defined hardness and established a method for measuring it.
In this sense, quantitative measurement often presupposes a qualitative or conceptual understanding of what is being measured. Conversely, detailed examination of quantitative changes may reveal previously unrecognized properties, differences, or phenomena.
Qualitative and quantitative approaches are not simply opposing methods. The important task is to understand their characteristics, select them according to the purpose of an investigation, and move between them when necessary.
How Are Qualitative and Quantitative Thinking Related to Human Cognition?
When people perceive the world, they distinguish objects and phenomena from one another. We may classify something as a metal, polymer, liquid, machine, or living organism. Identifying what something is, how it differs from other things, and what functions it performs involves qualitative recognition.
At the same time, we can compare objects in terms of size, weight, temperature, speed, or number. When such comparisons are expressed numerically according to established rules, they become quantitative information.
However, the distinction is not always clear-cut.
Temperature can be measured quantitatively, yet a change in temperature may cause a substance to undergo a phase transition from liquid to gas. Similarly, concepts such as human health or customer satisfaction encompass multiple characteristics but can be represented numerically through specific measurement procedures.
An important distinction must therefore be maintained between the properties and quantitative differences that exist in the objects being studied and the concepts and measurement scales through which humans represent them.
For example, measurements of length and mass correspond to physical properties. However, measuring customer satisfaction or organizational creativity requires decisions about how these concepts are defined and which questions or observations will represent them.
Expressing something numerically does not mean that the concept being measured has been objectively or definitively established.
Conversely, information expressed in words can support rigorous knowledge when it is based on reliable observations and clear evidence.
The simple contrast between qualitative analysis as subjective and ambiguous and quantitative analysis as objective and precise is misleading. In both cases, what matters is the object of investigation, how it is conceptualized and measured, and what evidence supports the conclusions.
Qualitative and Quantitative Approaches in Research
In market research and social research, a common explanation is that qualitative research is used to discover problems and generate hypotheses, while quantitative research is subsequently used to test those hypotheses.
For example, a company developing a new product may begin by interviewing customers to investigate usage problems and unmet needs. The findings are then used to formulate hypotheses, followed by a questionnaire survey to estimate how widely those problems or needs are shared.
This sequence is understandable and can be useful in practice. However, real investigations do not always proceed in such a neat order.
Researchers may begin by analyzing large volumes of numerical data, discover an unexpected pattern, and then conduct interviews to understand the reasons behind it.
Alternatively, qualitative and quantitative investigations may proceed in parallel, with their findings compared and integrated throughout the project.
Qualitative research is not limited to hypothesis generation, nor is quantitative research limited to hypothesis testing.
Exploratory statistical analysis may generate new hypotheses. Qualitative observations may also challenge or contradict existing hypotheses.
The Flexibility of Qualitative Research
Qualitative approaches such as interviews and field observations allow researchers to discover new perspectives while an investigation is underway.
Suppose a company begins customer interviews with the assumption that a product’s high price is its principal sales problem.
As the interviews proceed, customers may reveal that their main concerns are actually difficulty understanding how to use the product or uncertainty about after-sales support.
The researcher can then revise subsequent questions to investigate usability and customer support in greater depth.
The scope of the investigation might also be expanded to include types of customers who were not originally considered.
This flexibility is an important strength of qualitative research.
Even when researchers do not fully understand which questions are important at the beginning of an investigation, they can revise their questions as new evidence emerges.
Methodologies such as Grounded Theory emphasize the iterative relationship between data collection and analysis, in which emerging findings inform subsequent sampling, observation, and questioning.
However, flexibility should not be confused with arbitrariness.
When research questions or analytical perspectives change, researchers should document why the changes were made and what evidence supports them. The validity of the resulting interpretations must still be examined.
The Constraints of Predefined Quantitative Research
Questionnaire surveys typically require researchers to define questions, response options, and measurement procedures before data collection begins.
For example, suppose a survey asks customers why they purchased a product and offers only four options: price, performance, design, and brand.
After completing the survey, researchers may realize that after-sales support was actually a major consideration.
However, because this factor was not included in the questionnaire, its importance cannot be measured directly from the collected responses.
An open-ended question might allow respondents to mention previously unanticipated factors, but there is no guarantee that every respondent will provide the relevant information.
Thus, the questions defined before data collection constrain the range of information that can subsequently be obtained.
Nevertheless, quantitative research is not necessarily rigid.
Pilot studies can be used to improve questionnaires. Surveys can be conducted in successive stages, and adaptive survey designs can modify subsequent questions according to earlier responses.
Furthermore, standardized questions have an important advantage: they improve comparability across respondents.
If questions are changed freely during a survey, responses collected before and after those changes may no longer be directly comparable.
Predefined research designs are therefore not merely limitations. They also serve to maintain consistency and comparability in measurement.
Does Systematic Qualitative Research Necessarily Produce Better Results?
Qualitative research often emphasizes systematic data collection and analysis.
Examples include standardizing interview questions, recording observations in predefined forms, assigning codes to statements, and organizing information into established categories.
These procedures have important advantages. They can reduce inconsistencies in recording, make information easier to compare, and improve the transparency and traceability of analytical decisions.
However, if being systematic is understood as forcing all information into predetermined formats, qualitative research may lose much of its flexibility.
For example, an investigation of manufacturing problems might begin with the categories Equipment, Materials, People, and Methods.
These categories can be useful, but the actual problem may involve communication between departments or inadequate management of design changes.
If observations that do not fit the established categories are forced into them, important findings may be overlooked.
The KJ Method emphasizes discovering meaningful groupings from the collected information rather than assigning observations immediately to predefined classifications.
Similarly, a foundational principle of Grounded Theory is that concepts and theoretical explanations should be developed through engagement with empirical observations rather than imposed uncritically from existing theories.
What matters is not necessarily the rigidity of the procedure but whether the analytical process can be examined and evaluated.
For example, preserving original observations and documenting why categories were changed or how interpretations were developed can improve traceability without sacrificing flexibility.
Rigor does not always mean mechanically repeating the same procedure. The ability to revise an interpretation in response to new evidence, while explaining that revision transparently, is also an important form of rigor.
Qualitative and Quantitative Approaches in Thinking and Information Analysis Methods
Thinking and analytical techniques are also commonly classified as qualitative or quantitative.
Brainstorming, the KJ Method, SCAMPER, TRIZ, and SWOT Analysis are generally regarded as qualitative methods.
Statistical hypothesis testing, regression analysis, Design of Experiments, and multivariate analysis are typically quantitative.
However, many techniques occupy intermediate positions or can incorporate both qualitative and quantitative elements.
| Method | Main Purpose | Relationship Between Qualitative and Quantitative Approaches |
|---|---|---|
| Brainstorming | Generate a range of ideas | Primarily qualitative, but ideas may also be counted or evaluated numerically |
| KJ Method | Discover meaningful groupings and relationships in information | Primarily qualitative, but the frequency of categories may be analyzed quantitatively |
| 5 Whys | Investigate causal hypotheses | Primarily qualitative; numerical evidence can be used to verify proposed causal relationships |
| Cause-and-Effect Diagram | Organize possible causes systematically | Primarily qualitative; effects can be investigated through experiments and statistical analysis |
| SWOT Analysis | Organize internal and external conditions | Primarily qualitative, but evaluation and prioritization can incorporate numerical measures |
| Analytic Hierarchy Process (AHP) | Support decisions involving multiple criteria | Converts judgments into numerical comparisons and calculates priorities |
| Design of Experiments (DOE) | Investigate factor effects and interactions | Primarily quantitative, but factor selection and hypothesis formulation require qualitative judgment |
| Multiple Regression | Estimate relationships between several variables and an outcome | Quantitative; model construction and interpretation depend on conceptual judgment |
| Principal Component Analysis (PCA) | Summarize information from multiple variables | Quantitative; the meaning of components requires interpretation |
| Factor Analysis | Estimate underlying factors shared by observed variables | Uses quantitative models to investigate structures interpreted as qualitative concepts |
| Cluster Analysis | Group observations according to similarity | Uses numerical methods to form groups whose meanings must be interpreted |
Consider AHP, for example.
A decision-maker may judge that performance is more important than price and express this preference through a numerical comparison.
Mathematical procedures then calculate relative priorities.
However, the result depends on the judgments entered by the decision-maker. An accurate calculation does not establish that those judgments are objectively correct.
A similar issue arises in PCA and factor analysis.
Mathematical procedures can produce components and factors, but the meaning assigned to those results depends on the original variables and the analytical purpose.
Conversely, qualitative information can be organized into categories, and the number of observations associated with each category can then be analyzed quantitatively.
Thus, the distinction between qualitative and quantitative cannot be determined solely by the name of a method.
It depends on several aspects, including the input information, analytical operations, outputs, and interpretation of the results.
Three Approaches to Integrating Qualitative and Quantitative Research
Methods for combining qualitative and quantitative research within a single investigation have been systematically developed under the name Mixed Methods Research.
Three representative designs are particularly useful for understanding different forms of integration.
Exploratory Sequential Design — From Qualitative to Quantitative
Researchers begin by exploring the problem through interviews, observations, or other qualitative methods.
The findings are used to identify important concepts, develop measurement instruments, or formulate hypotheses.
A subsequent questionnaire survey or statistical investigation examines how widely the identified relationships or characteristics occur.
This approach is especially useful when the important variables or research hypotheses are not sufficiently understood at the beginning.
Explanatory Sequential Design — From Quantitative to Qualitative
Researchers begin by analyzing survey results or other numerical data.
They then identify distinctive outcomes, unexpected patterns, or relationships that require further explanation.
Interviews or field observations are conducted to investigate why those patterns may have occurred.
For example, a survey might reveal that customer satisfaction is high while customer retention remains low.
Follow-up interviews could investigate customers’ actual circumstances and behaviors to identify explanations that were not apparent from the numerical data alone.
Convergent Design — Qualitative and Quantitative Approaches in Parallel
Researchers collect qualitative and quantitative information during approximately the same period and then compare or integrate the findings.
For example, a product team might analyze usage logs while also conducting user interviews.
The quantitative data reveal how people use the product, while interviews explore their experiences and explanations.
Agreement between the results may strengthen an interpretation.
However, disagreement can be equally informative.
Users might report that a product is easy to operate, while usage logs reveal that they repeatedly undo or restart operations.
This discrepancy could prompt a reconsideration of questionnaire wording, usage circumstances, or difficulties that users do not consciously recognize.
Nevertheless, merely conducting two kinds of research does not necessarily constitute meaningful integration.
The important question is how knowledge obtained through one approach changes the questions, interpretation, or subsequent investigation conducted through the other.
Practical Example — Investigating Problems in Customer Support
Suppose a company observes an increase in customer support inquiries.
An initial analysis of support records reveals that inquiries are concentrated around a particular product.
At this stage, the company has identified a quantitative problem: an unusually high number of inquiries.
However, the underlying reasons remain unclear.
Researchers then examine the text of support records and interview customers.
They discover that the primary problem is not general difficulty operating the product. Instead, customers struggle to find necessary information when handling certain exceptional situations.
Based on this finding, the inquiries can be classified into categories such as Routine Operations, Exceptional Procedures, and Functional Defects.
The company can then calculate the number of inquiries and average handling time for each category.
Next, 5 Whys or Cause-and-Effect Diagrams may be used to explore why inquiries involving exceptional procedures occur so frequently.
The company might develop hypotheses about improvements to the user interface or the presentation of support information.
Where appropriate, experiments could compare alternative interfaces or information formats and measure their effects on inquiry frequency and task completion time.
The investigation moves repeatedly between qualitative and quantitative approaches:
Detect a Quantitative Anomaly → Investigate the Content Qualitatively → Develop New Categories → Analyze Category Frequencies → Formulate Causal Hypotheses → Test Through Experiments
If the original survey or data collection system had been restricted to predetermined categories, the problem of finding information for exceptional procedures might never have been recognized.
On the other hand, qualitative investigation alone might not establish how frequently the problem occurs or how significantly it affects operations.
Using both approaches makes it possible to investigate the nature of the problem and its prevalence and consequences.
The Meaning and Limitations of Quantification
Converting qualitative information into numerical representations can be extremely valuable.
For example, observing that customers experience difficulties with product operation does not reveal how widespread the problem is.
If researchers define what constitutes an operational difficulty and establish appropriate measurement procedures, they can investigate its frequency and severity quantitatively.
However, quantification requires decisions about what counts as the same kind of phenomenon and what should be counted.
For example, the number of customer inquiries may appear to be a straightforward measure.
But if the same customer contacts support three times about the same problem, should this be counted as three inquiries or one underlying problem?
The result depends on the definition.
Similarly, when customer satisfaction is measured on a five-point scale, the difference between ratings of five and four does not necessarily have the same meaning as the difference between three and two.
Even when numerical calculations follow precise rules, the meaning of the numbers depends on the underlying concepts and measurement procedures.
Therefore, integrating qualitative and quantitative analysis requires more than converting words into numbers.
It is important to preserve the relationship between numerical representations and the original observations from which they were derived.
If analysts can return from statistical results to the underlying cases, they can reexamine contextual information that may have been lost during aggregation or quantification.
Quantitative Change and Qualitative Change — Perspectives from Hegel and Marxist Philosophy
Philosophical discussions of quality and quantity provide another perspective on their relationship.
In Science of Logic, G. W. F. Hegel examined quality, quantity, and the concept of Measure, which concerns their relationship.
One idea in this discussion is that quantitative changes may occur within certain limits without altering an object’s defining quality, while changes beyond particular thresholds may produce a qualitative transformation.
This idea was subsequently developed within Marxist philosophy.
Friedrich Engels explicitly discussed the transformation of quantity into quality, and vice versa, as a principle of dialectical thinking.
A familiar example is the relationship between temperature and the physical state of water.
Under specified pressure conditions, changes in temperature can lead to transitions between solid, liquid, and gaseous states.
In this case, quantitative changes in temperature are related to qualitative changes in the physical state of matter.
This illustrates an important point: qualitative characteristics and quantitative conditions are not necessarily independent.
However, combining qualitative and quantitative research is not the same issue as explaining how quantitative changes produce qualitative changes in natural phenomena.
The former concerns methods of human cognition and investigation.
The latter concerns philosophical interpretations of processes occurring in reality.
Furthermore, there is no need to accept the claim that all phenomena undergo qualitative transformations according to a single universal dialectical law.
The actual conditions and mechanisms of change must be investigated in each case.
When these distinctions are maintained, philosophical discussions of quality and quantity can provide a useful perspective for reconsidering the limitations of a rigid qualitative–quantitative dichotomy.
Integrating Qualitative and Quantitative Analysis in the AI Era
The development of large language models and machine learning creates new opportunities to work with qualitative and quantitative information within the same analytical environment.
For example, LLMs can extract concepts and characteristics from documents, interview transcripts, and open-ended customer responses.
Once the extracted information has been structured, it can be summarized by category or subjected to statistical analysis.
Text embeddings can also represent documents as numerical vectors, allowing similarities to be analyzed using clustering and dimensionality reduction.
These numerical representations may help reveal meaningful groupings and previously unrecognized characteristics.
Conversely, when statistical analysis or machine learning identifies an unusual cluster or pattern, analysts can return to the original documents and cases to investigate its meaning and context.
Consider the exploration of new technological themes in research and development.
A possible workflow is:
Collect Documents → Extract Technical Elements → Represent Characteristics Numerically → Analyze Similarities and Structures → Identify Interesting Groups → Reexamine Original Documents → Form New Technical Hypotheses
Qualitative interpretation and quantitative analysis occur repeatedly throughout this process.
However, categories generated by AI and numerical similarities derived from embeddings do not necessarily represent valid semantic relationships.
For example, documents from different contexts may appear similar because they share technical terminology.
Conversely, technologies based on related underlying principles may be described using different terminology and therefore remain distant in a particular embedding space.
The important task is not to treat machine-generated numbers or categories as final conclusions, but to retain the ability to return to the original evidence and examine its meaning.
It is also desirable to allow analytical perspectives, extraction criteria, classification schemes, numerical representations, and analytical procedures to be revised when new findings emerge.
AI-assisted integration of qualitative and quantitative analysis is not merely about converting language into numbers. Its greater value lies in enabling movement between numerical patterns and new questions, and between those questions and renewed examination of the original information.
Qualitative and Quantitative Thinking in ThinkNavi
ThinkNavi does not treat qualitative and quantitative analysis as two entirely separate product categories.
Instead, it approaches them as different but complementary ways of supporting thinking and problem-solving.
For example, the KJ Method may be used to explore and organize an unclear problem.
5 Whys and Cause-and-Effect Diagrams may help generate possible causal explanations.
Design of Experiments can be used to test selected causal hypotheses, while multiple regression and multivariate analysis can investigate relationships among several variables.
Decision Matrices and AHP may help compare alternatives when decisions involve multiple criteria.
Organizing these methods solely according to whether they are qualitative or quantitative would obscure the relationships between them.
A more useful approach is to organize methods according to what users are trying to discover or accomplish.
Nevertheless, ThinkNavi does not eliminate the distinction between qualitative and quantitative approaches.
Different methods require different types of data, assumptions, analytical procedures, and standards of interpretation.
ThinkNavi aims to explain these characteristics while allowing users to connect different techniques as their investigations develop.
LLM-assisted analysis may further support natural-language problem formulation, information extraction, hypothesis generation, interpretation of numerical results, and suggestions for additional investigation.
These are directions for future development and do not imply that all such functions are currently available.
The important question is not whether qualitative or quantitative analysis is superior.
Nor is it necessary to apply both approaches to every problem.
What matters is the ability to revise the questions and classifications established at the beginning of an investigation when new evidence appears. Moving from qualitative understanding to quantitative examination, and then returning from numerical findings to renewed qualitative understanding, may reveal knowledge that neither approach would have uncovered alone.
Related Topics
- What Is Statistics? — Discovering Patterns in Data and Understanding an Uncertain World
- Multivariate Analysis — Discovering Relationships and Structures Among Multiple Variables
- The KJ Method | Building Meaningful Structures from Fragmented Information
- Grounded Theory
- Concept Formation | How Categories Are Made
- Brainstorming — Generating More Ideas by Suspending Judgment
- Five Whys | Exploring Root Causes by Repeatedly Asking “Why?”
- Cause-and-Effect Diagram (Fishbone Diagram) — Systematically Organizing Factors That May Contribute to a Problem
- Design of Experiments (DOE) — Investigating Multiple Factors and Their Interactions Through Systematic Experimentation
- AHP | Pairwise Comparisons & Consistency
- Abduction | Reasoning Toward Explanations
- Scientific Method | Hypotheses, Tests, and Discovery
- Dialectics — Dialogue, Contradiction, and the Development of Ideas
- Concept Exploration
References and Further Reading
- John W. Creswell and Vicki L. Plano Clark, Designing and Conducting Mixed Methods Research, SAGE Publications.
- Barney G. Glaser and Anselm L. Strauss, The Discovery of Grounded Theory: Strategies for Qualitative Research, Aldine, 1967.
- Michael D. Fetters, Leslie A. Curry, and John W. Creswell, “Achieving Integration in Mixed Methods Designs—Principles and Practices,” Health Services Research, 48(6), 2013, pp. 2134–2156.
- Nick Pilcher and Martin Cortazzi, “‘Qualitative’ and ‘Quantitative’ Methods and Approaches Across Subject Fields,” Quality & Quantity, 58, 2024, pp. 2357–2387.
- “Practical Guidance to Qualitative Research: Part 3—Sampling, Data Collection and Analysis,” European Journal of General Practice, 2018.
- Robert Adcock and David Collier, “Measurement Validity: A Shared Standard for Qualitative and Quantitative Research,” in Working with Concepts, Cambridge University Press.
- G. W. F. Hegel, Science of Logic — Quantity.
- Friedrich Engels, Anti-Dühring, Chapter XII: Quantity and Quality.
- Friedrich Engels, Dialectics of Nature — Dialectics.