How Do You Choose Between Quantitative, Qualitative, and Mixed-Methods Content Analysis?

Should You Count What Documents Say, Interpret What They Mean, or Do Both?

You have assembled a corpus of university policies, institutional reports, social media posts, or other documentary materials. You want to investigate how these materials address generative AI, academic integrity, institutional responsibility, or another research phenomenon.

One possibility is to count how frequently particular concepts appear. Another is to examine how documents construct meaning, assign responsibility, or justify institutional decisions.

You might also want to combine these approaches. Perhaps you want to establish how many university policies require AI disclosure while examining how those requirements frame students' responsibilities.

Would counting themes make the study quantitative? Does interpreting passages make it qualitative? If you do both, can you legitimately describe the study as mixed methods?

The decision requires more than selecting software or deciding whether to present frequencies. It concerns the research question, the nature of the evidence, the analytical procedures, and the kinds of conclusions the study can support.

Choose the Approach According to Whether You Need Measurement, Interpretation, or Integrated Evidence

Understanding the Methodological Differences Between Content Analysis Approaches

What Is Content Analysis?

Content analysis is a family of systematic approaches for examining communication and recorded materials.

Depending on the methodology, researchers may classify content, measure recurring features, interpret meanings, or examine relationships among categories.

Relevant materials may include policies, reports, interview transcripts, media content, websites, social media posts, and other documentary or communicative records.

Krippendorff (2018) emphasizes the importance of making inferences from texts or other meaningful matter to the contexts of their use.

However, content analysis does not refer to one universally standardized procedure.

Researchers should specify whether their approach is quantitative, qualitative, or an explicitly integrated combination.

What Is Quantitative Content Analysis?

Quantitative content analysis involves systematically identifying and measuring defined features of content.

For example, a researcher may code whether university AI policies contain explicit disclosure requirements, verification obligations, privacy provisions, or institutional support mechanisms.

The resulting data may be summarized through frequencies, percentages, or other appropriate statistical procedures.

Quantitative content analysis generally requires sufficiently clear categories and recording rules to support consistent measurement.

However, numerical precision does not automatically establish conceptual validity.

The categories must adequately represent the phenomena the researcher intends to measure.

What Is Qualitative Content Analysis?

Qualitative content analysis involves systematically examining textual or other communicative material to develop an understanding of meanings, patterns, categories, and contextual relationships.

For example, a researcher may investigate how university AI policies construct the responsibilities of teachers and students.

The analysis might examine whether responsibility is framed as individual compliance, professional judgment, shared accountability, or institutional support.

Qualitative content analysis can use deductive, inductive, or combined category-development procedures.

It does not necessarily require avoiding all numerical summaries.

Its defining emphasis is the interpretation of content in relation to the research question and analytical context.

What Is Mixed-Methods Content Analysis?

Mixed-methods content analysis involves an intentional combination of quantitative and qualitative analytical components within a coherent research design.

For example, a study may estimate the prevalence of specific AI policy provisions across universities and then examine how selected institutions articulate those provisions.

The two components should contribute distinct evidence and be integrated to address the overall research question.

However, mixed methods is not established merely by reporting theme frequencies alongside qualitative quotations.

The researcher should explain what each component contributes, how the results are connected, and what understanding emerges from their integration.

The terminology should remain consistent with the methodological tradition being used.

How Do the Three Approaches Compare?

These distinctions describe methodological orientations rather than absolute boundaries. Specific procedures vary across content analysis traditions.

Start with the Research Question

The research question should determine the analytical approach.

For example, asking what proportion of university AI policies require disclosure is primarily a quantitative question.

Asking how policies construct student responsibility is primarily interpretive.

Asking whether different policy arrangements are associated with distinct representations of responsibility may justify integrating quantitative classification with qualitative interpretation.

Researchers should avoid choosing mixed methods merely because it appears more comprehensive.

Methodological complexity is justified only when it contributes necessary evidence.

What Is Manifest Content?

Manifest content refers to relatively explicit or directly observable features of communication.

For example, a policy may contain the statement that students must disclose generative AI assistance.

A researcher could code the presence of that explicit requirement.

However, even manifest coding requires decisions about terminology, context, and equivalent expressions.

Policies may communicate the same obligation using different wording.

Researchers should therefore avoid assuming that manifest content can always be measured without interpretation.

What Is Latent Content?

Latent content concerns underlying meanings, assumptions, or interpretive patterns that are not necessarily expressed through one explicit statement.

For example, a policy may repeatedly frame AI use through surveillance, misconduct prevention, and disciplinary responsibility.

A researcher might interpret this as a compliance-oriented representation of AI governance.

Such an interpretation requires attention to multiple passages and their context.

Graneheim and Lundman (2004) discuss manifest and latent content in qualitative content analysis.

Researchers should make the interpretive basis transparent rather than presenting latent meanings as directly observable facts.

Can Quantitative Content Analysis Examine Meaning?

Yes, provided the relevant meanings can be operationalized through defensible coding categories.

For example, researchers might classify policy statements according to whether responsibility is assigned primarily to students, teachers, or institutions.

These categories involve interpretation even though the resulting classifications may be analyzed numerically.

The important issue is whether coding rules are sufficiently clear and whether the categories adequately represent the intended concepts.

Quantitative analysis should not be equated with counting words without contextual understanding.

Can Qualitative Content Analysis Include Counts?

Yes. Researchers may use frequencies to describe how often particular categories occur within the analyzed material.

For example, a qualitative analysis may report that institutional support was discussed in fewer documents than student disclosure obligations.

However, the presence of numerical summaries does not automatically establish a separate quantitative methodological component.

The distinction depends on whether measurement and statistical inference constitute an intentional part of the research design.

Counts should also be interpreted cautiously because frequency does not necessarily indicate importance.

Why Word Frequency Is Not the Same as Conceptual Prevalence

Counting words can provide useful descriptive information, but concepts may be expressed through different terminology.

For example, policies may discuss accountability without using the word "accountability."

Conversely, the word may appear in a document without establishing a substantive accountability mechanism.

Researchers should distinguish lexical frequency from the prevalence of analytically defined concepts.

Automated text searches may support coding, but they do not automatically replace conceptual interpretation.

What Is Deductive Coding?

Deductive coding begins with categories derived from theory, previous research, policy frameworks, or other established conceptual sources.

For example, a study of AI governance might begin with predefined categories concerning transparency, accountability, privacy, and human oversight.

Researchers then examine how these categories appear in the corpus.

Deductive coding may be useful when the research question concerns an established framework or predefined propositions.

However, the framework should not prevent researchers from recognizing relevant content that falls outside the initial categories.

What Is Inductive Coding?

Inductive coding develops categories through engagement with the material rather than relying exclusively on a predetermined framework.

For example, a researcher examining university AI policies may discover recurring distinctions between student responsibility, instructor discretion, and institutional infrastructure.

These patterns may inform category development.

However, inductive analysis is not free from researcher assumptions or prior knowledge.

Researchers should document how categories were developed, refined, and applied.

Inductive and deductive approaches can also be combined when methodologically appropriate.

How Do You Choose the Unit of Analysis?

Content analysis requires decisions about what constitutes the analytical unit.

Possible units include words, sentences, paragraphs, policy provisions, documents, institutions, posts, or communication episodes.

For example, a study may code individual policy provisions but compare results at the university level.

The number of coded passages should not automatically be treated as the number of independent institutional observations.

Researchers should distinguish the unit of coding from the unit of analysis.

The question of identifying the unit of analysis in content analysis is particularly important when documents differ in length or several documents originate from one institution.

How Does Document Sampling Affect the Approach?

Quantitative content analysis intended to estimate population characteristics requires a sampling strategy appropriate to that inference.

For example, estimating the prevalence of AI disclosure requirements across all universities in a jurisdiction requires defensible institutional coverage.

Qualitative content analysis may instead use purposive sampling to examine contrasting representations in depth.

Mixed-methods studies may use different sampling strategies for their quantitative and qualitative components.

However, the relationship between those samples should be explained.

The choice between document sampling and complete analysis should therefore be coordinated with the analytical approach.

What Does Coding Reliability Mean in Quantitative Content Analysis?

Coding reliability concerns the consistency with which defined categories are applied.

For example, two coders may independently classify whether policy provisions require AI disclosure.

Agreement statistics can provide evidence about the consistency of those classifications.

However, reliability depends on the coding design, category distribution, and appropriate statistical measure.

High agreement does not establish that the categories validly represent the intended concepts.

Researchers should evaluate both consistency and conceptual appropriateness.

Does Qualitative Content Analysis Require Inter-Rater Reliability?

Requirements differ across qualitative content analysis traditions.

Some approaches emphasize consistent application of a coding framework and may use agreement assessments.

Others emphasize reflexive interpretation, contextual understanding, and transparent analytical decisions rather than treating coder agreement as the principal quality criterion.

Researchers should identify their methodological orientation and apply quality procedures consistent with it.

Using a reliability statistic without explaining its purpose may create an appearance of rigor without improving the interpretation.

How Do You Integrate Quantitative and Qualitative Results?

Integration involves connecting the two analytical components to develop an understanding that neither provides independently.

For example, a quantitative analysis may identify that explicit AI disclosure requirements appear in a large proportion of the sampled policies.

A qualitative analysis may then examine how those requirements are framed and whether institutional support accompanies them.

The integrated interpretation could distinguish the prevalence of formal obligations from the different ways responsibility is constructed.

However, integration should be planned and demonstrated rather than assumed because both types of results appear in one article.

What Are Common Mixed-Methods Arrangements?

One arrangement involves conducting quantitative and qualitative analyses of the same corpus and integrating the findings.

Another involves using quantitative coding to identify patterns or cases for subsequent qualitative examination.

A third involves qualitative exploration to develop categories that are later measured across a larger corpus.

These arrangements differ in sequence and analytical priority.

Researchers should specify whether the components are concurrent or sequential and explain how one informs the other.

The design should follow the research question rather than a generic preference for combining methods.

When Does Counting Themes Become a Mixed-Methods Study?

There is no universal threshold at which reporting frequencies transforms qualitative content analysis into mixed methods.

A qualitative study may count coded occurrences as a descriptive aid without developing a distinct quantitative inferential component.

Conversely, a study may intentionally use systematic quantitative coding alongside qualitative interpretation and integrate both forms of evidence.

The distinction concerns methodological purpose, analytical procedures, and integration rather than the mere presence of numbers.

Researchers should describe what they actually did instead of selecting a mixed-methods label for its perceived sophistication.

How Does Document Purpose Affect Content Analysis?

Documents created for different purposes may contain different amounts and types of relevant information.

For example, an institutional policy may specify formal obligations, while a public announcement summarizes selected provisions.

Raw theme counts may therefore reflect document function or length rather than institutional priorities.

Researchers should consider how documents produced for different purposes should be analyzed before treating their coded content as directly comparable.

Contextual differences may be important in both quantitative and qualitative approaches.

Can Automated Text Analysis Replace Human Coding?

Automated tools can support searching, classification, frequency analysis, and other content-processing tasks.

However, their suitability depends on the research question, the material, and the validity of the resulting classifications.

For example, automated identification of the word "disclosure" may miss equivalent policy language or include irrelevant uses.

Machine learning and generative AI tools may also produce classification errors or inconsistent interpretations.

Researchers should validate automated procedures against appropriate evidence and document their use.

Software output should not be treated as methodologically valid simply because it was generated automatically.

How Do You Choose Between the Three Approaches?

Begin by specifying whether the study requires numerical measurement, interpretive understanding, or both.

Then consider the documentary population, unit of analysis, coding requirements, and intended conclusions.

If the principal question concerns prevalence, quantitative content analysis may be appropriate.

If the question concerns how meaning is constructed, qualitative content analysis may be preferable.

If both forms of evidence are necessary and can be meaningfully integrated, mixed methods may be justified.

Researchers should avoid assuming that the most complex approach is necessarily the strongest.

How Should the Analytical Approach Be Reported?

The methods section should identify the content analysis tradition, research question, corpus, analytical units, coding procedures, and relevant quality assessments.

Quantitative studies should explain category definitions, measurement procedures, sampling, and statistical analysis.

Qualitative studies should explain category development, interpretation, reflexivity, and how findings were grounded in the material.

Mixed-methods studies should additionally explain the relationship between components and how integration contributed to the conclusions.

Transparent reporting should make clear why the selected approach was necessary and what the evidence can establish.

Choosing a Content Analysis Approach for University AI Governance Policies

The example illustrates that the same documentary corpus can support different analytical approaches. The correct choice depends on what the researcher wants to establish, not simply the format of the documents or the availability of coding software.

Common Misconceptions About Content Analysis Approaches

How to Select the Appropriate Content Analysis Approach

Begin by identifying the kind of knowledge the research question requires.

Then choose analytical procedures that can produce that evidence and support the intended interpretation.

The strongest approach is the one whose analytical procedures correspond to the research question and whose findings can be interpreted within the limits of the documentary evidence.

Check Your Content Analysis Design Before Coding

Frequently Asked Questions About Quantitative, Qualitative, and Mixed Content Analysis

Choose Content Analysis According to the Evidence Your Question Requires

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.