Can Generative AI Help Design a Mixed-Methods Study?

Can AI Help You Combine Quantitative and Qualitative Research Without Creating Two Separate Studies?

You are planning a study examining how university students use generative AI for academic learning. You want to measure patterns of use and academic outcomes, but you also want to understand why students use the technology differently.

You ask generative AI to recommend a mixed-methods design. It suggests administering a questionnaire, conducting interviews, analyzing both datasets, and discussing the findings together.

The plan sounds reasonable, but is it genuinely mixed methods? Does the qualitative component explain the quantitative findings, or are the two components simply placed beside each other?

Mixed-methods research requires more than collecting numerical and textual data. Its methodological value depends on how the different forms of evidence contribute to answering the research question and how their findings are meaningfully integrated.

AI Can Help Plan Mixed Methods, but Integration Must Be Intentionally Designed

How AI Can Support Mixed-Methods Design and Integration

What Makes a Study Mixed Methods?

Mixed-methods research involves the intentional use and integration of quantitative and qualitative approaches within a study or coordinated research program.

Creswell and Plano Clark (2018) emphasize the importance of collecting and analyzing both forms of data and integrating their findings to address research questions.

The integration requirement is important because two methodological components do not automatically form a coherent mixed-methods investigation.

For example, a researcher may administer a quantitative survey and conduct qualitative interviews. If the findings are reported independently without a meaningful relationship between them, the study may be better described as using multiple methods rather than demonstrating substantive mixed-methods integration.

AI can help organize the components, but researchers must establish their complementary purposes.

Why Is Integration More Important Than Simply Using Two Methods?

Integration is the process through which quantitative and qualitative components inform one another or contribute to a combined understanding of the research problem.

Fetters, Curry, and Creswell (2013) describe integration as a central feature of mixed-methods research, including approaches involving connecting, building, merging, and embedding.

For example, quantitative results may identify an unexpected pattern of AI use among students. Qualitative interviews can then investigate how students understand the technology, providing explanations for the observed pattern.

The qualitative findings do not merely supplement the survey with quotations. They contribute to interpreting the quantitative results.

What Mixed-Methods Designs Can AI Help Compare?

Several established designs differ in the timing, purpose, and relationship of their quantitative and qualitative components.

These structures are useful starting points rather than rigid templates. Their implementation depends on the research questions and the methodological contributions expected from each component.

When Is a Convergent Mixed-Methods Design Appropriate?

A convergent design may be useful when researchers want quantitative and qualitative evidence about related dimensions of a phenomenon within a similar research period.

For example, a study of university students' generative AI use might collect questionnaire responses about frequency and purposes of use while conducting interviews about students' academic practices.

The researcher could then examine whether the reported patterns correspond with, expand upon, or differ from the experiences described in interviews.

However, integration requires more than comparing findings that happen to concern the same topic.

The researcher should establish which concepts or questions can be meaningfully examined across the components and how agreement, divergence, or complementary insights will be interpreted.

When Is an Explanatory Sequential Design Appropriate?

An explanatory sequential design begins with a quantitative component whose findings inform a subsequent qualitative investigation.

Suppose a survey identifies that students with similar levels of AI literacy report substantially different levels of academic engagement.

The researcher may select interview participants based on those quantitative patterns and investigate possible explanations for the differences.

The connection between the phases is deliberate. The quantitative findings inform whom to interview, what questions to ask, or which patterns require closer investigation.

AI can help propose interview sampling criteria and questions, but the researcher must verify that the qualitative inquiry addresses the actual quantitative results rather than merely repeating the original questionnaire.

When Is an Exploratory Sequential Design Appropriate?

An exploratory sequential design may begin with qualitative inquiry to identify concepts, practices, or dimensions that subsequently inform quantitative measurement or investigation.

For example, researchers may interview instructors about how they evaluate the appropriateness of generative AI use in assessment.

The resulting qualitative findings might inform the development of a questionnaire examining the prevalence or distribution of identified considerations across a larger population.

However, translating qualitative findings into questionnaire items requires careful construct definition, item development, and appropriate validation procedures.

AI-generated items should not be assumed to possess content validity simply because they reflect themes identified during interviews.

The second phase should follow from the substantive findings of the first rather than from a predetermined list of variables imposed by AI.

Can an Experimental Study Also Use Mixed Methods?

Yes. A randomized or quasi-experimental evaluation may incorporate qualitative inquiry to understand implementation, participant experiences, or contextual processes.

For example, a randomized evaluation of AI-generated writing feedback might estimate differences in writing performance and use interviews to investigate how students engaged with the feedback.

The qualitative component could help interpret adherence, unexpected outcomes, or implementation difficulties.

However, qualitative explanations do not automatically establish the causal mechanisms underlying quantitative treatment effects.

Researchers should distinguish between interpreting participant experiences and demonstrating a causal pathway.

This distinction is particularly relevant when selecting a study design for a causal research question.

How Can AI Help Plan Integration?

AI may assist in identifying where quantitative and qualitative components should interact.

Fetters and colleagues (2013) describe several approaches to integration that provide a useful basis for planning.

AI can suggest integration opportunities, but researchers should decide which approach directly serves the research questions.

What Is a Joint Display, and Can AI Help Create One?

A joint display presents quantitative and qualitative evidence together in a structured visual or tabular format to support integrated interpretation.

Guetterman, Fetters, and Creswell (2015) discuss the use of joint displays in mixed-methods research.

For example, a joint display might present quantitative engagement patterns alongside qualitative explanations from students reporting different experiences of AI tutoring.

These findings are hypothetical illustrations, not empirical results.

AI can help organize a joint display, but it should not invent qualitative quotations, quantitative findings, or integrated conclusions. The final interpretation must follow from the actual data.

What If the Quantitative and Qualitative Findings Disagree?

Divergent findings are not automatically evidence that one component is wrong.

For example, survey results may indicate high satisfaction with an AI tutoring system, while interviews reveal frustration with particular features.

The difference may reflect measurement limitations, variation among participants, different dimensions of satisfaction, or contextual circumstances.

AI may help generate possible explanations, but these possibilities must be examined against the data rather than adopted as convenient reconciliations.

A rigorous mixed-methods interpretation should preserve meaningful disagreement when the evidence does not support a unified conclusion.

Can AI Overlook Methodological Incompatibilities?

Yes. AI may combine quantitative and qualitative procedures without examining their philosophical assumptions, sampling logic, or analytical purposes.

Mixed-methods research does not require every component to share identical epistemological assumptions. Nevertheless, researchers should explain how the approaches can contribute coherently to the investigation.

For example, a qualitative component intended to explore unexpected meanings should not be reduced to confirming predetermined quantitative categories without methodological justification.

Similarly, qualitative findings should not be treated as statistically representative merely because the interviews follow a large survey.

AI-generated designs should therefore be reviewed for methodological coherence rather than evaluated solely by their apparent comprehensiveness.

Does Mixed Methods Always Produce Stronger Evidence?

No. Combining approaches can improve understanding when the components address complementary questions, but it also introduces additional demands.

Researchers must consider the expertise required for both components, the quality of each dataset, the timing of phases, participant recruitment, and the procedures for integration.

A poorly implemented mixed-methods study may provide less defensible evidence than a focused single-method investigation.

AI should not recommend mixed methods merely because the researcher has access to questionnaires and interviews.

How Can You Develop a Mixed-Methods Design With AI?

AI may support each planning stage, but the researcher's methodological rationale must determine the final design.

Using Mixed Methods to Explain Different Patterns of AI Tutoring Engagement

The researcher recognizes that the qualitative interviews should be designed after examining the quantitative findings. Otherwise, the second phase may fail to investigate the patterns that motivated it.

Suppose the survey identifies students with high AI literacy but unexpectedly low engagement. Interviews might explore whether those students use AI primarily for task completion rather than sustained learning.

The decision: The researcher retains the explanatory sequential design because the qualitative phase has a specific explanatory role arising from the quantitative results.

The resulting qualitative explanations remain interpretive findings rather than proof of causal mechanisms. The researcher also avoids assuming that interview accounts represent the experiences of all students with similar survey scores.

The example demonstrates how integration can be planned around a genuine research need rather than added after two datasets have already been collected.

Common Mistakes in AI-Generated Mixed-Methods Designs

Ask AI What the Two Components Will Accomplish Together

Before requesting a mixed-methods design, determine why the research problem cannot be adequately addressed through one methodological approach alone.

AI can then help compare design structures and examine whether the planned integration will produce a meaningful contribution.

A Prompt for Planning Mixed-Methods Integration

Researchers should also examine whether the proposed approach is feasible within the available resources and whether the intended conclusions are compatible with the methods used.

For a complex design, independent methodological review should examine the quality of each component and the adequacy of integration.

The value of mixed methods lies in what the combined evidence helps researchers understand, not in the number of methods included in the proposal.

Before Accepting an AI-Generated Mixed-Methods Design

Frequently Asked Questions About AI and Mixed-Methods Research

Mixed Methods Requires Integrated Reasoning, Not Merely Mixed Data

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.