Can Generative AI Build a Mixed-Methods Joint Display?

Can AI Turn Quantitative and Qualitative Findings Into a Meaningful Joint Display?

You have completed the quantitative and qualitative analyses of a mixed-methods study. Your survey results are presented in tables, while your interview findings are organized into themes. Now you need to show how the two sets of evidence relate to one another.

A joint display could help, but constructing one requires decisions about which findings belong together, how their relationships should be represented, and what conclusions can reasonably be drawn from their combination.

Can generative AI build this display for you? More importantly, can it distinguish meaningful integration from a table that simply places numbers beside quotations?

Yes, AI Can Draft a Joint Display, but Its Analytical Relationships Need Verification

How Joint Displays Work and Where Generative AI Can Help

What Is a Mixed-Methods Joint Display?

A mixed-methods joint display is a visual arrangement that brings quantitative and qualitative information together to facilitate integration. It may take the form of a table, matrix, diagram, or another structured representation.

Rather than presenting each research strand independently, the display makes relationships between them explicit. Depending on the study design, these relationships may involve agreement, complementarity, contradiction, explanation, or the development of one strand from another.

For example, a researcher examining student engagement in online learning might place survey findings about participation alongside interview themes describing how students experience different learning activities. A further column could explain what the combined evidence suggests.

The essential feature is not the table itself. It is the analytical relationship represented by the display.

What Information Does AI Need to Construct a Joint Display?

Generative AI can arrange information into a convincing table even when the supplied evidence is incomplete. That makes preparation particularly important.

Before requesting a joint display, researchers should provide the study's mixed-methods design, research questions, relevant constructs, and verified findings from each strand.

For quantitative results, this may include sample sizes, descriptive statistics, estimated effects, confidence intervals, or other outputs relevant to the study. For qualitative findings, it may include theme definitions, supporting excerpts, variation across participants, and contextual qualifications.

The model should also know whether the strands involve the same participants, linked subsamples, or independent samples. Without this information, it may suggest relationships at the individual level that cannot be established from the available evidence.

When preparing these materials, researchers should distinguish between integrating quantitative and qualitative findings and merely summarizing them in a common format.

What Should a Joint Display Contain?

There is no universal joint-display template. Its structure should follow the integration purpose and the study's design.

For a convergent study, a practical structure might include the research question, quantitative finding, qualitative finding, relationship between strands, and integrated interpretation.

Additional columns may be appropriate for methodological qualifications, discrepant cases, or evidence identifiers. A more elaborate display is not necessarily a better one. Each column should contribute to the analytical purpose.

Can AI Build Different Joint Displays for Different Mixed-Methods Designs?

Yes. AI can adapt a display's proposed structure when the researcher supplies the study's design and integration purpose.

These are illustrative arrangements, not mandatory formats. Some joint displays represent integration during instrument development or sampling rather than the final comparison of results.

AI should therefore not select a display structure solely because it resembles a familiar table from a published article.

Can AI Determine Whether Findings Converge or Diverge?

AI can propose relationship labels, but researchers must assess whether those labels are justified.

Convergence generally concerns findings that support compatible interpretations. Complementarity occurs when strands contribute different but mutually informative perspectives. Divergence involves findings that appear inconsistent or contradictory, while some findings may have no defensible direct relationship.

For example, a high average satisfaction score and an interview theme about flexible scheduling may be compatible. However, they do not necessarily measure the same construct, and the qualitative finding does not automatically explain the numerical result.

Researchers should verify any proposed convergence between findings and examine apparent divergence or contradiction before accepting the relationship column.

Can a Joint Display Contain Meta-Inferences?

Yes. Joint displays are often used to develop or communicate interpretations that arise from examining the strands together.

A mixed-methods meta-inference should reflect what the combined evidence supports, rather than merely restating a statistical result or qualitative theme.

AI may suggest a candidate interpretation, but it may also introduce an unsupported causal explanation, overlook uncertainty, or overstate the representativeness of qualitative evidence.

A defensible display should make it possible to trace the interpretation back to the findings from which it was developed.

What Are the Risks of AI-Generated Joint Displays?

The most consequential risk is that an orderly table may make weak analytical relationships appear established.

• Artificial matching: AI may pair findings because they contain similar terms rather than because they address comparable constructs.
• Selective representation: It may omit findings that complicate an otherwise consistent interpretation.
• Invented evidence: It may generate plausible statistics, participant quotations, or theme descriptions that were never supplied.
• False precision: It may assign definitive labels such as convergence or contradiction where the relationship remains uncertain.
• Unsupported explanations: It may treat a qualitative account as proof of the mechanism behind a quantitative association.
• Loss of context: It may remove qualifications concerning sampling, timing, measurement, or analytical limitations.

Should Researchers Use AI to Format or Interpret the Joint Display?

These are related but distinct tasks.

Formatting a display from an already verified integration matrix is generally a more constrained task. Asking AI to decide which findings belong together and what they mean introduces greater interpretive uncertainty.

Researchers can reduce this risk by separating the workflow into two stages: first establish and verify the analytical relationships, then use AI to improve the display's organization and presentation.

If AI contributes to the analytical stage, its suggestions should be treated as provisional until independently evaluated.

Creating a Joint Display for an Online Learning Study

Step 1: Organize the Verified Findings

The researcher supplies three quantitative findings: 76% of surveyed students report overall satisfaction, 81% consider flexibility beneficial, and 59% experience connectivity difficulties at least occasionally.

The qualitative analysis identifies appreciation for flexible scheduling, difficulties participating during unstable connections, and differences between satisfaction with the overall arrangement and experiences during particular learning activities.

Step 2: Ask AI to Construct a Preliminary Joint Display

The researcher asks the model to align findings by construct, identify the nature of each proposed relationship, and avoid introducing evidence not included in the supplied material.

Step 3: Evaluate the Proposed Relationships

The researcher examines whether the survey items and interview themes genuinely address comparable aspects of online learning.

The flexibility row may support convergence if both strands concern sufficiently similar evaluations. The connectivity row offers a different relationship: the survey reports how commonly difficulties were experienced, while the interviews provide accounts of their consequences.

The satisfaction row requires additional caution. The researcher cannot infer from the aggregate survey percentages that every satisfied respondent also experienced difficulties. Nor can interview accounts establish the prevalence of such combinations in the wider sample.

Step 4: Revise the Integrated Interpretation

The researcher may conclude that the evidence suggests broadly favorable evaluations of online learning alongside persistent practical barriers. The interviews help characterize those barriers and clarify why overall satisfaction should not be interpreted as uniformly positive experiences across all activities.

The joint display makes these relationships visible without claiming that the qualitative evidence establishes the causes of survey responses.

Step 5: Retain an Auditable Version

Each quantitative entry should be traceable to the relevant statistical output, and each qualitative entry should correspond to a documented theme or supporting excerpt.

The researcher retains the original results, preliminary AI output, corrected display, and rationale for any interpretive changes. This supports subsequent verification of the integrated analysis.

Common Mistakes When Creating Joint Displays With AI

How to Build a Defensible AI-Assisted Joint Display

Use AI primarily to make the integration process more transparent and manageable. Begin with the methodological purpose of the display rather than asking the model to invent a structure from an unorganized collection of results.

A Practical Prompt for Generating a Joint Display

Once the model produces a display, check the entries against the original analyses. Revise the integration column yourself, particularly where the model has suggested explanatory or causal relationships.

A joint display should make your reasoning inspectable. It should not substitute a model's confidence for the researcher's methodological judgment.

Before Using an AI-Generated Joint Display, Verify These Details

Frequently Asked Questions About AI-Generated Joint Displays

AI Can Build the Display, but the Integration Must Be Defensible

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.