AI Qualitative Analysis of STEM Education

According to Pre-service mathematics teachers’ integration of design thinking in STEM lesson plans: Insights from a qualitative case study, published by PLOS One on November 2, 2026, 55 senior pre-service mathematics teachers produced 24 mathematics-focused STEM lesson plans, but design thinking was integrated unevenly.

The study shows why AI-enabled qualitative research can help education researchers compare lesson-plan artifacts and open-ended responses: the strongest pattern was frequent use of Define, Ideate, and Prototype phases, while Empathy and Test were less consistently implemented.

• The PLOS One study reported that Define appeared in 95.8% of the 24 lesson plans, Ideate in 83.3%, and Prototype in 91.7% in 2026.
• The PLOS One study found that Empathy appeared in only 20.8% of lesson plans, while Test appeared in 54.2% in 2026.
• The PLOS One study reported that 18 of 24 lesson plans emphasized mathematics and that 18 included creative application in 2026.
• The PLOS One study found that 43 of 55 participants believed mathematics teachers should be able to develop design thinking lesson plans in 2026.
• The PLOS One study reported 90% initial inter-rater agreement for a 25% sample of questionnaire coding and 93.1% initial agreement for lesson-plan evaluation in 2026.

The PLOS One study examined how senior pre-service mathematics teachers integrated design thinking into STEM lesson plans during a 15-week course at a state university in Türkiye during the 2021 to 2022 academic year.

According to the study by Rüveyda Karaman Dündar, participants created 24 group lesson plans and completed individual open-ended questionnaires. The lesson plans addressed topics including solar-system models, periscopes, irrigation systems, recycling bins, and mathematical optimization.

The PLOS One researchers used content analysis for the lesson plans and an inductive four-stage thematic approach for the questionnaires: familiarization, initial coding, category development, and theme construction. Two researchers independently coded a 25% questionnaire sample, representing 14 of the 55 participants across eight prompts.

The study distinguished between a design-thinking phase being mentioned and a phase being substantively implemented. According to the study’s coding protocol, robust implementation required meaningful mathematical reasoning, quantitative analysis, or evidence-based justification rather than a superficial label.

The study’s author wrote that the most challenging part was “aligning the design with other disciplines and finding suitable learning outcomes, ” quoting participant PMT 25. This statement illustrates why qualitative analysis must connect codes to the context in which participants describe their planning decisions.

AI-enabled qualitative research can help education researchers separate what participants intended to do from what lesson-plan artifacts actually contained.

The PLOS One study found a clear difference between participants’ positive views of design thinking and the uneven presence of its phases in lesson plans. Participants described benefits such as engagement, creativity, contextual learning, and mathematical modeling, while the lesson-plan analysis showed that Empathy and Test were often absent or superficial.

Researchers can use thematic analysis to identify perceived benefits and barriers, frequency analysis to count phase occurrence and recurring concerns, and cross-segment analysis to compare participant beliefs with artifact-level implementation. In this study, the key comparison is between individual questionnaire responses and group lesson plans.

The study’s participants described practical constraints as well as pedagogical value. Participant PMT 16 said, “I think design is necessary, but it should not take too much time because there is a program that needs to be covered.” Participant PMT 45 similarly described implementation in crowded classrooms as “confusing, time-consuming, and costly.” These quotations show why counts should be paired with verbatim evidence and contextual interpretation.

An AI workflow should therefore support, rather than obscure, the researcher’s audit trail. Researchers should preserve source documents, inspect representative excerpts, define codes clearly, compare segments consistently, and treat model-generated themes as analytical suggestions that require human review.

The November 2, 2026 PLOS One study shows that design thinking can make mathematics more visible in STEM lesson planning, but it also reveals gaps between intended integration and enacted practice. The study’s 24 lesson plans and 55 participant responses demonstrate the value of combining artifact analysis, thematic coding, frequency analysis, quotations, and cross-segment comparison. AI-enabled qualitative research can make that workflow faster and more traceable when researchers retain human oversight and inspect the evidence behind every theme. Researchers analyzing STEM education data can try Evidano for free to organize documents, compare qualitative segments, and generate evidence-based analyses.