This post translates a 21 July 2026 PLoS One qualitative study of 47 Saudi science teachers into a repeatable workflow and shows how Evidano can accelerate transcript coding, cross-segment comparisons, and creation of culturally contextualized materials. Read the full study at PLoS One article.
Key Takeaways
Evidano is an AI-powered qualitative data analysis platform that accelerates transcript coding, cross-segment comparisons, and the creation of culturally contextualized materials. The PLoS One study (21 Jul 2026) of 47 Saudi primary and upper-primary science teachers documents widespread teacher use of generative and adaptive AI to personalize curricula and embed local cultural resources.
- A qualitative study published 21 July 2026 observed 47 teachers in Al-Qassim and Al-Ahsa using tools like ChatGPT and DALL·E to localize science lessons.
- Teachers used AI for personalization in 46/47 cases (97.9%) and to integrate cultural texts in 40/47 cases (85.1%).
- The study used open and axial coding with inter-rater reliability κ = 0.86 and coding agreement = 95%; findings are qualitative and non-causal.
Fast take + source
This section summarizes the study and its source. A qualitative study published 21 July 2026 documents how 47 primary and upper-primary science teachers used AI tools, such as ChatGPT, DALL·E, and adaptive systems, to personalize curricula and embed local cultural resources in science lessons.
- Study: Alrashood et al., PLoS One article (published 21 Jul 2026).
- Data: n=47 teachers in Al-Qassim and Al-Ahsa; classroom observations (25 Aug–30 Oct 2025) plus semi-structured interviews.
- Methods: open and axial coding; inter-rater reliability κ = 0.86; coding agreement = 95%; qualitative, non-causal claims.
Findings snapshot
| Date / Item | Metric | Value | Source note |
|---|---|---|---|
| Publication date | Published | 21 July 2026 | PLoS One article |
| Sample | Teachers observed/interviewed | 47 (27 female, 20 male) | Private primary schools, Al-Qassim & Al-Ahsa |
| Observation window | Classroom observation dates | 25 Aug–30 Oct 2025 | 45-minute lessons, Grades 6–9 |
| AI for personalization | Teachers using AI to localize content | 46/47 (97.9%) | Theme One |
| AI + cultural texts | Teachers using AI to integrate heritage/proverbs | 40/47 (85.1%) | Theme Two |
| Interactive activities | Teachers designing AI-supported activities | 42/47 (89.4%) | Theme Three |
| Curriculum development | Teachers using AI in curriculum work | ~87.2% reported | Theme Four |
| AI assessment | Teachers using AI for culturally aware assessment | ~72.3% reported some use | Theme Six |
How the study works (methods in plain English)
This section explains the study methods in plain English. The study used a purposive sample of experienced science teachers, with classroom observations followed by semi-structured interviews, and framed analysis as inductive thematic qualitative work (open and axial coding).
- Data sources: 45-minute observed lessons (Grades 6–9) and 45–55 minute interviews with the same teachers.
- Analysis: independent coding by two analysts, 95% agreement, inter-rater observation reliability κ = 0.86.
- Tools referenced in practice: generative models (ChatGPT for explanations and prompts; DALL·E for cultural visuals), adaptive systems, and intelligent assessments.
Limitations: The study reports perception-based qualitative findings without standardized student achievement measures; results are context specific and non-causal.
So what for researchers, curriculum teams, and policy
For qualitative researchers
Researchers should use mixed-method follow-ups to test learning gains: the paper flags perception and observational evidence but not learning outcomes, so design paired qualitative and quantitative studies to test impact.
Replicate the coding approach: report inter-rater metrics (κ, percent agreement) and treat reported frequencies as 'participants who mentioned/used' rather than occurrence counts.
For curriculum & UX teams
Curriculum and UX teams should include human review when using AI-generated materials: the study shows AI can generate culturally contextualized examples and visuals quickly, but human review prevents bias or homogenization.
Embed teacher mediation into workflows: the study shows gains occur when teachers adapt AI outputs to local practice, so design teacher workflows that edit and validate AI content.
For PD & policy leads
PD and policy leads should invest in teacher AI literacy and culturally sensitive prompts, because technical tools alone are insufficient.
Require human-in-the-loop checks for assessment and curricular materials and include ethics oversight; the study had KFU-REC-2025JUN-EA000985 approval on 12 Jun 2025.
Do more, faster with Evidano (mapped to this study)
Problem: Long transcripts + manual coding
Use Evidano to ingest interview transcripts and observation notes, run thematic and frequency analysis, and export hierarchical codes and subcodes so you get themes (for example, 'cultural texts' and 'local examples') in minutes.
Problem: Multilingual/local terms & cultural markers
Use Evidano transcription and translation with custom dictionaries to preserve local terminology and cultural references, which is useful when AI outputs reference heritage practices.
Problem: Cross-segment comparisons
Use Evidano cross-segment analysis to compare regions, teacher experience levels, or gender (for example, 27 female vs 20 male teachers) and surface where cultural integration varies.
Problem: Stakeholder-ready artifacts
Use Evidano to create word clouds, co-occurrence networks, and exportable reports with quotes linked to coded segments so teachers and policymakers can validate AI-generated curricular drafts.
Security & governance
Follow secure practices for sensitive datasets: Evidano uses end-to-end encryption and private LLMs tuned for qualitative research, and the platform does not use your data to train third-party models, which matters for restricted datasets like those in this study.
Checklist: 7-step workflow to reproduce the study’s practical outputs with Evidano
This checklist shows 7 steps to go from raw audio and notes to insight and culturally contextualized materials using the study as a template.
- 1) Collect: record lessons and interviews and obtain consent, noting KFU ethics practices as a model.
- 2) Transcribe: upload audio to Evidano and apply a custom dictionary for local terms and PII redaction.
- 3) Translate (if needed): use Evidano translation with cultural term mapping.
- 4) Auto-code: run AI-assisted open and axial coding, then review and refine the codebook.
- 5) Cross-segment: run frequency and co-occurrence analyses by region, teacher experience, or classroom type.
- 6) Generate materials: use AI chat over your corpus to draft culturally contextualized lesson prompts and visuals, then export for teacher review.
- 7) Validate & iterate: triangulate with human review, run inter-rater checks, and document κ and agreement metrics.
FAQ: qualitative analysis of AI in education
When should I use qualitative analysis of AI in education?
Use qualitative analysis when you need to understand how and why teachers or learners use AI to localize or mediate learning, best combined with classroom observation and interviews as in the PLoS One study.
How do I compare segments reliably?
Compare segments reliably by reporting raw counts and proportions and by using cross-segment frequency and co-occurrence tools; the study reports examples such as 46/47 = 97.9% and includes inter-rater reliability metrics.
Is AI safe for culturally sensitive data?
AI can be used safely when you include human-in-the-loop review, secure storage, and platforms that guarantee no third-party model training; the study recommends these governance practices and Evidano provides end-to-end encryption and private LLMs.
Wrapping up & next steps
This section summarizes the study findings and suggested next steps: the PLoS One study (21 Jul 2026) documents observation-backed examples of AI enabling culturally responsive science teaching (n = 47).
Try a pilot: upload a week of lesson audio and interviews to Evidano, run an automated codebook, and export a stakeholder brief. For a trial, Try Evidano for free.