Ideas and Thoughts

This chapter documents key research ideas and conceptual developments. Entries are logged chronologically with a unique ID.

IDEA-001: AACSB International Comparison + Advanced Methodology Integration (2026-04-08)

Background

Hwang et al. (2025), published in JBMC, conducted a first-mapping study limited to the top 40 US AACSB schools. Using Content Analysis + Comparative Mapping, they derived three clusters (Explicit CT Leaders, Implicit CT Integrators, Nominal CT Adopters), but the study has two key limitations: no international comparison, and reliance on qualitative mapping alone.

Core Idea: Three Extensions

Extension 1 — Scope: US vs China vs Korea Comparison

Rationale for three-country selection:

Country AACSB Accredited Comparative Value
US ~550+ Baseline of prior research; leading AI curricula
China ~60+ (growing rapidly) State-led AI strategy; unique AI tool ecosystem (DeepSeek, etc.)
Korea ~20+ Digital talent development policy; manageable census size

Additional contextual variables (beyond existing 4-dimension coding):

  • National AI Policy Level (National AI Strategy Index)
  • AACSB accreditation tenure (years)
  • Faculty AI competency (proportion of AI-related faculty)
  • Use of domestic vs. foreign AI tools (ChatGPT vs. DeepSeek vs. domestic)

Sampling strategy:

  • US: Use existing 40 schools (Hwang et al. data reused)
  • China: 20–30 schools (AACSB-accredited census or top-ranked)
  • Korea: 15–20 schools (near full census of AACSB-accredited)

Extension 2 — Methodology: QCA + Topic Modeling + Network Analysis

2-A. QCA (Qualitative Comparative Analysis)

Purpose: “What combination of conditions produces Explicit CT Leaders?” — identifying combinatorial causality

Design:

  • Outcome: Higher CT Integration (Explicit = 1, others = 0)
  • Conditions: Country, Program Level, Curricular Theme diversity, Pedagogical Approach type (Active/Passive), Assessment Mode (Reflective/Traditional), National AI Policy Strength
  • csQCA (dichotomous) or fsQCA (continuous CT Level with calibration)

Expected output: Country-invariant sufficient conditions vs. country-specific pathways (equifinality)

2-B. Topic Modeling (STM)

Purpose: Inductively discover latent topics beyond the 5 pre-defined themes → deductive/inductive triangulation

Design:

  • Corpus: AI-related course description/syllabus text from all sampled schools
  • Recommended model: STM — prevalence ~ country + program_level directly models cross-national topic differences
  • Alternatives: LDA (easier to interpret), BERTopic (semantic similarity)
  • Implementation: R stm package

Analysis steps:

  1. Topic extraction (k = 10–20)
  2. Cross-national/program comparison of topic distributions
  3. Match extracted topics with 5 pre-defined themes → identify newly discovered themes
2-C. Network Analysis

Purpose: Visualize and compare connection patterns among themes, pedagogy, and assessment

Design:

  • Network Type 1 — Bipartite (school × coding attributes): community detection vs. existing 3 clusters
  • Network Type 2 — Co-occurrence (attribute × attribute): weighted network by co-occurrence frequency; cross-national sub-network comparison; centrality analysis for key categories
  • Network Type 3 — Country Comparison: statistical test of cross-national network structure differences via QAP

Extension 3 — Integrated Research Design

Phase 1: Data Collection & Coding
  ├── US: Reuse existing data
  ├── China: Collect AACSB data + coding
  └── Korea: Collect AACSB data + coding

Phase 2: Qualitative Comparative Mapping (extending existing method)
  ├── Derive clusters per country
  └── Cross-national cluster distribution comparison

Phase 3: Quantitative Analysis (new methods)
  ├── Topic Modeling (STM) → Latent themes + cross-national differences
  ├── Network Analysis → Connection patterns + cross-national comparison
  └── QCA → Condition combinations for Explicit CT Integration

Phase 4: Integration & Triangulation
  ├── Qualitative mapping ↔ Quantitative clustering
  ├── Pre-defined themes ↔ Topic Modeling
  └── Network communities ↔ QCA pathways

Expected Contributions

  1. First international AI–CT comparative study: Extends US-centric research to Asia
  2. Mixed methods refinement: Qualitative mapping + QCA + STM + Network Analysis triangulation
  3. Policy implications: Illuminates the pathway from national AI policy to business education curricula
  4. Methodological contribution: Combining Comparative Mapping with QCA is a novel approach in curriculum research

Practical Considerations

Issue Response
Multilingual coding Secure Korean/Chinese coders; translation verification protocol
Data accessibility Chinese university syllabi may have low public availability → website + direct contact
QCA case count fsQCA minimum 15–20 cases; verify sufficiency per country
Topic Modeling corpus Course descriptions alone may be small → include full syllabus text
Research timeline 2026-05-30 deadline → Phase 1–2 priority; Phase 3 can be a separate follow-up paper

Publication Separation Strategy (Option)

  • Paper 1: International Comparative Mapping (Phase 1–2) — extending existing methodology
  • Paper 2: Mixed Methods in-depth analysis (Phase 3–4) — QCA + STM + Network
  • Or: Single integrated paper covering Phase 1–4

Created: 2026-04-08 Related: Hwang et al. (2025) JBMC, docs/01_기존자료/기존분석/