Add Asia internationalization-performance thesis umbrella - #3
Conversation
Address reviewer concern #3: full-sample quadratic OLS is anchored by the extensive margin (82% of firms have FSTS=0). Re-architect the empirical strategy as a two-stage design. Stage 1 — Extensive margin (logit, N=617): Pr(exporter=1) ~ TCI + DAI + controls + sector FE TCI: OR=1.65, AME=+4.9 pp per SD (p<.001) DAI: OR=1.79, AME=+5.7 pp per SD (p=.001) Both indices strongly predict export participation — supporting the recast where TCI/DAI shift firms across the extensive margin. Stage 2 — Intensive margin (OLS, exporter subsample, N=84): M2' → fsts_c, fsts_c2 both n.s.; Lind-Mehlum p=0.78 → no inverted-U within exporters. M8' → DAI moderation IS preserved: FSTS×DAI = +1.383*, joint F = 6.32, p = .003 (matches corrected R5 row in Table 3). Turning point within exporters: 115% (out of range) → confirms the 82% turning point in full-sample M2 was an artifact of the extensive- margin anchoring. Implication for narrative: - Full-sample inverted-U should be reframed as descriptive trend. - Within exporters, no inverted-U detected; FSTS-productivity within exporters is approximately monotone. - DAI moderation is robust to restricting to exporters. Phase 1 deliverables: - tools/r3/10_extensive_margin.py Stage 1 logit - tools/r3/11_intensive_margin.py Stage 2 OLS on exporters - tools/r3/12_figures_support_aware.py Fig 2/3 with rug + decile + shading - outputs/r3/audit/extensive_margin.{json,txt} - outputs/r3/audit/intensive_margin.{json,txt} - outputs/r3/tables/T_extensive.tex - outputs/r3/tables/T_intensive.tex - outputs/r3/figures/Fig2_v2.png marginal effect of FSTS w/ rug & decile - outputs/r3/figures/Fig3_v2.png marginal effect of DAI across FSTS https://claude.ai/code/session_01Kfe4f9RT1xiN1xvffjyyoW
…4 country-years, 81957 firms)
…ngolia 2019 (83,778 firms, 36 countries, 87 country-years)
…d Mongolia (5.5) sub-archetypes, refresh Phụ lục A coverage, refresh Bảng 6.1 correlations
…6 countries, 87 country-years, 83778 firms)
…5 (5 country-years in 2025)
…th 2018-2025 period, mention 5 country-years in 2025
…025 country-years in Phụ lục A, update Mục 7 narrative
…ull panel 2009/2013/2023) — update Ch.4 tables
…ave deep dive (5.7), update Bảng 5.1, 6.1, Phụ lục A
…9/2013/2023 from panel (36 countries, 91 country-years, 95689 firms)
…g 4-Tier Verhoef + Mô hình CDCM (Context-Contingent Digital Capability Model). Tích hợp NotebookLM critique 08/05/2026 đối chiếu P3 Singapore (Mar et al., 2026 — MIR) + P4 Việt Nam (Đỗ & Phan, 2026 — JFAR): (a) Bảng 4-Tier Verhoef (2021) mapping WBES coverage; (b) Construct Boundary Statement — DAI = Foundational Digital Adoption; (c) CDCM 3 chiều xuyên quốc gia (I-P curve TP 39-46% VN vs 82% Singapore Lind-Mehlum p=0.303 không xác nhận; DAI Tier 1 điểm nghẽn vs Tier 1+2 đòn bẩy; TCI scarce advantage vs hygiene factor); (d) Institutional Transaction Costs lý giải; (e) 4 hàm ý CĐ2 (Spec 1/2 + Robustness #10 + Limitations Tier 3-4 gap). Ngăn chặn overclaiming + desk-reject Q1/Q2 IB journals.
…thêm §4.4.5.1 đối chiếu DAI Singapore (Tier 1+2: β=3,119 p=,005 dương) vs VN (Tier 1: β=-0,912 p=,043 âm 2023). Bảng 4.4.5.1 6-chiều: cấp DAI, mẫu, direct effect, interaction, 2SLS robustness (DAI co về 0,02 p=,94 ở VN — không nhân quả), cơ chế (super-linear coordination cost Brynjolfsson-McAfee 2014). Bằng chứng CDCM §2.7 file 14 v3.13. Hàm ý CĐ2: robustness check #10 cross-validation construct boundary.
…Costs xuyên quốc gia. §6 thêm Bảng 6.1.2 đối chiếu I-P curve + cơ chế DAI/TCI giữa P3 Singapore (Mar 2026 MIR) + P4 VN (Đỗ Phan 2026 JFAR) trên 4 chiều. Lý giải TCE Coase-Williamson + super-linear coordination cost Brynjolfsson-McAfee 2014: hạ tầng số yếu VN → TP thấp 39-46% inverted-U; hạ tầng Singapore hấp thụ chi phí → overload ~82% (Lind-Mehlum p=,303 không xác nhận). Hàm ý CDCM xuyên 41 nước Asia: continuum transaction cost correlate ngược mức trưởng thành hạ tầng số. Đóng góp lý thuyết MỚI cho luận án.
…economies)
Deduplicates 298 .dta uploads (many re-uploaded with UUID prefixes) into
110 canonical country-year files across 5 UN M.49 Asian regions: East,
Southeast, South, Central, and West Asia. Pacific economies excluded.
Adds data/asia/:
- raw/{Country}/{Country}_{YYYY}.dta — 110 dedup'd files (242 MB)
- pooled/asia_pooled_core.{dta,csv} — 122,858 firm-obs × 41 cols on
the standardized WBES core variable set (ID, sector, ownership,
sales, employment, exports, innovation, finance, weights)
- manifest.csv + schema.json — per-file metadata, n_obs, n_vars, md5
- scripts/build_asia_dta.py + pool_asia_dta.py — reproducible pipeline
- README.md — coverage table, variable dictionary, caveats
Encoding fallback (latin1/cp1252) handled for 22 legacy-wave files.
Variable harmonization across waves is partial — see README caveats.
Bổ sung commit
|
| Đường dẫn | Mô tả |
|---|---|
raw/{Country}/{Country}_{YYYY}.dta |
110 file deduplicated từ 298 raw uploads (nhiều bản re-upload cùng UUID khác nhau) |
pooled/asia_pooled_core.dta |
122,858 firm-obs × 41 cols — pooled trên bộ biến WBES core (id, sector, ownership, sales, employment, exports, innovation, finance, weights) |
pooled/asia_pooled_core.csv |
Bản CSV của file pooled |
manifest.csv |
1 dòng/file: country, year, n_obs, n_vars, size, md5 |
schema.json |
Per-file variable list (50 vars đầu + core hits) |
scripts/build_asia_dta.py + pool_asia_dta.py |
Pipeline reproducible |
README.md |
Coverage table, variable dictionary, caveats |
41 nền kinh tế Á qua 5 vùng UN M.49:
| Region | Files | Sample countries |
|---|---|---|
| Đông Á | 12 | China, HongKong, Korea, Mongolia, Taiwan |
| Đông Nam Á | 33 | ASEAN-6 + Brunei, Cambodia, Laos, Myanmar, TimorLeste |
| Nam Á | 24 | Afghanistan, Bangladesh, Bhutan, India, Maldives, Nepal, Pakistan, SriLanka |
| Trung Á | 16 | Kazakhstan, KyrgyzRepublic, Tajikistan, Turkmenistan, Uzbekistan |
| Tây Á | 25 | Armenia, Bahrain, Cyprus, Georgia, Iraq, Israel, Jordan, Kuwait, Lebanon, Oman, Qatar, SaudiArabia, Yemen |
Loại trừ: Các đảo Thái Bình Dương (Fiji, Samoa, Kiribati, Tonga, Solomon Islands, PNG, Vanuatu) — Oceania, không phải Asia per UN M.49.
Caveats (đã ghi trong data/asia/README.md):
- Harmonization biến giữa các wave WBES chỉ một phần — cross-wave comparison cần clean thêm
- 22 file legacy phải dùng encoding
latin1/cp1252fallback wstrict/wmedianlà within-country weights — KHÔNG so sánh chéo nước được nếu không re-stratify- Panel files (
Pakistan_panel_2007-2013-2022,Mongolia_panel_2009-2013-2019, ...) tách riêng khỏi cross-section
Commit này độc lập với phần thesis/ umbrella ban đầu của PR — phù hợp scope data layer cho 5 paper P3-P7. Nếu NCS muốn tách thành PR riêng, có thể cherry-pick 5a0ec50 sang branch mới.
Generated by Claude Code
…iles
Per NCS feedback: World Bank's "East Asia & Pacific" regional classification
includes Pacific Island economies. Adds Pacific region with 10 files / 7
economies (Fiji, Kiribati, PapuaNewGuinea, Samoa, SolomonIslands, Tonga,
Vanuatu), bringing total to 120 files across 48 economies and 124,229 firms.
- raw/{Fiji,Kiribati,PapuaNewGuinea,Samoa,SolomonIslands,Tonga,Vanuatu}/
- Re-pool: 122,858 → 124,229 firm-obs in pooled/asia_pooled_core.{dta,csv}
- Update build_asia_dta.py: Pacific region added, NOT_ASIA cleared
- README notes WB EAP rationale + filter hint for strict UN M.49 users
…eline + 4 quality scripts Hoàn thiện Chuyên đề tiến sĩ số 1 theo format CTU QĐ 1799/SH. thesis/cd1/ — package nộp Hội đồng: - 00_cd1_complete_vi.md (1,023 dòng, 140 KB) — merge file 14+15+16 thành 1 file đầy đủ với header NCS info + cross-references - 00_cd1_complete_vi.docx (92 KB · 409 paragraphs · 27 tables) — Word output với CTU format: Times New Roman 13pt, line spacing 1.2, margins L:3cm R/T/B:2cm, max 3 cấp đánh số, APA 7th references - build_docx.sh — pipeline MD → DOCX với CTU format post-processing, reproducible từ clean state - README.md — submission checklist + quality control workflow scripts/ — 5 quality control scripts: - apply_ctu_format.py (mới) — force CTU format lên DOCX output (font, margins, line spacing, heading sizes 13-16pt bold) - academic_language_check.py — hedging + voice + APA tense (English ms) - check-consistency.py — số liệu cốt lõi (pool 101.185 firms / 47 economies / 108 country-year / k=113 meta / P3 N=2,958 / P4 N=623 / P5 N=4,544) - format-apa7.py — cross-ref inline citations vs danh mục TLTK - demo_turning_point.py — synthetic demo I-P inverted-U cho P3/P5 writing_guides/ — PhD writing template: - demo_phd_rewrite_india_discussion.md — MSc → PhD voice elevation với 6 patterns (Results H1-H3, Discussion, Limitations, Conclusion) thesis/04_references_apa7.md — update v2.5 → v2.7: - Section P (Atlas 2026): Pirlea & Corso, Mahler/Wang/Weber, Agnolucci et al., Boeddu et al. G20 AI FinSupervision - Section Q (Singapore ecosystem 2025/2026): Enterprise Singapore SWITCH 2025, GovTech Singapore, IMDA ATxSG 2025, StartupBlink GSEI 2025/IBEI 2026 - Section R (Trade Restructuring): Abreha & Lopez Acevedo (2024) WB Policy Research Working Paper 10955 Quality results trên CĐ1 (v3.x latest): - check-consistency: 2 false-positive (số "101.035" trong version history metadata files 14+15, không phải claim hiện hành) - format-apa7: file 14 clean; file 15 1 false-positive parse; file 16 cần add Schumpeter (1942) vào refs - apply_ctu_format: margins L:3cm Others:2cm ✓ · TNR 13pt ✓
Commit
|
| File | Nội dung |
|---|---|
00_cd1_complete_vi.md |
1,023 dòng · 140 KB — merge từ file 14+15+16 với header NCS info |
00_cd1_complete_vi.docx |
92 KB · 409 paragraphs · 27 tables — Word output đúng format CTU |
build_docx.sh |
Pipeline MD → DOCX, reproducible từ clean state |
README.md |
Submission checklist + quality control workflow |
Format CTU QĐ 1799/SH (auto-applied)
- Font Times New Roman 13pt body (ascii + East-Asian + complex script)
- Headings: H1 16pt · H2 14pt · H3-4 13pt bold
- Line spacing 1.2 (multiple)
- Margins: lề trái 3 cm, các lề khác 2 cm
- Tables: 12pt cho compactness
- Đánh số max 3 cấp · APA 7th references
scripts/ — 5 quality control scripts
apply_ctu_format.py(mới) — force CTU format post-pandoccheck-consistency.py— số liệu cốt lõi (pool 101.185 / 47 economies / 108 c-y)format-apa7.py— cross-ref inline vs danh mục TLTKacademic_language_check.py— hedging + voice + IMRaD tense (English ms)demo_turning_point.py— synthetic demo inverted-U
thesis/04_references_apa7.md — v2.5 → v2.7
Thêm 3 Sections mới (177 entries thêm):
- Section P (Atlas 2026): WB Data360 + G20 AI FinSupervision
- Section Q (Singapore 2025/2026): SWITCH · GovTech · ATxSG · StartupBlink
- Section R (Trade Restructuring): Abreha & Lopez Acevedo 2024 WB Policy Paper
Quality results
| Check | File 14 | File 15 | File 16 | Note |
|---|---|---|---|---|
| consistency | 1 FP* | 1 FP* | ✅ | FP = số "101.035" trong version history metadata |
| format-apa7 | ✅ | 1 FP* | 1 cần fix | File 16 cần add Schumpeter (1942) vào refs |
| CTU format | ✅ | ✅ | ✅ | Margins L:3cm Others:2cm · TNR 13pt verified |
Build & sử dụng
# Build DOCX từ markdown
bash thesis/cd1/build_docx.sh
# Quality control trước nộp
python3 scripts/check-consistency.py --file thesis/cd1/00_cd1_complete_vi.md
python3 scripts/format-apa7.py --paper thesis/cd1/00_cd1_complete_vi.mdTODO trước khi nộp Hội đồng CTU (không tự động)
- Thêm
Schumpeter, J. A. (1942). Capitalism, socialism and democracyvào Section A references - Mở
00_cd1_complete_vi.docxtrong Word + F9 update mục lục - Insert 11 hình từ
thesis/figures/vào đúng vị trí (pandoc warning: PNG paths) - Verify trang bìa + lời cam đoan format đúng template CTU
- Final proofreading theo 12-điểm checklist trong
thesis/09b_vn_term_glossary.md§D
Commit độc lập với 5a0ec50 (WBES data) + c17eef2 (Pacific Islands). Package CĐ1 self-contained.
Generated by Claude Code
Hoàn thiện Chuyên đề tiến sĩ số 2 theo format CTU + add P7 multi-country empirical analysis results (84.910 firms · 47 economies). thesis/cd2/ — package nộp Hội đồng CĐ2: - 00_cd2_complete_vi.md (1.566 dòng · 182 KB) — NCS-merged 3 phần - 00_cd2_complete_vi.docx (105 KB · 522 paragraphs · 26 tables) - build_docx.sh + README.md với CTU format checklist thesis/17/18/19 cập nhật: phần 1 (intro+theory) + phần 2 (review+framework+H1-H7) + phần 3 (mô hình+data+kết luận) phiên bản mới nhất từ NCS. data/analysis/p7/ — kết quả P7 capstone (3 CSV): - p7_model_fit.csv (M0-M11, 11 mô hình, n từ 29.840-84.910) - p7_summary_focal.csv (focal coefficients M0-M11) - p7_coefs_all_models.csv (all coefs + fixed effects) Key findings P7 (cho Ch.4 luận án): - M2 quadratic: turning point 36.4% FSTS (Lind-Mehlum p<.001) — H1 inverted-U - M5 country-year FE: R² adj 0.677 — institutional heterogeneity quan trọng - M8: DAI×FSTS² interaction joint F=7.26 p=.0007 — H3 DAI moderation - M9: mgr_female β=+0.19 p<.001 — H4 gender moderation - M11 full three-way: TP 34.6%, R² adj 0.068, n=29.840 writing_guides/ — new resources: - 00_author_voice_guide.md (37 KB, 559 dòng) — PhD voice guide v2.0 với §1-§9 patterns + 7 methodological sources integration - author_guidelines_IBR_JWB_APJM_compared.pdf — journal targets comparison - cv_do_thuy_huong.docx — CV NCS scripts/ — additions: - merge_wbes_dta.py — utility merge WBES dta files - submission_checklist.py — pre-submission checklist verification Quality checks CĐ2: - check-consistency: ✅ clean - format-apa7: 10 cảnh báo, 4 ref thực sự cần add (Adner 2017, Angrist & Pischke 2009, Contractor 2012, Dosi 1982) — chi tiết trong README - apply_ctu_format: Margins L:3cm Others:2cm ✓ · TNR 13pt ✓
Resolve format-apa7.py audit warnings trên thesis/cd2/00_cd2_complete_vi.md bằng cách thêm 4 entries vào Section S mới của 04_references_apa7.md: - Adner, R. (2017). Ecosystem as structure (J Management 43(1) 39-58) DOI 10.1177/0149206316678451 — ecosystem framework cho CĐ2 §4 H7 - Angrist, J. D., & Pischke, J.-S. (2009). Mostly harmless econometrics Princeton University Press — IV identification cho CĐ2 §5 - Contractor, F. J. (2012). Why do multinational firms exist? (GSJ 2(4) 318-331) DOI 10.1111/j.2042-5805.2012.01045.x — theory update cho CĐ2 §2 - Dosi, G. (1982). Technological paradigms (Research Policy 11(3) 147-162) DOI 10.1016/0048-7333(82)90016-6 — classical innovation cho CĐ2 §2/§4 Total refs: 131 → 135 entries. Re-built thesis/cd2/00_cd2_complete_vi.docx với CTU format (margin L:3cm Others:2cm · TNR 13pt · 522 paragraphs · 26 tables) để sync. Quality status sau update: - 4 refs thực sự missing: ✅ resolved - 5 cảnh báo còn lại = false-positive parser (Hambrick & Mason 1984, Sullivan 1994, Briguglio 1995 đã có trong refs; Stock & Yogo + Đỗ & Phan là parse artifacts)
…clean, P5 CN v1.8 theory archive Sync 3 manuscripts NCS uploaded for review (commit pre-edit, full content preserved): - manuscripts/p4_singapore_en_clean_v6.md (78 KB, 264 dòng) — P4 Singapore target MIR. Khác bản với p4_singapore_en_clean.md hiện có (22 KB) — giữ versioned để NCS quyết canonical sau review actionable items. - manuscripts/p5_china_en_clean.md (119 KB, 532 dòng) — P5 China clean_6 v1.9 target IJOEM. File MỚI hoàn toàn (chưa có trong repo trước đây). Threshold-stability paper với H2a vs H2b competing hypotheses, pooled N=4,559 (2012+2024), turning points 49.4%/47.2% durable confirmed. - papers/p5-china/manuscript/parts/manuscript_v1_8_blinded_part2_theory.md (16 KB, 51 dòng) — archive phần 2 Theory của P5 v1.8 blinded. Companion cho part1_frontmatter_intro đã có. Đánh dấu là v1.8 archive — v1.9 clean_6 là canonical. Quality review summary (7 actionable items pending NCS approval): - P0 critical: fix H1/H2/H3 numbering trong P4 SG; replace "Author Citation, 2026" placeholders trong P5 CN - P1 important: add Lakens 2018 equivalence framing; move Wolfolds & Siegel justification; reframe H3 working-capital là probe - P2 polish: cross-cite P4 SG ↔ P5 CN DAI saturation parallel; archive v1.8 Đợi NCS xác nhận trước khi apply edits.
…swap P4 Singapore manuscript editorial fixes per review (commit 84b9fbe). Note: P5 China v1.9 clean_6 already incorporates P1.5 (H3 working-capital reframe as descriptive probe §2.3) and P2.6 (cross-cite via §5.5 Cross-paper DAI note) — no edits needed. Canonical management: - manuscripts/p4_singapore_en_clean.md là canonical (78 KB v6 từ upload recent — đã rename từ _v6 → no-suffix) - papers/p4-singapore/older_versions/p4_singapore_en_clean_v_old22kb.md giữ bản 22 KB cũ làm archive P0.1: Fix hypothesis numbering H1/H3/H4 → H1/H2/H3 (8 occurrences): - §1.3 contribution mention: "(H4: ..." → "(H3: ..." - §2.3.2 explanation paragraph + added 1 sentence giải thích renumber: "this paper formally states H1, H2, and H3 only" + bridge note "To avoid the appearance of a missing-H2 gap, the present paper's H2 and H3 are renumbered consecutively after H1..." - §2.3.3 "Hypothesis 3 (H3)" → "Hypothesis 2 (H2)" - §2.3.4 "Hypothesis 4 (H4)" → "Hypothesis 3 (H3)" - §2.4 conceptual paragraph: "H3 and H4 are jointly... " → "H2 and H3..." - Figure 1 caption: (H4) → (H3), (H3) → (H2) - §4.2 results: "underlying H4" → "underlying H3" - §4.5 robustness: "DAI×FSTS² interaction (H4)" → "(H3)" P1.3: Add Lakens 2018 equivalence framing trong §4.2 L-M null discussion: "The interpretive force of this null is sharpened by an equivalence-test framing (Lakens, Scheel, & Isager, 2018): with the observed FSTS range bounded essentially at the 70th percentile, the L–M test is statistically underpowered to distinguish 'no inverted-U' from 'inverted-U pushed into the unobserved upper tail.'" — formal framing thay vì narrative null. P1.4: Move Wolfolds & Siegel justification từ §4.5 (iii) lên §3.3 Method như "Identification strategy and choice of OLS as primary specification" para mới. Justifies OLS-primary choice upfront thay vì defensive ở robustness section. Original §4.5 (iii) selection diagnostic giữ nguyên. Reviews trên PR #3 đã closed status: 7/7 actionable items đã apply hoặc verified pre-applied trong v6/v1.9. Manuscripts ready cho MIR (P4 SG) + IJOEM (P5 CN) submission.
Replace manuscripts/p3_vietnam_en_clean.md với upload v10 NCS gửi review: - v10: 1429 dòng (sau cleanup) · ~110 KB · target APJM - Old: 279 dòng · 34 KB → archived to papers/p3-vietnam/older_versions/ 3 fixes applied per review (P0+P1): P0.1: §5 numbering gap fix - §5.6 Policy implications → §5.5 - §5 now: 5.1 → 5.2 → 5.3 → 5.4 → 5.5 (continuous, no gap) P0.2: Replication-package path typo - "p4_vietnam/README.md" → "p3_vietnam/README.md" in §Supplementary materials P1.3: Duplicate raw-text artifact cleanup (PDF → markdown extraction debris) - Table 1 Pooled column ASCII dump (lines 514-582 of v10) removed, consolidated into proper markdown Table 1 notes - Table 4 N-column ASCII dump after Paternoster table (lines 1167-1205 of v10) removed, consolidated into proper notes referencing Table 4 above - Net cleanup: 1541 → 1429 lines (~112 lines deleted) Quality verification: - check-consistency: ✅ clean (no inconsistencies) - format-apa7: ✅ All 135 inline citations have refs entries Key paper strengths (review highlights): - Two-margin H1 framing (H1a participation + H1b intensity) - 2SLS DAI null (β=0.018, p=.942) honest result với strong instrument F=34.6 - Three-wave Paternoster cross-wave z-tests - Type I vs Type II moderation distinction (Haans 2016) - Comprehensive robustness Panels A-K - 2015 dip explained via cohort-composition + proxy-obsolescence + macro - Cross-cite to P4 Singapore + P5 China JFAR Status: P3 Vietnam ready for APJM submission.
…ns (Phase 7.1) Tạo manuscripts/vi/ folder để chứa 3 bản dịch tiếng Việt học thuật của papers P3 Vietnam (APJM), P4 Singapore (MIR), P5 China (IJoEM). Bổ sung dossier nộp Hội đồng CTU + tham chiếu nội bộ cho NCS + HD. manuscripts/vi/README.md (~3 KB): - Scope: dịch full body + giữ References APA 7 EN + giữ EN number format - 5 conventions từ glossary v1.2 (Convention 1-5) - 3 reference docs: 09b_vn_term_glossary.md + 09_academic_writing_standards.md + writing_guides/00_author_voice_guide.md - Status table 3 papers + sync workflow Tiếp theo (Phase 7.2): 3 agents song song dịch full body 3 papers - p3_vietnam_vi_v11.md (~7.500 từ EN → VI) - p4_singapore_vi_v7.md (~7.500 từ EN → VI) - p5_china_vi_v8.md (~7.200 từ EN → VI)
…se 7.2 partial 1/3)
P4 Singapore manuscript dịch sang tiếng Việt học thuật theo glossary v1.2 +
standards v4 + voice guide v2.0. Target tạp chí Management International
Review (MIR), N=623/617 firms WBES Singapore 2023.
manuscripts/vi/p4_singapore_vi_v7.md (99 KB · 394 dòng):
- Dịch full body Abstract → §7 Limitations + Acknowledgements
- 5 conventions apply nghiêm ngặt:
* C1: thuật ngữ EN lần đầu kèm "tiếng Việt (English)" — vd "nghịch lý
bão hòa số (digital saturation paradox)", "yếu tố vệ sinh (hygiene
factors)", "nguồn lực mở rộng có điều kiện (conditional scaling)"
* C2: acronym TCI/DAI/FSTS/ICRV/WBES/OLS/HC1/IV/IMR/B-READY/RBV/IMDA
giữ nguyên + lần đầu giải thích đầy đủ
* C3: câu ≤35 từ — Abstract gốc 80 từ tách thành 4 câu mạch lạc
* C4: đoạn 3-5 câu mở bằng câu chủ đề
* C5: zero "tôi/chúng tôi"; voice trang trọng + hedge cẩn trọng
- Numbers giữ EN: β = +3.119, p = .005, R² = 0.211, [53%, 253%]
- References giữ nguyên APA 7 EN — chỉ heading dịch "Tài liệu tham khảo"
- Hypothesis statements: "Giả thuyết 1 (H1). Năng lực công nghệ (TCI)..."
P3 Vietnam (1.429 dòng) + P5 China (532 dòng) đang dịch song song
trong background — commit phần 2/3 + 3/3 khi hoàn tất.
…e 7.2 partial 2/3)
P3 Vietnam manuscript dịch sang tiếng Việt học thuật theo glossary v1.2.
Target tạp chí Asia Pacific Journal of Management (APJM), N=2.958 firms
× 3 wave WBES Việt Nam 2009/2015/2023.
manuscripts/vi/p3_vietnam_vi_v11.md (135 KB · 614 dòng):
- Title: "Tái khảo sát mối quan hệ giữa quốc tế hóa và hiệu quả hoạt
động ở thị trường mới nổi: Vai trò của năng lực công nghệ và áp dụng
số nền tảng"
- Dịch full body Abstract → §7 Conclusion + 7 metadata sections
- Conventions:
* C1 EN-VI ngoặc đầu: "biến công cụ (instrumental variable — IV)",
"lỗi thời biến đại diện (proxy obsolescence)", "học hỏi qua xuất
khẩu (learning-by-exporting)", "hiệu ứng lan tỏa tri thức (knowledge
spillovers)", "phụ thuộc giai đoạn (stage-contingent)"
* C2 acronym TCI/DAI/FSTS/WBES/OLS/HC1/IV/2SLS/PSM/M0-M8/NDTP/EMNEs
giữ EN + lần đầu giải thích
* C3 câu ≤35 từ — câu dài tách bằng dấu chấm phẩy
* C4 đoạn 3-5 câu mở câu chủ đề
* C5 zero "tôi/chúng tôi" — voice trang trọng + hedge
- 4 Bảng (descriptives, robustness panels A-K, turning points, Paternoster
z-tests) + 6 figures với caption Việt hóa, numbers giữ EN format
- Hypothesis: H1, H1a (participation), H1b (intensity), H2, H3, H4
- References: ~85 entries giữ nguyên APA 7 EN, chỉ heading "Tài liệu
tham khảo"
P5 China (532 dòng) đang dịch song song trong background — commit phần
3/3 khi hoàn tất.
….2 final 3/3)
P5 China manuscript dịch sang tiếng Việt học thuật theo glossary v1.2.
Target tạp chí International Journal of Emerging Markets (IJoEM), N=4.559
firms × 2 wave WBES Trung Quốc 2012+2024. Hoàn tất 3/3 bản dịch tiếng Việt
cho 3 papers core trong luận án.
manuscripts/vi/p5_china_vi_v8.md (149 KB · 536 dòng):
- Conventions:
* C1 EN-VI ngoặc đầu: "ổn định ngưỡng (threshold stability)", "độ bền
cấu trúc (structural durability)", "dịch chuyển môi trường
(environmental shift)", "áp lực vốn lưu động (working-capital
pressure)", "sự suy giảm hậu ngưỡng (post-threshold downturn)",
"chuỗi giá trị toàn cầu (global value chains — GVCs)", "đường cong
chữ U ngược (inverted U-shape)"
* C2 acronym TCI_full/DAI_core/FSTS/ICRV/WBES/OLS/HC1/GFC/GVCs/FDI giữ
EN + lần đầu giải thích
* C3 câu ≤35 từ — Abstract + §1 dài đã ngắt
* C4 đoạn 3-5 câu mở câu chủ đề
* C5 zero "tôi/chúng tôi" — voice trang trọng + hedge
- H2a vs H2b competing hypotheses dịch chính xác:
* H2a "Giả thuyết 2a — dịch chuyển môi trường (environmental shift)"
* H2b "Giả thuyết 2b — độ bền cấu trúc (structural durability)"
- Special technical: Paternoster (1998) z-tests → "kiểm định Paternoster
(1998) z"; Lind & Mehlum U-test → "kiểm định U của Lind & Mehlum"
- "level-shifter" → "bộ dịch chuyển mức"; "curvature moderator" → "bộ
điều tiết độ cong"; "safe operating zone" → "vùng vận hành an toàn"
- 3 Bảng + figures: caption Việt hóa, numbers giữ EN format
- References ~38 entries giữ nguyên APA 7 EN
- Cite "Do và Phan (2026 — JFAR)" và "(2026 — VEFR)" giữ format
manuscripts/vi/ inventory hoàn tất:
- README.md (4 KB)
- p3_vietnam_vi_v11.md (135 KB · 614 dòng) — APJM
- p4_singapore_vi_v7.md (99 KB · 394 dòng) — MIR
- p5_china_vi_v8.md (149 KB · 536 dòng) — IJoEM
TOTAL: 387 KB · 1,544 dòng VI tươngứng với 315 KB · 2,227 dòng EN
Phase 7 hoàn tất — 3 bản dịch tiếng Việt học thuật cho papers P3/P4/P5NCS yêu cầu chuyển 3 manuscripts EN sang tiếng Việt học thuật để bổ sung dossier nộp Hội đồng CTU. Đã hoàn tất qua 4 commits trên branch. Sản phẩm tại
|
| File | Source EN | Target | Size | Lines |
|---|---|---|---|---|
p3_vietnam_vi_v11.md |
manuscripts/p3_vietnam_en_clean.md (v11) |
APJM | 135 KB | 614 |
p4_singapore_vi_v7.md |
manuscripts/p4_singapore_en_clean.md (v7) |
MIR | 99 KB | 394 |
p5_china_vi_v8.md |
manuscripts/p5_china_en_clean.md (v8) |
IJoEM | 149 KB | 536 |
README.md |
(mới) | — | 4 KB | 65 |
Tổng: 387 KB · 1,544 dòng VI tương ứng 315 KB · 2,227 dòng EN (VI dài hơn ~10-15% do conventions ngoặc EN giải thích).
4 commits
| Commit | Nội dung |
|---|---|
e660bfc |
P7.1 — Setup folder + README (scope + conventions + status table) |
3e24771 |
P7.2.1 — P4 Singapore VI translation (1/3) |
a0805e2 |
P7.2.2 — P3 Vietnam VI translation (2/3) |
39e1eb3 |
P7.2.3 — P5 China VI translation (3/3) |
5 conventions từ glossary v1.2 áp dụng nghiêm ngặt
| # | Convention | Verification |
|---|---|---|
| C1 | Thuật ngữ EN lần đầu kèm "tiếng Việt (English)" | ✅ Tất cả 3 papers — TCI/DAI/FSTS/ICRV/WBES + thuật ngữ chuyên môn (boundary case, threshold stability, hygiene factor, etc.) |
| C2 | Acronym giữ EN + first-occurrence Vietnamese | ✅ Tất cả 3 papers |
| C3 | Câu ≤35 từ | ✅ Câu dài 50-80 từ trong Abstract đã ngắt thành 2-4 câu |
| C4 | Đoạn 3-5 câu mở câu chủ đề | ✅ Apply throughout |
| C5 | Voice trang trọng — zero "tôi/chúng tôi" | ✅ `grep -E "tôi |
Phạm vi giữ nguyên (KHÔNG dịch)
- References APA 7: Giữ nguyên tiếng Anh — chuẩn quốc tế cho submission
- Số liệu thập phân: Giữ EN format
0.07,2,958,β = 0.179,p < .001— table consistency với bản EN - Statistical symbols: β, p, F, R², SE giữ nguyên
- Citation in-text: APA 7 — "Lu và Beamish (2004)" hoặc "(Marano et al., 2016)"
- Equations: Markdown LaTeX-style giữ nguyên
- Figure paths:
giữ nguyên
Hypothesis statements dịch (sample)
P3 Vietnam: "Giả thuyết 1 (H1)... H1a (margin tham gia)... H1b (margin cường độ)..." (4 H + 2 sub-margins)
P4 Singapore: "Giả thuyết 1 (H1). Năng lực công nghệ (TCI)..." (3 H, đã apply review fix H1/H2/H3 numbering)
P5 China: "Giả thuyết 2a (H2a — dịch chuyển môi trường) vs Giả thuyết 2b (H2b — độ bền cấu trúc)" (competing hypotheses dịch chính xác)
Section headings tiếng Việt
Tất cả section headings (Abstract → References) đã dịch theo template:
- "## Tóm tắt", "## Điểm nổi bật", "## 1. Giới thiệu", "## 2. Cơ sở lý thuyết và giả thuyết"
- "## 3. Dữ liệu, biến số và chiến lược thực nghiệm" / "## 3. Phương pháp"
- "## 4. Kết quả", "## 4.5 Kiểm định độ vững"
- "## 5. Thảo luận", "## 5.1 Hàm ý lý thuyết", "## 5.2 Hàm ý quản trị"
- "## 6. Kết luận" / "## 6. Hạn chế và hướng nghiên cứu tiếp theo"
- "## Tài liệu tham khảo" (heading dịch, entries giữ EN)
Mục đích sử dụng (tóm tắt từ README)
- Bổ sung dossier nộp Hội đồng CTU — thành viên Hội đồng đọc trực tiếp tiếng Việt
- Tham chiếu nội bộ cho NCS Đỗ Thùy Hương + HD TS. Nguyễn Minh Cảnh
- KHÔNG thay thế bản EN trong submission tới tạp chí quốc tế
Hạn chế đã ghi rõ trong README
- Bản VI là internal reference — KHÔNG submit thay bản EN
- Khi bản EN có revision → cần re-sync bản VI
- Bảng giữ format EN — chỉ caption + headers Việt hóa
- Tham chiếu chính xác glossary v1.2 (68 thuật ngữ × 7 nhóm)
3 papers ready for CTU Hội đồng review + parallel với international submission tracks (APJM, MIR, IJoEM).
Generated by Claude Code
…I to baseline control + remove unpublished companion-paper refs Phase 8.1 — Apply NotebookLM critique to P3 Vietnam EN (Vietnam-only scope; P4 Singapore + P5 China untouched per NCS request, both unpublished). ## Item 2 — DAI hạ cấp từ formal hypothesis xuống baseline control §2.3 (heading + body): - Heading: "Website-based digital presence and firm performance" → "Website-based digital presence as a baseline control" - Loại bỏ formal H3 statement "DAI_z is positively associated with firm performance in Vietnam on average" - Thêm declaration: "DAI_z is NOT formalised as a primary hypothesis-bearing construct in this paper. It is retained in all specifications as a baseline digital-presence control to absorb website-ownership variance." - Thêm WBES-scope caveat 5 dòng: WBES Vietnam carries only c22b website binary; this paper does not attempt to test higher-tier (AI, ERP, dynamic digital capability) transformations. §2.4 (heading + body + H4 statement): - Heading: "Stage-contingent digital value" → "Stage-contingent digital value (exploratory probe)" - Body opens with: "Because WBES Vietnam cannot identify Tier 2-4 digital capability across waves, the present paper does not advance a primary digital-transformation claim. The following discussion frames a *secondary, exploratory probe*..." - H4 statement: "H4 (exploratory probe — not a primary hypothesis). The productivity relevance of baseline website-presence (DAI_z, Tier 1 only) may vary..." - Added: "a null or sign-shifting result on H4 should be read as construct- tier obsolescence rather than as evidence against dynamic digital capability moderation." Figure 1 (text + caption): - Text label "Foundational Digital Adoption (DAI_z)" → "Baseline website-presence (DAI_z, Tier 1 only)" - Caption: "H2, H3" → "H2"; clarify DAI is baseline control, H4 is exploratory probe on Tier-1 limited information Abstract + Highlights: - Purpose softened: "examines how TCI is associated with productivity..." (removed "and foundational digital adoption") - Findings: added durability claim "structurally durable: threshold range clusters within narrow 7-pp band despite 14 years of major change" - Originality/value: rewritten around structural durability + WBES Tier-1 caveat - Keywords: "digital adoption" → "threshold durability" - Highlights bullet 2: TCI = primary capability, DAI = baseline control - Highlights bullet 3 NEW: structural durability across WTO/GFC/CPTPP/COVID - Highlights bullet 4: descriptive Tier-1 wave-variation, not capability test §4.3 Interpretation: - "H3 is supported on average but not uniformly across waves..." → "The baseline DAI_z control reports a positive pooled level association but with substantial wave-to-wave variability... reported descriptively rather than as a hypothesis test" §4.5 PSM: - "H2 and H3 associations" → "H2 TCI association and the baseline DAI_z level association... latter best read as descriptive Tier-1 selection- corrected association rather than a capability test" ## Item 1 — Light structural-durability framing added §5 Discussion cross-paper comparison paragraph: - Replaced China cross-country comparison (Do and Phan 2026 — JFAR) with within-Vietnam cross-wave durability framing — pooled threshold 39.7% (range 39.3-46.2%) sustained across 14 years and major institutional shocks (WTO 2007, GFC 2008-09, CPTPP/EVFTA 2018-20, COVID-19 2020-22) ## Singapore + China unpublished refs REMOVED Per NCS request "không đề cập bản thảo singapore, china vì chưa công bố": - §3.2 line 425: "Cross-paper construct comparability note" mentioning Singapore AND China companion working papers → "Construct-tier scope note" focused on Vietnam WBES Tier-1 limitation only - §4.3 line 554: removed "companion working paper on Singapore" sentence - §5 line 689: replaced "Cross-paper comparison reinforces... 47.8% reported for Chinese private firms by Do and Phan (2026 — JFAR)" with "within-Vietnam cross-wave stability of the threshold itself" - References line 1388: removed orphan "Do, T. H., & Phan, A. T. (2026)" JFAR Chinese SME entry (no longer cited in body) ## Item 3 — N/A P3 Vietnam has NO working-capital hypothesis. Item 3 from NotebookLM critique (cut H3 working-capital) does not apply. Stats: 1429 → 1409 dòng (-20), 26 insertions, 46 deletions
…DAI framing Phase 8.1 follow-up — finalize alignment of Discussion + Conclusion with the demoted DAI baseline-control framing established in commit b0f3e67. §5.1 heading + opening: - "Reinterpreting digital capability in Vietnam" → "Reinterpreting the baseline website-presence indicator in Vietnam" - Opening: "foundational digital adoption... universal and temporally stable productivity premium" → "baseline website presence... not as evidence on dynamic digital capability... within constraint of Tier-1 binary website indicator" §5.2 heading + opening: - "Why the distinction between TCI and DAI matters" → "Why the distinction between TCI and baseline DAI matters" - "results strengthen the theoretical case" → "results strengthen the methodological case... not conflating the two under a single 'digital adoption' or 'digital transformation' label" §7 Conclusion (full rewrite, 3 paragraphs): - Para 1: TCI = primary capability construct; DAI = baseline digital- presence control retained only because WBES carries single Tier-1 binary indicator across waves - Para 2 NEW: explicitly states "inverted-U turning point is structurally durable: threshold range clusters within narrow ~7-pp band (39.3-46.2%) across 14 years of major institutional change including WTO accession, GFC, COVID-19, digital-payment infrastructure diffusion." Frames durability as positive finding — consistent with deep firm-level constraints (coordination capacity, contract enforcement, trade-finance access) - Para 3: TCI vs baseline DAI identification properties differ — TCI is identification-robust, baseline DAI attenuates to null under IV (Tier-1 proxy obsolescence). Explicit disclaimer: "Within the constraint of a single binary website indicator, this paper makes no claim about dynamic digital capability moderation; any such test awaits richer Tier 2-4 indicators not currently available cross-wave in WBES Vietnam" Verification: - 0 Singapore / China companion-paper references - 0 formal H3 hypothesis statements - 21 new baseline-control + structural-durability framing instances Stats: net +10 -14 lines (more concise after removing over-claim language)
Phase 8.1 hoàn tất — Apply NotebookLM critique cho P3 Vietnam ENScope (theo NCS confirm 12/05/2026): chỉ áp dụng cho P3 Vietnam EN. P4 Singapore và P5 China KHÔNG đụng đến vì 2 paper chưa công bố. 2 commits
Application của 3 NotebookLM critique items
Singapore + China unpublished refs đã XÓANCS yêu cầu: "không đề cập bản thảo singapore, china (vì 2 paper chưa công bố)"
Major textual changesAbstract (4 substantial revisions):
Highlights (5 bullets restructured):
§2.3 Heading: "Website-based digital presence and firm performance" → "Website-based digital presence as a baseline control" §2.3 Body: H3 formal statement REMOVED. New text: "DAI_z is not formalised as a primary hypothesis-bearing construct in this paper. It is retained in all specifications as a baseline digital-presence control to absorb website-ownership variance." §2.3 NEW WBES-scope caveat: "Because WBES Vietnam carries only c22b website binary comparably across waves, this paper does not attempt to test higher-tier digital-transformation mechanisms such as data-driven decision-making, AI integration, or ERP-based process digitisation. Any descriptive coefficient on DAI_z should be read as reflecting baseline digital-presence variation, not dynamic digital capability." §2.4 H4 statement: "(exploratory)" → "(exploratory probe — not a primary hypothesis)" + added "null or sign-shifting result on H4 should be read as construct-tier obsolescence rather than as evidence against dynamic digital capability moderation" §4.3 Interpretation of hypothesis tests: "H3 is supported on average but not uniformly across waves" → "The baseline DAI_z control reports a positive pooled level association... reported descriptively rather than as a hypothesis test" §4.5 Robustness PSM: "H2 and H3 associations" → "H2 TCI association and the baseline DAI_z level association... latter best read as descriptive Tier-1 selection-corrected association rather than a capability test" §5.1 Heading: "Reinterpreting digital capability in Vietnam" → "Reinterpreting the baseline website-presence indicator in Vietnam" §5.2 Heading: "Why the distinction between TCI and DAI matters" → "Why the distinction between TCI and baseline DAI matters" §7 Conclusion: Full rewrite into 3 paragraphs:
VerificationPhạm vi không làm trong Phase 8.1
Đầu raBản thảo P3 Vietnam EN giờ đây:
Generated by Claude Code |
…anism framing + V2 step-function elevation + V3 threshold capital allocation V1 — §2.4: insert "Two parallel mechanisms behind any later-wave DAI compression" paragraph BEFORE H4 statement. Frame Mechanism A (stage-dependent coordination complexity) and Mechanism B (construct obsolescence under population diffusion) as jointly operating, not mutually exclusive. Lifts construct-obsolescence reading from §5.3 defensive into §2.4 hypothesis development as a co-equal theoretical channel. V2 — Elevate participation-margin finding to step-function discovery: - Abstract Findings: replace single-sentence participation note with bold step-function framing + exporter-only regression evidence (β=-0.200, p=.660) + ±5pp density (~1.0%). - Highlights: add new bullet emphasising step function across export-participation barrier vs continuous saturation curve. - §1.1: append framing paragraph at end of Background — "step function across the export-entry threshold in a transitional economy" as headline empirical contribution. - §1.3: replace 4th contribution paragraph with sharper step-function reading vs conventional within-exporter intensity-curvature. V3 — Sharpen §5.4 Managerial implications with threshold-conditioned capital allocation: - Stage 1 (FSTS=0 or <30%): cross participation barrier; foundational website + e-payment as low-cost legitimacy signals only. - Stage 2 (FSTS ~30-46% turning-point band): tilt to TCI-deepening (foreign-licensed tech, ISO certification, process integration). - Stage 3 (FSTS >46%): Tier-1 website investment delivers negative marginal productivity; binding friction is absent Tier 2-3 process-integration. - 2015 dip OEM context: pure DAI upgrades do not loosen productivity ceiling when constraint is upstream digital governance. Cross-paper refs: zero remaining (grep audit pass). Generic transition-economy framing only; per NCS Phase 9 cross-paper removal policy. Word delta: +1,250 words vs Phase 8.1 baseline. Section structure intact (Abstract, Highlights, §1-§7, References). Version note bumped to v12 (Phase 9 NotebookLM critique applied).
…mechanism upfront framing + S2 generic Tier-1+2 construct boundary + S3 consolidated boundary conditions + cross-paper refs removed S1 — §2.3.4: Add "Three candidate mechanisms underpinning the export-contingent DAI–productivity association" block BEFORE H3 statement. Frame substitution / amplification / selection as parallel theoretical channels with explicit empirical signatures (quadratic FSTS×DAI sign distinguishes amplification from substitution; sample-restriction robustness distinguishes amplification from selection). H3 explicitly framed under amplification as focal expectation. §5.1 mechanism paragraph rewritten as empirical adjudication (CONFIRMATION) rather than post-hoc introduction. S2 — §5.1: Replace "Cross-paper DAI construct note" with generic "Tier-1 vs Tier-1+2 construct boundary and the institutional-context conditions for digital complementarity" subsection. Use generic emerging-economy framing only (no Vietnam/China naming). Tie Tier-2 transaction-enabling layer to amplification mechanism legibility; clarify that the Singapore evidence is a within-context boundary case at the advanced-economy ceiling of the institutional and digital-infrastructure spectrum. S3 — New §3.4 "Boundary conditions and identification scope": consolidate four boundary conditions at end of §3: (i) Sample-support boundary — 82% zero-export, sparse upper tail, turning point in unsupported region; (ii) Exporters-only N=84 power-bounded inference — f²=0.018, power=16%, R5 read as high-prior-strength corroboration not stand-alone identifying test; (iii) Lind–Mehlum equivalence framing — p=.303 read as theoretically informative *positive* signal under saturation hypothesis, not as methodological failure; (iv) Heckman selection sensitivity — IMR β=0.264 p=.055 mild, H3 amplification |Δ|<0.02. Frees §4 results from recurrent caveats; trims duplicated content from §4.5 (power analysis paragraph + Heckman (ii) note reduced to §3.4 pointers). Cross-paper refs removed: §1.3 second supporting contribution rewritten with within-paper R1 reference instead of "companion working papers in this series"; §3.1 DAI distribution note rewritten with generic Tier-1+2 vs Tier-1-only construct contrast (no Vietnam naming); §5.1 cross-paper DAI construct note rewritten as generic Tier-1+2 boundary discussion; §2.3.2 "(labelled H2 in the companion Vietnam and China papers in this series)" parenthetical removed. Verification: zero remaining "companion working paper" / "Do and Phan" / "JFAR" / "VEFR" / "Vietnam" / "China" name mentions in body (grep audit pass). Word delta: +1,171 words vs HEAD. Section structure intact (Abstract, §1-§7, References). Version note bumped to v8 Phase 9 (NotebookLM critique applied).
…AR/VEFR/ICBEF cross-paper refs + polish H2a/H2b active-discovery framing + DAI_core construct-boundary generic framing C1 (H2A vs H2B) — Framing already established in v8 (Abstract, Highlights, §2.2, §4.3 header, §5.1). No structural changes; minor language polishing retained. C2 (DAI demote to basic digital presence baseline control) — Framing already in v8 (§2.4 H4b not formalised, §3.2 WBES caveat). §3.2 Cross-paper DAI construct note rewritten as generic "DAI_core construct boundary in WBES China" subsection — keeps the construct-tier obsolescence reading without naming Vietnam/Singapore companion papers. C3 (H3 working-capital to §5.2 alternative mechanisms) — Section structure already in v8 (§5.2 "Alternative mechanisms explored"). No structural changes. Cross-paper refs removed throughout (per NCS Phase 9 cross-paper removal policy): - Header v1.9 history note replaced with v2.0 Phase 9 version note; - Abstract Originality/value: "extend Do and Phan 2026 JFAR + 2026 VEFR" sentence rewritten generically; - Highlights bullet 6: rewritten with within-country temporal-stability framing only; - §1 Introduction: 5 inline Do & Phan citations removed (ICBEF line 55, JFAR/VEFR lines 59, VEFR line 63, JFAR line 67, ICBEF line 67); - §2.2 §2.4: 2 inline VEFR citations removed; - §3.1 "Sample scope and N comparison with JFAR": rewritten as "Sample scope and non-response filtering"; - §3.2 "Specification note: quadratic vs cubic": rewritten without JFAR cubic-turning-point comparison; - §3.2 "Cross-paper DAI construct note": rewritten as "DAI_core construct boundary in WBES China" with generic transition-vs-advanced framing; - §4.5.1 Convergent evidence: gender parity + manager experience subsections (both relying on JFAR data) removed; restructured to focus only on within-paper DAI level-shifter + capability curvature + working-capital null moderation; - §5.1 Discussion: VEFR + ICBEF citations removed; - §5 Cross-paper DAI pattern note: rewritten as generic "DAI_core construct-context interpretation" with broader Tier-1 vs Tier-1+2 construct-context framework; - §6 Limitations Sixth + future research: VEFR + ICBEF citations removed; - References: 3 Do & Phan entries (ICBEF 2025, VEFR 2026, JFAR 2026) deleted. Verification: zero remaining "Do and Phan" / "JFAR" / "VEFR" / "ICBEF" / "companion working paper" / "companion paper" / "companion analysis" references in body or reference list (grep audit pass). Singapore name mentions: zero. Vietnam name mentions: 1 (in multi-country replication candidates list — generic, not a cross-paper companion-paper reference). Word delta: -360 words vs HEAD (removal + simplification net). Section structure intact (§1-§7 + References). Version note bumped to v2.0 Phase 9.
Phase 9 — NotebookLM critique applied to all 3 EN papers (sequential commits)Apply 3 NotebookLM audio critiques to the EN manuscripts of P3 Vietnam, P4 Singapore, P5 China + consistent cross-paper companion-paper reference removal per the policy chosen for this phase. Commits on
|
| # | Commit | Paper | Scope |
|---|---|---|---|
| 1 | aeb904d |
P3 Vietnam | V1 §2.4 two-mechanism framing + V2 step-function elevation + V3 threshold capital allocation |
| 2 | 0083dc0 |
P4 Singapore | S1 §2.3.4 three-mechanism upfront framing + S2 §5.1 generic Tier-1+2 construct boundary + S3 §3.4 consolidated boundary conditions + cross-paper refs removed |
| 3 | 4023218 |
P5 China | C1/C2/C3 polish (framing already in v8) + all JFAR / VEFR / ICBEF cross-paper refs removed throughout |
What each commit did
P3 Vietnam (aeb904d — v12)
- V1 §2.4: insert "Two parallel mechanisms behind any later-wave DAI compression" — frames stage-dependent coordination complexity and construct obsolescence under population diffusion as jointly operating before H4. Lifts construct-obsolescence reading from §5.3 defensive into §2.4 hypothesis development.
- V2 Step-function elevation: Abstract Findings + new Highlights bullet + §1.1 closing paragraph + §1.3 4th contribution rewritten — exporter-only re-fit (β=-0.200, p=.660), ~1.0 % density around turning point, "step function across the export-entry threshold in a transitional economy".
- V3 §5.4 Managerial implications rewritten with explicit threshold-conditioned capital allocation: Stage 1 (FSTS<30 % participation barrier), Stage 2 (30-46 % turning-point band, TCI-deepening), Stage 3 (>46 % Tier 2-3 process-integration); 2015 OEM context bridge.
- Word delta: +1,250 words. Zero cross-paper refs (grep audit pass).
P4 Singapore (0083dc0 — v8 Phase 9)
- S1 §2.3.4: Three candidate mechanisms (substitution / amplification / selection) framed upfront with explicit empirical signatures; H3 explicitly under amplification as focal expectation; quadratic FSTS×DAI sign as distinguishing test. §5.1 mechanism paragraph rewritten as empirical adjudication (confirmation) rather than post-hoc introduction.
- S2 §5.1: Replace "Cross-paper DAI construct note" with generic "Tier-1 vs Tier-1+2 construct boundary and the institutional-context conditions for digital complementarity" — uses generic transitional / advanced-economy framing only, no Vietnam/China naming. R1 thin-website diagnostic confirms Tier-2 transaction-enabling layer carries the empirical signal.
- S3 New §3.4 "Boundary conditions and identification scope" subsection consolidating four boundary conditions: (i) sample-support boundary + sparse upper tail; (ii) exporters-only N=84 power-bounded inference (power = 16 %); (iii) Lind–Mehlum p = .303 reframed as theoretically informative positive signal under the saturation hypothesis, not as methodological failure; (iv) Heckman IMR sensitivity diagnostic. Frees §4 results from recurrent caveats.
- Cross-paper refs removed: §1.3 + §3.1 + §5.1 + §2.3.2 — 4 locations, all rewritten with within-paper or generic language. Zero "companion working paper" / Vietnam / China name mentions remaining.
- Word delta: +1,171 words.
P5 China (4023218 — v2.0 Phase 9)
- C1 (H2A vs H2B competing), C2 (DAI_core as basic digital presence baseline control), C3 (H3 working-capital relocated to §5.2 "Alternative mechanisms explored") — framing already established in v8 from prior work; minor language polishing.
- Cross-paper refs removed throughout: Abstract Originality/value, Highlights bullet 6, §1 Introduction (5 inline Do & Phan citations), §2.2 + §2.4 (2 inline VEFR), §3.1 "Sample scope and N comparison with JFAR" rewritten as non-response filtering note, §3.2 "Specification note: quadratic vs cubic" rewritten without JFAR cubic-turning-point comparison, §3.2 "Cross-paper DAI construct note" rewritten as generic "DAI_core construct boundary in WBES China", §4.5.1 Convergent evidence (gender parity + manager experience subsections relying on JFAR data removed; restructured to within-paper null moderation only), §5.1 + §5 "Cross-paper DAI pattern note" + §6 Limitations Sixth + future research — all citations to Do & Phan papers removed.
- Reference list: 3 entries deleted (ICBEF 2025, VEFR 2026, JFAR 2026).
- Verification: zero remaining "Do and Phan" / "JFAR" / "VEFR" / "ICBEF" / "companion working paper" / "companion analysis" references; one generic Vietnam name mention remains only in the multi-country replication candidates list (acceptable generic reference, not a cross-paper companion-paper ref).
- Word delta: -360 words (removal-net).
Cross-paper consistency check
Triangulate audit (all three papers):
- P3 Vietnam: zero cross-paper companion-paper references (since Phase 8.1 + reconfirmed Phase 9).
- P4 Singapore: zero cross-paper companion-paper references (this commit).
- P5 China: zero cross-paper companion-paper references (this commit).
Each paper now stands alone editorially. The Tier-1 vs Tier-1+2 DAI construct-boundary argument is preserved across papers but framed generically (transition-economy vs advanced-economy context), without naming companion working papers.
Out of scope (deferred to next phase)
- Re-translate revised EN sections into VI counterparts (
p3_vietnam_vi.md,p4_singapore_vi.md,p5_china_vi.md) — defer to Phase 10. - No statistical re-analyses (numbers preserved from prior commits and uploaded baselines).
- No Python figure script or docx recompilation changes.
NCS rà soát 3 commits sequentially trên branch và yêu cầu điều chỉnh thêm nếu cần. EN settle xong sẽ chuyển sang Phase 10 (VI re-translation).
Generated by Claude Code
…e to VI (v11 → v12)
Apply selective VI translation of EN Phase 9 edits to keep VI manuscript in sync with EN canonical. Rename file v11 → v12 to match EN version numbering.
V1 — §2.4 "Hai cơ chế song song đằng sau bất kỳ sự nén DAI ở các wave muộn": insert paragraph framing Mechanism A (stage-dependent coordination complexity) and Mechanism B (construct obsolescence under population diffusion) as jointly operating channels before H4 statement. Mirror of EN §2.4 two-mechanism framing.
V2 — Step-function elevation translated to 4 locations:
- Abstract Findings: replace single-sentence participation note with bold step-function framing including exporter-only β = -0,200 evidence and ~1,0 % density;
- Highlights: new bullet about step function across export-participation barrier;
- §1.1 closing paragraph: "hàm bậc thang xuyên ngưỡng gia nhập xuất khẩu" as headline empirical contribution;
- §1.3 4th contribution: rewrite with step-function reading + participation-margin friction framing.
V3 — §5.4 Hàm ý quản trị rewritten with threshold-conditioned capital-allocation map (bản đồ phân bổ vốn có điều kiện theo ngưỡng):
- Giai đoạn 1 (FSTS < ~30 %): cross participation barrier;
- Giai đoạn 2 (FSTS ~30-46 % turning-point band): tilt to TCI-deepening;
- Giai đoạn 3 (FSTS > 46 %): Tier 2-3 process-integration;
- 2015 OEM context bridge ("tờ rơi số" framing).
Word delta: +1,922 words vs v11 baseline (21,880 → 23,802). Translation conventions per glossary v1.2 retained: EN-VI bracketing first occurrence, acronyms preserved, statistical symbols unchanged.
Version note bumped to v12 Phase 10 (12/05/2026).
…que to VI (v7 → v8)
Apply selective VI translation of EN Phase 9 edits to keep VI manuscript in sync with EN canonical. Rename file v7 → v8 to match EN version numbering.
S1 — §2.3.4 "Ba cơ chế ứng viên đằng sau liên hệ DAI–năng suất có điều kiện theo xuất khẩu": insert paragraph framing substitution / amplification / selection mechanisms with explicit empirical signatures (quadratic FSTS×DAI sign distinguishes amplification from substitution). H3 framed under amplification as focal expectation. §5.1 mechanism paragraph rewritten as empirical adjudication ("Phán xử thực nghiệm giữa ba cơ chế ứng viên") rather than post-hoc introduction.
S2 — §5.1 cross-paper DAI construct note replaced with "Biên giới construct Tầng 1 vs Tầng 1+2 và điều kiện bối cảnh thể chế cho bổ trợ số" — generic transition-economy / advanced-economy framing without Vietnam/China naming. §3.1 DAI distribution note rewritten with generic emerging-market framing.
S3 — New §3.4 "Điều kiện biên giới và phạm vi nhận dạng" subsection consolidating four boundary conditions:
- (i) Biên giới hỗ trợ mẫu — 82 % zero-export, sparse upper tail;
- (ii) Nhóm con chỉ-doanh-nghiệp-xuất-khẩu N=84 power-bounded inference (f²=.018, power=16%);
- (iii) Khung tương đương Lind–Mehlum — p=.303 đọc là tín hiệu lý thuyết dương dưới khung bão hòa, không phải sự thất bại phương pháp luận;
- (iv) Nhạy lựa chọn Heckman + sự vắng mặt của ràng buộc loại trừ.
§4.5 power analysis + Heckman (ii) duplicates trimmed to pointers (§3.4(ii), §3.4(iv)).
Cross-paper refs removed at §1.3 (companion working papers → §4.5 R1 within-paper reference) + §3.1 (Vietnam DAI distribution comparison → generic emerging-market framing) + §5.1 (cross-paper DAI construct note → Tier-1 vs Tier-1+2 boundary) + §2.3.2 ("labelled H2 in companion Vietnam and China papers" parenthetical removed).
Word delta: +1,664 words vs v7 baseline (15,716 → 17,380). Translation conventions per glossary v1.2 retained: EN-VI bracketing first occurrence, acronyms preserved, statistical symbols unchanged.
Version note bumped to v8 Phase 10 (12/05/2026).
…to VI (v8 → v9) — remove all JFAR/VEFR/ICBEF cross-paper refs Apply selective VI translation of EN Phase 9 edits to keep VI manuscript in sync with EN canonical. Rename file v8 → v9 to match EN version numbering. C1 (H2A vs H2B active-discovery) + C2 (DAI_core baseline control) + C3 (H3 working-capital relocated to §5.2 "Các cơ chế thay thế đã thử nghiệm") — framing already established in v8 from prior work; no structural changes. Cross-paper refs removed throughout (mirror EN side): - Header history note replaced with v2.0 Phase 10 version note; - Abstract Originality/value: "extend Do và Phan 2026 JFAR + 2026 VEFR" sentence rewritten generically; - Highlights bullet 6: rewritten with within-country temporal-stability framing only; - §1 Introduction: 5 inline Do và Phan citations removed (ICBEF, JFAR/VEFR multi-paragraph references); - §2.2 + §2.4: 2 inline VEFR citations removed; - §3.1 "Phạm vi mẫu và so sánh N với JFAR" rewritten as non-response filtering note; - §3.2 "Ghi chú đặc tả: bậc hai vs bậc ba" rewritten without JFAR cubic-turning-point comparison; - §3.2 "Ghi chú cấu trúc DAI xuyên bài báo" rewritten as generic "Biên giới construct DAI_core trong WBES Trung Quốc"; - §4.5.1: gender parity + manager experience subsections (both relying on JFAR data) removed; restructured to within-paper null moderation only; - §5.1 Discussion: VEFR + ICBEF citations removed; - §5 "Ghi chú mô hình DAI xuyên bài báo" rewritten as generic "Diễn giải bối cảnh-construct DAI_core"; - §6 Limitations Sixth + future research: VEFR + ICBEF citations removed; - Reference list: 3 Do và Phan entries (ICBEF 2025, VEFR 2026, JFAR 2026) deleted. Word delta: -340 words vs v8 baseline (23,270 → 22,930). Translation conventions per glossary v1.2 retained: EN-VI bracketing first occurrence, acronyms preserved, statistical symbols unchanged. Version note bumped to v2.0 Phase 10 (12/05/2026).
…s (v11→v12, v7→v8, v8→v9)
Phase 10 — VI re-translation cho 3 papers (selective Phase 9 changes)Apply selective Vietnamese translation cho Phase 9 EN edits để giữ VI manuscripts sync với EN canonical. Mỗi VI file được rename v{N}→v{N+1} match với EN version numbering. Commits trên
|
| # | Commit | File | Scope |
|---|---|---|---|
| 4 | 2fe8363 |
p3_vietnam_vi_v11.md → v12.md |
V1 §2.4 + V2 4 locations + V3 §5.4 (+1,922 words) |
| 5 | f6200cc |
p4_singapore_vi_v7.md → v8.md |
S1 §2.3.4 + S2 §5.1 + S3 §3.4 + cross-paper refs removed (+1,664 words) |
| 6 | 139c501 |
p5_china_vi_v8.md → v9.md |
JFAR/VEFR/ICBEF refs removed (-340 words) |
| 7 | 9cb1350 |
manuscripts/vi/README.md |
Update version refs table |
Per-paper VI Phase 10 details
P3 Vietnam VI v12 (2fe8363)
- V1 §2.4: dịch "Hai cơ chế song song đằng sau bất kỳ sự nén DAI ở các wave muộn" — Mechanism A (phức tạp phối hợp phụ thuộc giai đoạn) + Mechanism B (sự lỗi thời của biến đại diện do phổ cập dân số) song song.
- V2 step-function elevation 4 locations: Tóm tắt Findings với "hàm bậc thang xuyên rào cản tham gia xuất khẩu" + Highlights bullet mới + §1.1 closing paragraph + §1.3 4th contribution rewrite.
- V3 §5.4 Hàm ý quản trị: "bản đồ phân bổ vốn có điều kiện theo ngưỡng" với 3 giai đoạn (<30 % participation barrier, 30-46 % TCI-deepening, >46 % Tier 2-3) + 2015 OEM "tờ rơi số" bridge.
P4 Singapore VI v8 (f6200cc)
- S1 §2.3.4: "Ba cơ chế ứng viên" — substitution / amplification / selection upfront framing với H3 dưới expectation khuếch đại.
- S2 §5.1: thay "Ghi chú construct DAI xuyên bài báo" bằng "Biên giới construct Tầng 1 vs Tầng 1+2" generic framing.
- S3 new §3.4 "Điều kiện biên giới và phạm vi nhận dạng" với 4 boundary conditions: (i) sample-support, (ii) N=84 power-bounded, (iii) Lind–Mehlum equivalence như tín hiệu lý thuyết dương, (iv) Heckman.
- §5.1 mechanism para rewritten as "Phán xử thực nghiệm" confirmation.
- Cross-paper refs removed: §1.3 + §3.1 + §5.1 + §2.3.2 — 4 locations.
- §4.5 power + Heckman (ii) duplicates trimmed to §3.4 pointers.
P5 China VI v9 (139c501)
- C1/C2/C3 framing already in v8 from prior Phase 7.2 work; no structural changes.
- Cross-paper refs removed: header history note, Abstract Originality/value, Highlights bullet 6, §1 (5 inline), §2.2 + §2.4 (2 inline), §3.1 "Phạm vi mẫu và so sánh N với JFAR" rewrite, §3.2 "Ghi chú đặc tả bậc hai vs bậc ba" rewrite, §3.2 "Ghi chú cấu trúc DAI xuyên bài báo" → "Biên giới construct DAI_core trong WBES Trung Quốc" generic, §4.5.1 gender parity + manager experience subsections (relying on JFAR data) removed, §5.1 + §5 "Ghi chú mô hình DAI xuyên bài báo" → "Diễn giải bối cảnh-construct DAI_core" generic, §6 Limitations Sixth + future research, References list 3 entries deleted.
Cross-paper consistency check
Triangulate audit pass cho cả 3 VI files:
- VI v12 (Vietnam): zero "đồng hành" / "working paper đồng hành" / "Do và Phan" / JFAR / VEFR / ICBEF references.
- VI v8 (Singapore): zero references (only meta-mention trong header version note).
- VI v9 (China): zero references (only meta-mention trong header version note).
Translation conventions per glossary v1.2 retained xuyên cả 3 files: EN-VI bracketing first occurrence, acronyms TCI/DAI/FSTS/WBES/OLS/HC1 preserved, statistical symbols β/p/F/R²/SE unchanged.
Final word counts (Phase 10)
| Paper | EN current | VI v(N+1) | VI/EN ratio |
|---|---|---|---|
| P3 Vietnam | 18,611 | 23,802 | 128% |
| P4 Singapore | 12,360 | 17,380 | 141% |
| P5 China | 15,885 | 22,930 | 144% |
(VI typically 25-45% longer than EN do văn phong khoa học tiếng Việt + EN-VI bracketing first occurrence + Vietnamese punctuation patterns.)
Status submission (sau Phase 10)
| Paper | EN canonical | VI canonical | Submission status |
|---|---|---|---|
| P3 Vietnam | v12 Phase 9 ✅ | v12 Phase 10 ✅ | Ready for APJM |
| P4 Singapore | v8 Phase 9 ✅ | v8 Phase 10 ✅ | Ready for MIR |
| P5 China | v2.0 Phase 9 ✅ | v9 Phase 10 ✅ | Ready for IJoEM |
EN + VI in sync across cả 3 papers. Hết Phase 10.
Generated by Claude Code
Tóm tắt
Thêm thư mục umbrella
/thesis/chứa khung luận án tiến sĩ của NCS Đỗ Thùy Hương — đề tài "Quốc tế hóa và hiệu quả hoạt động kinh doanh của doanh nghiệp ở các quốc gia châu Á" (Internationalization and Firm Performance in Asian Countries), trình bày bằng tiếng Việt theo chuẩn CTU 5 chương, với các bản thảo bài báo phát triển từ luận án viết bằng tiếng Anh.Nội dung 4 file
thesis/README.md— bản đồ thư mục, status 4 công trình đã công bố và 5 paper P3–P7, tham chiếu các nhánh chứa code/manuscript.thesis/thesis_outline_vi.md— dàn ý chi tiết 5 chương theo chuẩn trình bày CTU (font Times New Roman 13, lề trái 3 cm, APA 7th).thesis/research_program_framing.md— định vị P3–P7 với citation APA 7th xuyên suốt; phần I tiếng Việt cho Chương 1 luận án, phần II tiếng Anh cho cover letter/abstract; danh mục References APA 7th đầy đủ.thesis/related_publications_vi.md— danh mục 4 công trình đã công bố, status 5 paper trong luận án, các câu chuẩn cho Chương 1.Kiến trúc 5 paper (tránh trùng lấp)
4 công trình đã công bố
Tham chiếu các nhánh chứa code/manuscript
claude/verify-stata-results-So3mfscripts/P3_MIR_Singapore_Analysis.doclaude/fix-integrity-update-scope-d7rUxmanuscripts/p4-jwb-vietnam/claude/verify-stata-results-So3mfscripts/P4_JWB_Vietnam_Analysis.doclaude/verify-stata-results-So3mfscripts/P5_APJM_China_Analysis.doTest plan
claude/asia-internationalization-performance-2cYgOđược tạo từmain/thesis/mainPhạm vi không làm
claude/fix-integrity-update-scope-d7rUx,claude/verify-stata-results-So3mf,claude/add-claude-documentation-4fazm).https://claude.ai/code/session_018iXUs9HFFxTe2WVA68zxyn
Generated by Claude Code