Qualitative Analysis of Corporate DEI Disclosures

The Conference Board’s Aug 20, 2025 report shows US public companies materially reframing DEI language, narrowing workforce and board demographic disclosures, and embedding oversight inside governance to reduce legal risk. This post gives UX researchers, policy analysts, and corporate reporting teams a step‑by‑step qualitative analysis workflow to reproduce those findings, spot industry differences (e.g., IT vs. health care), and produce stakeholder-ready outputs faster using Evidano (see www.evidano.com). Read the original report at www.corpgov.law.harvard.edu/2025/08/20/dei-in-transition-2025-corporate-diversity-disclosure-trends/ for source tables; below we translate its numbers into an actionable qualitative research plan.

Fast take + source

What happened: In 2025 many large US public firms cut public DEI language and some demographic disclosures while increasing formal board committee oversight of DEI, driven by legal rulings (notably the 2023 Supreme Court affirmative action decision and a Dec 2024 Fifth Circuit decision affecting Nasdaq rules) and shifting investor signals.

Primary source: Conference Board summary posted Aug 20, 2025 (original: www.corpgov.law.harvard.edu/2025/08/20/dei-in-transition-2025-corporate-diversity-disclosure-trends/).
Why it matters to researchers: language shifts and selective disclosure create a heterogeneous corpus where qualitative coding, cross-segment comparison, and timeline analysis reveal governance intent and stakeholder risk.
Quick win: use an AI-enabled platform to ingest 10‑K/10‑Q, proxy statements, and sustainability reports, map thematic changes (e.g., removal of “equity”), and quantify disclosure shifts by industry.

Findings snapshot (key numbers)

Date / Metric	Value	Implication	Source
Aug 20, 2025, firms reviewed	Top 100 US public companies (Form 10-K review)	Sample for framing and language-change analysis	Conference Board (Aug 2025)
Material messaging changes in 2025	53% of firms	Widespread reframing of DEI language	Conference Board (Aug 2025)
Removal of the word “equity”	≈33% of firms	Term viewed as politically sensitive; replaced or omitted	Conference Board (Aug 2025)
Removed “DEI” acronym	11% of firms	Softening of explicit DEI labels	Conference Board (Aug 2025)
Removed DEI subsection / councils	16% / 10% respectively	Less visible public structuring of DEI	Conference Board (Aug 2025)
Decline in board gender disclosure (S&P 500)	−30 percentage points	Sharp pullback in board-level demographic transparency	Conference Board (Aug 2025)
Russell 3000, no board gender data	47.6%	Nearly half disclose no board gender figures	Conference Board (Aug 2025)
Russell 3000, no racial/ethnic board data	54.7%	Majority withhold racial/ethnic board demographics	Conference Board (Aug 2025)

What happened, plain English

Between 2024 and 2025 companies adjusted public DEI communications to reduce legal exposure and reputational risk: many removed specific words (e.g., “equity”), dropped DEI subsections, or reframed goals as inclusion, belonging, or broader human capital priorities.

Drivers: 2023 Supreme Court decision on affirmative action, Dec 2024 Fifth Circuit ruling affecting Nasdaq diversity rules, shifts in investor stewardship (some asset managers eased firm targets).
Disclosure pattern: external messaging became more cautious even as many firms maintained internal data collection; board committee reporting on oversight increased as a defensive governance signal.
Sector variance: IT firms led the pullback (≈78% made changes in the S&P 100), while healthcare showed more continuity (≈31% changed), reflecting stakeholder/regulatory pressures in patient-facing sectors.

How to run a qualitative analysis of corporate DEI disclosures

Step 1; Build the corpus

Collect 10-Ks, proxy statements, sustainability/ESG reports, and DEI webpages for your target universe and years (e.g., 2023–2025). Use web scraping to pull press releases and corporate statements for timeline context.

Evidano tip: import PDFs and URLs directly into www.evidano.com and enable automatic transcription/OCR for image PDFs; set a custom dictionary for firm names and legal terms.

Step 2; Preprocess & normalize

Standardize tokens (e.g., DEI, diversity, inclusion, equity) and normalize variations (’equity’ vs. ‘pay equity’). Flag sections by document type (Management discussion, Governance, Compensation).

Evidano tip: use translation and dictionary features to harmonize terminology across filings and to redact PII where needed.

Step 3; Thematic coding (mixed method)

Create a codebook that captures: explicit DEI terms, reframed language (’inclusion’, ’belonging’), governance signals (committee mentions), disclosure actions (removed/added subsections), and compensation linkages.

Evidano tip: import a seed codebook and run AI-assisted coding to label quotes, then validate with a 200‑item sample and adjust thresholds; Evidano produces theme frequencies and co-occurrence networks automatically.

Step 4; Segmentation and cross-tab analysis

Segment by index membership (S&P 500 vs Russell 3000), industry, and company size. Compare pre/post 2025 language usage, and compute percentage-point changes in disclosure rates (e.g., female director reporting).

Evidano tip: run cross‑segment analyses to surface where language change clusters (e.g., IT vs health care) and export charts for stakeholder reporting.

Step 5; Governance mapping and narrative extraction

Map where DEI oversight moved (nominating/gov committee vs compensation vs full board) and extract representative quotes explaining the change. Use timeline analysis to align shifts with regulatory events (Dec 2024, 2023).

Evidano tip: extract clickable quotes and build a narrative report with hierarchical codes → subcodes for governance vs. workforce vs. compensation themes.

7‑step reproducible checklist

Run this checklist to reproduce the Conference Board-style analysis:

1) Define universe and years (e.g., S&P 500, 2023–2025).
2) Ingest filings, proxies, and DEI pages into Evidano; enable OCR and custom dictionary.
3) Normalize terminology and create a seed codebook for DEI language and governance mentions.
4) Auto-code with human validation (200–300 checks) to set precision thresholds.
5) Produce theme frequency tables and co-occurrence networks; generate industry segment comparisons.
6) Extract governance oversight changes and quantify percentage-point disclosure shifts.
7) Export visuals and a 1‑page stakeholder brief for legal/IR/HR teams.

FAQ: qualitative analysis of corporate DEI disclosures

Q: How do I compare disclosure rates reliably across companies?

A: Normalize by document type and year, compute presence/absence flags (e.g., DEI subsection present = 1), and report percentage-point changes with confidence intervals where sample size allows.

Q: What about legal sensitivity and ethics?

A: This is a research and reporting exercise, not legal advice. Preserve consent and confidentiality for interview data; when handling sensitive demographic records, follow internal legal guidance and redact PII. Evidano encrypts data and does not use it to train external models.

Q: Can we reproduce the Conference Board metrics?

A: Yes, by assembling the same document universe and applying consistent coding rules for terms, subsections, and committee mentions. Automated ingestion + AI-assisted coding speeds the process and enhances consistency.

Wrapping up: next steps and CTA

If you need to map how DEI language, board transparency, and compensation linkages changed in 2025, use the checklist above to build a reproducible qualitative study and surface governance risk signals.

Try a pilot: ingest a 50‑company sample, run the seed codebook, and generate a dashboard that compares disclosure changes by industry.
Start on www.evidano.com to import documents, run AI-assisted coding, and export stakeholder-ready visuals in hours instead of weeks.

Keep reading

Browse all articles