Patent comparisons and expanded technical review
This extension uses the retained USPTO/PatentsView release through December 2025, acquired on 27 October 2026. It compares nonwithdrawn utility grant events in 2015–2025. Pregrant documents are not pooled with grants. All seven strategic areas remain present; their CPC screens remain provisional analyst definitions.
The national eligible set contains 3,545,902 utility grants. Of these, 1,716,604 have at least one inventor in US50+DC, representing 1,613,139.476 inventor-fractional grants. Arkansas reconciles to 4,653 whole grants and 3,485.499 fractional grants. A fractional grant divides the region's inventor sequences by all named inventor sequences. Unknown, foreign and territory locations retain their denominator weight. Regional whole counts overlap and must not be added.
What the comparison adds
These ratios compare each provisional technology screen's share of regional patent activity with its share of national activity, using the same counting rules. A ratio of two means twice the share within the patent record. It does not mean twice the market opportunity, innovation performance or readiness. No compatible population denominator was retained, so no per-capita comparison is claimed.
| Area | Current CPC narrow ratio | Current CPC broad ratio | At-issue narrow ratio | At-issue broad ratio |
|---|---|---|---|---|
| ARA1 · Population Health Innovations and Clinical Research | 0.57 | 0.74 | 0.56 | 0.77 |
| ARA2 · Power Electronics, Advanced Packaging, and Grid Management Systems | 0.69 | 0.65 | 1.01 | 0.91 |
| ARA3 · Next Generation Transportation and Logistics Systems | 1.80 | 1.70 | 1.75 | 1.74 |
| ARA4 · Materials Engineering Applications | 0.77 | 0.83 | 0.78 | 0.86 |
| ARA5 · Precision Agriculture | 3.67 | 1.55 | 3.75 | 1.61 |
| ARA6 · Food Production and Integrative Health | 4.02 | 1.91 | 4.01 | 2.06 |
| ARA7 · Supply Chain, Retail, and Consumer Analytics | 3.31 | 3.35 | 3.53 | 3.56 |
The food, transport/logistics and supply/retail screens have above-national shares under both boundaries and classification views. This supports focused technical/partner questions, subject to organizational concentration and scope review. The broad agriculture screen and narrow A01 screen have materially different ratios, and A01 includes substantial non-precision activity. Power's narrow ratio changes from about 1.01 at issue to 0.69 under current classifications; classification movement alone can change the apparent comparative position. Lower ratios for a provisional screen do not establish absence of a capability or a reason to reject a strategic area.
Equal five-year grant-event windows (2015–2019 and 2020–2024) are supplied separately with annual panels. They preserve the same field and geography definitions and avoid comparing periods of different length. Grant-event movement is not first-priority inventive momentum; family/priority analysis is a separate companion workstream.
Expanded primary-document review
The new selection has 84 probability-selected area/document pairs: four from each area's narrow, broad-only and outside-broad strata. Seven separately labeled purposeful keyword challenges test plausible missed scope. There are 90 unique primary patent documents across 91 pairs. Abstracts and first independent claims were retained with hashes and reviewed by the Codex analyst under an explicit technical-relevance protocol. This is not independent external domain validation. The small outside-basket sample has low power to detect missed matches. The table shows selected-set yields only; it is not a population accuracy or recall estimate.
| Area | Narrow: direct / reviewed | Broad-only: direct / reviewed | Outside broad: direct / reviewed | Purposeful outside challenge |
|---|---|---|---|---|
| ARA1 | 3 / 4 | 0 / 4 | 0 / 4 | direct |
| ARA2 | 3 / 4 | 0 / 4 | 0 / 4 | adjacent |
| ARA3 | 4 / 4 | 4 / 4 | 0 / 4 | direct |
| ARA4 | 3 / 4 | 3 / 4 | 0 / 4 | direct |
| ARA5 | 0 / 4 | 0 / 4 | 0 / 4 | outside |
| ARA6 | 4 / 4 | 0 / 4 | 0 / 4 | direct |
| ARA7 | 4 / 4 | 1 / 4 | 0 / 4 | outside |
Adjacent and uncertain decisions are retained separately in the full review table. Zero direct matches in a tiny sample does not establish zero relevant patents in its population.
Concrete scope corrections include an asphalt-aging method inside A61, organ cold storage and ordinary crop varieties inside A01, industrial biotechnology/cancer therapy inside broad food C12, and healthcare-demand forecasting outside the retail/supply-sector question. Purposeful challenges identify medication-adherence records outside health's basket, indoor drone navigation outside transport's basket, core-shell nanoparticles outside materials' basket, and food-safety risk tracking outside food's basket. These examples support better inclusion/exclusion rules without silently redefining the screens to improve comparative results.
Report implications
- Strengthen the applied-food and supply/logistics discussion with comparable patent-record composition, using carefully labeled supporting notes rather than a seven-area league table.
- Report comparative strength with both classification vintages and boundaries; power and agriculture demonstrate why a single code screen can mislead.
- Retain explicit materials and human-health applications even where generic code baskets miss cross-sector technical examples.
- Keep AI/ML, analytics and trusted systems as shared methods. A healthcare-demand forecast illustrates analytics beyond the retail/supply area.
- Use technical examples to sharpen partner questions about users, access and operating results. Patent activity does not establish commercialization, demand or accepted partner responsibilities.
Reproduction and evidence
Run PYTHONPATH=src python3 -m ara_research.patent_benchmarks --output data/analysis/NEW_ID, then --phase download --output ..., followed by --phase finalize --output .... Completed runs refuse overwrite. Source catalog/file releases, canonical source URLs, input hashes, code/definition hashes, primary text hashes and the full output index are in the manifest and validation record. The review protocol and decisions are copied into this run.
Supporting tables: inventor-geography-grant-panel.csv; provisional-technology-share-panel.csv; equal-window-screen-comparisons.csv; expanded-technical-review.csv; selected-review-yields.csv. Raw HTML and internal national identifiers remain internal, outside website export.