Data quality & coverage
How much of the picture is actually filled in, and how the pipeline got there. This system links independent public records by name (precision over recall), so coverage is partial by design — this page makes that explicit rather than hiding it.
Enrichment coverage by institution
Of salaried teaching faculty, how many carry a matched federal-grant profile (NIH+NSF lead PI) and an OpenAlex research profile. Low grant coverage is expected — most faculty aren’t federal lead PIs — and reflects real funding concentration, not a gap.
| Institution | Teaching faculty | With federal grants | With research profile |
|---|---|---|---|
| Georgia Tech | 2.8K | 687 24% | 1.4K 49% |
| UGA | 5.5K | 290 5% | 2.1K 38% |
| Georgia State | 7K | 142 2% | 1.3K 19% |
| Kennesaw State | 5.4K | 34 1% | 963 18% |
Teaching-to-salary linkage
The share of instructors matched to a salary record, as measured per source. Rates differ by how names arrive (full names link far better than last-name-only PDFs).
About 84.8% of distinct instructors were linked to a salary record.
CSV term: about 81% of instructors linked. Historical PDF terms: about 57% of instructors linked.
About 91.6% of sections carry a linked instructor, and roughly 71% of distinct instructor names link to a salary record — the strongest linkage of the institutions covered, thanks to full names.
About 90% of sections carry a linked instructor, and the large majority of distinct instructor names link to a salary record.
Pipeline normalization & matching log
Every automated change the pipeline made is logged. 11K entries in total · 559 uncertain matches held for review (not auto-linked).
Matching is tuned for precision over recall— a missing link is preferred to a wrong one — so unmatched people (instructors absent from the salary export, PIs too common-named to accept, faculty who don’t publish) are omitted rather than guessed. Every analysis inherits this coverage; the per-institution “About this data” pages carry the full methodology and caveats.