Skip to content
JDTool
← Research

2026 Study · California scorecard

In California, 30 of 32 large-employer postings showed a pay range

California is the first state with a large enough sample for its own scorecard. The protocol requires at least 30 determinate postings before a state rate is quoted, and California is the first to reach it: 17 determinate California postings from the follow-up study, plus 15 newly collected California postings, for 32 in total.

30 of the 32 showed a pay range — a 93.8% pass rate (Wilson 95% confidence interval 79.9–98.3%). The 2 failures are both from the follow-up study, both the same role — a bilingual retail sales consultant posting in Fullerton — with no pay range in the posting. California requires a pay range only, so the grade is binary on pay: a full-text posting with a pay range passes, a full-text posting without one fails.

This is a first scorecard, not a representative sample of California postings. It describes 32 postings from two large employers, nothing more.

Key findings

1. 30 of 32 California postings showed a pay range

The cumulative California sample is 32 determinate postings: 30 pass, 2 fail, a 93.8% pass rate (Wilson 95% confidence interval 79.9–98.3%). The interval is wide because the sample is small — the honest reading is that California postings from large employers usually show a pay range, with real uncertainty about the exact rate.

2. The 15 newly collected postings all passed

The targeted California collection added 15 rows — 10 from a large California healthcare employer and 5 from a large national retailer — and all 15 showed a pay range. One of them, a facilities-technician posting in Tracy, was first graded a fail because no text fetch found pay; a live re-check found the range ($23.00–$42.00/hr) in the Role Details header of the official career page, and the row was corrected to a pass. Without that correction the new collection would have been 14 of 15.

3. The baseline was corrected before the scorecard was built

The published follow-up page reported California as 24 determinate (22 pass, 2 fail). A reconciliation against the graded dataset found that figure cannot be verified from the row-level record: the complete 300-row graded file shows California as 17 determinate (15 pass, 2 fail) with 11 indeterminate rows, and no mechanism in the data explains the extra 7 passes. The follow-up page was corrected to the auditable counts, and this scorecard uses the auditable baseline — 17, not 24. The headline results of the follow-up are unchanged, because the correction changes only how rows are attributed to states, not how any posting was graded.

4. One role counts three times — and the scorecard still clears the threshold without it

The same per-diem Certified Nurse Midwife role appears as three separate requisitions with identical title, pay ($102.95–$109.22/hr), posted date, facility, and description. All three were counted as separate rows, following the follow-up study's precedent of counting repeated identical-title postings as separate rows. If the three are de-duplicated to one, the cumulative sample is 30 determinate (28 pass, 2 fail), a 93.3% pass rate (Wilson 95% confidence interval 78.7–98.2%) — still exactly at the 30-row threshold. The scorecard reports the counted-as-collected figure (32) and states the de-duplicated alternative here so the reader can see both.

5. The sample is pass-heavy and concentrated

The 15 newly collected rows come from two employers: a single healthcare facility's nursing and management roles, and a national retailer's corporate and technical roles. That is what the most-recent-postings design returned for these two employers, but it is not a cross-section of California employers or industries. A third employer in a different industry would broaden the base, and the scorecard should be read as a first measurement, not a state verdict.

Breakdown

SourceDeterminatePassFail
Follow-up study, California17152
Targeted California collection15150
Cumulative California32302

Raw counts only — no rate is quoted for either employer group, because neither reaches 30 determinate rows (10 and 5). Indeterminate rows: 11 in the follow-up study's California set, 0 in the targeted collection. The de-duplicated alternative is 30 determinate (28 pass, 2 fail), 93.3% (Wilson 95% confidence interval 78.7–98.2%).

What this adds

The follow-up study measured 12 states but quoted no state rate, because no state reached 30 determinate rows. This page is the first state to clear that bar, and it exists to answer one question: what does the California sample look like once it is large enough to quote? It does not re-argue the follow-up study's case — that is on the follow-up study's page.

Grade your postings in bulk → Check a single posting →

Methodology — summary

The sample. 32 determinate California postings from large employers (1,000+ employees): 17 from the follow-up study's graded dataset and 15 from a targeted California collection on 2026-11-09 (the 10 most recently posted roles per employer from their public career pages, excluding internships, contractor roles, and internal-only transfers). Determinate means the posting text was substantial enough to grade; indeterminate and ungradable postings are excluded from the rate, not treated as passes.

The California rule. Graded only against the frozen requirement dataset (DATA_VERSION 2026-10-14). California requires a pay range for employers with 15+ employees, at least one in California; benefits and deadline are not required in California. A single starting rate is not a pay range. The full research protocol is on the methodology page.

Dual grading and the live re-check correction. An automated pass plus human review of every failing row. One row was first graded a fail because two text fetches, the search index, and an independent text extraction all found no pay; a live-browser re-check of the official career page found the range in the Role Details header, which text extraction does not capture. The row was corrected to a pass before this scorecard was built, and the correction is disclosed here rather than hidden — the same fetch-miss mechanism as the follow-up study's correction, in reverse.

Baseline correction (November 9, 2026). The follow-up study's published California row (24 determinate, 22 pass, 2 fail) was corrected to the graded dataset (17 determinate, 15 pass, 2 fail, 11 indeterminate), because the 24 could not be verified from the row-level record — the review that produced it left no audit trail. This scorecard uses the auditable 17. The correction changes only state attribution, not how any posting was graded.

The triplicate. Three identical requisitions for the same per-diem role were counted as three rows, following the follow-up study's precedent. The de-duplicated alternative (30 determinate, 28 pass, 2 fail) is stated in finding 4 above.

Anonymization. Employers are not named. The two collection employers are referred to only by role type and location. Anonymized datasets are available on request from the press contact.

Honest limits

This restraint is deliberate. A study that overclaims gets one news cycle and zero repeat citations; a study that's honest about its limits gets cited every year.

Press

For questions about the data or the method, contact jobdescriptiontool@gmail.com.

jobdescriptiontool.com is a free tool for writing job descriptions and checking them against state pay-transparency requirements, with a pay-transparency checker covering 15 U.S. jurisdictions.