Data Methodology

Data Sources

PlainEmployers combines three federal government datasets into unified employer profiles, giving job seekers and researchers a multi-dimensional view of any employer's compensation, safety, and stability record:

1. DOL H-1B Labor Condition Applications

The Department of Labor Employment and Training Administration (DOL/ETA) requires employers to file Labor Condition Applications (LCAs) when sponsoring H-1B visa workers. These filings, public under the Freedom of Information Act and downloadable directly from DOL's OFLC Performance Data page - include job titles (SOC codes), prevailing and actual wage rates (sourced from the BLS Occupational Employment and Wage Statistics (OEWS) program), work locations, and the number of workers involved. PlainEmployers processes 234,110 H-1B filings from 36,919 H-1B sponsors. Those sponsors are a subset of the live 290,575-employer multi-source corpus (H-1B + OSHA + WARN), not the full directory. While the wage data specifically covers visa-sponsored positions, it remains a reliable public proxy for an employer's disclosed professional-role pay structure.

2. OSHA Injury and Illness Data

The Occupational Safety and Health Administration (OSHA) collects annual injury and illness data from employers meeting federal reporting thresholds via its Injury Tracking Application (ITA). Enforcement records, citations, penalties, abatement deadlines, are published in OSHA's Enforcement Database (OIS). The key metric used is the DART rate (Days Away, Restricted, or Transferred) - the number of cases involving lost worktime per 100 full-time workers. OSHA publishes industry-average DART rates by NAICS code, enabling comparison of individual employer safety performance against the industry norm. A DART rate below the industry average indicates better-than-typical safety performance; a rate above suggests elevated workplace risk. NAICS classifications come from the U.S. Census Bureau NAICS reference.

3. WARN Act Notices

The Worker Adjustment and Retraining Notification (WARN) Act requires employers with 100+ employees to provide 60 days' advance notice before mass layoffs or plant closures. State workforce agencies publish these notices as public records, and the DOL/ETA WARN Act clearinghouse aggregates state-by-state contacts. PlainEmployers tracks 65,731 WARN notices spanning multiple years, sourced from state workforce-agency originals (with LayoffData.com used only as an archival backstop where direct state feeds are not available). Each notice includes the employer name, location, number of affected workers, type of event (layoff, closure, or relocation), and the effective date.

Classification standards used

PlainEmployers groups employers by industry using the U.S. Census Bureau NAICS (North American Industry Classification System) and by metropolitan area using Census/OMB Metropolitan Statistical Area (MSA) definitions. These are classification schemes applied to the three primary datasets above, not additional metrics; PlainEmployers does not currently ingest Census American Community Survey demographic data or BLS Quarterly Census of Employment and Wages industry-employment data.

Open data downloads

The compiled H-1B and WARN employer roster is free to download as CSV or JSON from /data under CC BY 4.0. National aggregates live on /statistics; the FY-over-FY filing-change extract is at /changes.csv. Downloads use the same live employer aggregates the profile pages render.

Corrections and changelog

Material corrections, source-handling changes, and methodology revisions are logged on the public changelog. To report a possible data error, use the data corrections form or see the corrections process.

Data Vintage and Update Frequency

H-1B disclosure data is released quarterly by the DOL Employment and Training Administration (ETA), typically within 30 days of the quarter's end. The live wage table currently holds DOL fiscal years FY2024 and FY2025; employer wage-trend charts compare those two years from the same employer_salaries population only (never OSHA or WARN rows in the ratio). OSHA injury and illness data is updated annually, with each release covering the prior calendar year. WARN notices are published on varying schedules by individual state workforce agencies, some states update daily while others publish monthly or quarterly. PlainEmployers refreshes each data source as new releases become available, and the database build date is displayed on the site.

When a chart plots H-1B volume against an OSHA DART rate (for example on industry pages), each axis keeps its own population and vintage. The chart is an exploratory join of employers that carry both figures, not a claim that the two federal releases share a fiscal year or category set.

Corpus placement (where an employer sits)

Every employer, metro, and industry page states where that entity sits in the PlainEmployers corpus using live SQL percentiles computed at request time:

  • Hiring: H-1B filing-count percentile among employers with at least one filing.
  • Safety: inverse DART percentile among employers with an establishment OSHA injury record (higher = safer than peers).
  • Stability: WARN notice-count percentile among all employers (higher = fewer notices than peers).
  • Wage: disclosed median wage percentile among credible per-position rows in employer_salaries.
  • Metro / industry: employer-count or total-filings percentile within the metros / industries tables.

These placement figures are derived only from the portal database, never from hardcoded fleet maxima, and each axis stays inside its own source population so a ratio never silently mixes H-1B, OSHA, and WARN vintages.

Processing Pipeline

  1. Raw data files are downloaded from each federal source, DOL/ETA bulk disclosure files for H-1B, OSHA establishment-specific injury data, and state WARN databases.
  2. Employer names are title-cased and cleaned for display (preserving standard abbreviations like Inc, Corp, and LLC), and an exact-name duplicate (the same legal name filed under a different worksite state, or in inconsistent capitalization) is merged into a single row before rankings are built. A genuinely different name, a subsidiary, a former legal name, or a related entity filed separately under DOL/OSHA/state records, is tracked as its own distinct employer; PlainEmployers does not currently perform fuzzy subsidiary or trade-name cross-linking across such variants.
  3. H-1B wage statistics are aggregated by employer: median salary, salary range (minimum to maximum disclosed wage), a per-job-title breakdown, and the metro/state footprint of each employer's filing locations.
  4. Safety grades (A through F) are calculated by comparing each employer's DART rate against the published OSHA industry average for their specific NAICS sector. The grading scale uses standard deviation bands to assign letter grades.
  5. Stability scores reflect WARN Act history, incorporating the frequency, recency, and scale of layoff notices relative to employer size where known.
  6. Industry and state-level aggregations are pre-computed for ranking pages, enabling users to compare employers within the same sector or geography.
  7. All data is loaded into a structured SQLite database serving employer profiles, state pages, industry rankings, and search functionality.

Coverage

  • Employers (multi-source corpus): 290,575 unique entities joined from H-1B, OSHA, and/or WARN records
  • H-1B sponsors: 36,919 employers with at least one Labor Condition Application (234,110 filings)
  • OSHA-record employers: 217,316 with establishment injury data on file
  • WARN notices: 65,731 mass layoff and closure notifications
  • Geographies: All 50 U.S. states and territories
  • Industries: All NAICS sectors represented in H-1B and OSHA data

Accuracy Commitment

PlainEmployers reproduces federal government data exactly as published. No subjective ratings or editorial opinions are added, grades and scores are calculated algorithmically from the underlying government data using transparent, consistent formulas. Employer name normalization is the only transformation applied, and the site clearly indicates when data is available from one, two, or all three federal sources for a given employer.

Limitations

  • H-1B salary data reflects filed wages for visa-sponsored positions only, which may not represent the full range of compensation offered to all employees at a company.
  • OSHA DART rates apply only to employers with reportable data, smaller employers or those below the reporting threshold (typically under 250 employees in most industries) may not appear in safety data.
  • Employer name normalization is imperfect and deliberately conservative, only an exact-name duplicate is merged, so a subsidiary, former legal name, or trade name that files under different text is tracked as a separate employer rather than linked to its parent.
  • WARN Act data relies on state agency publication, which varies in timeliness and completeness. Some states provide decades of history; others publish only recent years.
  • PlainEmployers is not affiliated with DOL, OSHA, or any government agency. The site aggregates publicly available government data for informational purposes only.

Download the H-1B and WARN employer roster cited on this page: employers.csv (CC BY 4.0).