Dror Poleg’s Data Dashboard

Pay gaps within occupations

By Dror Poleg

This chart tracks how the gap between high- and low-paid workers has changed within selected occupations. It uses weighted Census ACS microdata for full-time, year-round workers and reports percentile ratios, which describe dispersion without identifying differences in experience, geography, hours, or job specialization.

Latest observation: 2024·Expected cadence: Annual
WorkInequality
What does it show?

Pay inequality is widening inside several professions, not only between different occupations.

Methodology

Six-profession view. Weighted 90th/10th percentile ratio of annual wage income (WAGP) per occupation, ACS 1-year PUMS person records, person weights (PWGTP). Universe: full-time (35+ usual hours) year-round (50+ weeks) workers with positive wage income. Occupations matched to OCCP codes via each vintage's data dictionary labels across all three coding eras (2000-, 2010-, 2018-based; physicians combine the post-2018 physician/surgeon codes; PUMS lumps judges with lawyers under 2100). Known breaks: 'Designers' includes all designers through the 2017 vintage, graphic designers only after; 'Marketing and sales managers' includes sales managers through the 2017 vintage. ADJINC omitted (cancels within-year in a ratio). WAGP top-coding compresses the far right tail, so high-wage occupations' p90 (notably physicians) is conservative; wage income only — partnership/self-employment income (e.g. law-firm partners) is excluded and unvested equity never appears, so measured spread is a floor. Underlying percentile levels (p10-p95) in data/wagespread-levels.json. Annual, ~1-year lag; no 2020 1-year release. Directional, not a precise continuous series.

Sources