Physician compensation data methodology
The compensation pages, the Offer Check and the contributed benchmark each read a different kind of number. They measure different things, so they are shown side by side with their definitions and never averaged into one figure. This page says what each dataset is, where it comes from, what it cannot tell you, and which figure on the site it powers.
Definition. Self-reported total compensation: base salary plus productivity, quality and other cash pay, as each survey’s respondents reported it. Neither publisher defines the term further in its methodology, and neither publishes a median — both publish an average.
What we add. Where both 2026 reports cover the same specialty for the same income year, we show their arithmetic midpoint, labelled as ours (27 of 53 specialties today). It is a convenience figure for readers who want one number; the two published figures beside it are the evidence. The rule suppresses a midpoint when the sources diverge by more than 25% — no pair currently does — and when the surveys measure different constructs (Oncology, where one survey pools it with Hematology). The year-over-year change is our arithmetic from Doximity’s two tables.
Limitations. Self-reported, national, one figure per specialty. Not adjusted for region, setting or hours except as each publisher states. Reproduced one figure per specialty with attribution; nothing beyond that is reproduced or derived.
Definition. Annual wages of employed physicians from the U.S. Bureau of Labor Statistics Occupational Employment and Wage Statistics survey (public domain, 17 U.S.C. § 105). Wages include production bonuses and exclude nonproduction bonuses, so a wage percentile reads lower than a total-compensation survey for the same specialty. Current vintage: May 2025, released 2026. Every row was checked against the BLS public API on 2026-09-11.
Mapping. 15 specialties have their own SOC code and get an Offer Check. 22 more fall under a catch-all occupation (“Physicians, All Other” 29-1229 or “Surgeons, All Other” 29-1249); their pages show that row as context, labelled catch-all, and the Offer Check does not run against it. The rest have no defensible federal row and show none. The mapping is published in code (lib/instruments/blsSlugs.ts) and on every page’s “Data on this page” block.
Top-coding. Through the May 2024 release BLS published no annual wage above $239,200; from May 2025 it publishes specific values above that line. Each percentile carries a top-coded flag per vintage; a capped value renders as “above the federal reporting cap”, never as a number. No percentile in the current release is capped.
Limitations. Wages, not total compensation; national; employed physicians only; broad occupation codes for most subspecialties.
Definition. Total compensation entered by physicians on this platform, by specialty × geography × practice setting. Every submitter is asked to verify their NPI against the CMS National Provider Registry (active NPI-1, individual, last name matches). Verification confirms only that a real clinician submitted; it never reveals who.
The floor. No figure of any kind — not a median, a rounded number or a range — is shown for a cell with fewer than 30 verified contributors, and percentiles are not computed below it. Verified and unverified submissions are counted and displayed separately, never averaged into one figure. Every benchmark surface says which it is: “30+ verified physicians” or “not enough verified data to display — n more needed”. The scarcity is the quality signal.
Never blended. Contributed total compensation is not weighted with, averaged into or adjusted by the surveys or the federal row. A weighting of a wage figure and a total-compensation figure has no defensible interpretation, and we do not publish one. Once a cell clears the floor its benchmark stands on its own, beside the others.
| Surface | Dataset | What it shows |
|---|---|---|
| /compensation index headline | Survey reference | Midpoint of the two 2026 reports, or the one source, named |
| /compensation/[specialty] hero | Survey reference | Same headline with each source’s own figure and year beneath it |
| Offer Check (on the specialty page) | Federal wage benchmark | Where a base salary sits between the BLS percentiles for the specialty’s SOC code |
| Catch-all context block | Federal wage benchmark | The catch-all occupation’s percentiles, labelled context, not a benchmark |
| Homepage compensation calculator | Survey reference | The specialty headline; verified contributions once a cell clears the floor |
| Contributed benchmark | Contributed | Nothing below 30 verified physicians per cell; then the cell’s own figures |
| Career visibility / offer board | Contributed | Cohort totals at 5 or more, never an individual, never a figure below the floor |
Related: editorial standards · no conflicts of interest