Federal wage data is a survey of employers, so it excludes the self-employed entirely, reports straight-time gross pay without employer benefit contributions, and describes a reference period roughly a year before publication. For an employed worker comparing an offer it is the best source available; for a contractor, a business owner or anyone pricing a whole package it is incomplete in ways the figure does not announce.
Five things the wage survey does not contain
Years of experience. Not collected, not published, not derivable. The tenth percentile mixes beginners with part-time-equivalent workers and low-paying regions, and nothing separates them.
Sex, race and ethnicity. The wage survey is an establishment survey and carries no worker demographics at all. Pay gap analysis comes from a different, household-based survey with a different methodology.
Remote work. No flag. Jobs are counted where the establishment sits, not where the worker does, which limits every geographic claim built on it.
Bonus, equity and most incentive pay. The survey covers wages and salaries including some production bonuses and commission, but not annual bonuses, not stock, not employer retirement or insurance contributions.
Employer identity or size. You cannot ask what large employers pay against small ones.
Why the omissions are deliberate
It is an establishment survey: employers report what they pay for positions. Employers can answer that accurately and cheaply. Asking them for worker demographics, remote status or individual tenure would produce worse data at higher cost, and response rates would fall.
So the omissions are the price of the survey’s biggest strength. It has a very large sample and consistent definitions precisely because it asks a narrow question well.
What that means when you use it
It answers “what does this occupation pay in this place” with real authority. It cannot answer “what should somebody with eight years of experience earn”, and any site claiming to answer that from this data is inferring rather than reporting.
Experience-based figures come from private compensation surveys, which have smaller samples, proprietary methods and paying customers. That is not disqualifying, but it is a different kind of source and worth recognizing as one.
The exclusions people trip over
Self-employed workers are not covered, which matters enormously for occupations where a large share work for themselves. Owners and partners of unincorporated firms are excluded. So are agricultural workers on small farms and some household workers.
For an occupation with heavy self-employment, the published figure describes only the employed portion, and that portion may be systematically different from the rest.
Suppression, and why some cells are blank
Where publishing a figure would identify an individual employer, or where the estimate does not meet quality standards, it is withheld. That is why some occupations show employment but no wage, and why some metros are missing for occupations you know exist there.
A blank is not zero and it is not an error. It usually means the occupation is small in that area, which is itself information about whether it is a market for you.
The annualization assumption
Annual figures for hourly occupations assume full-time year-round work — 2,080 hours. For occupations with substantial part-time or seasonal employment, the published annual figure describes a full-time equivalent rather than what people actually take home.
That inflates apparent annual pay in retail, food service and parts of healthcare relative to lived experience, and it is a common source of confusion when somebody compares the table against their own year.
How to fill the gaps honestly
For experience: private salary surveys or job postings in your metro, treated as weaker evidence. For demographics: the household survey, cited as such. For total compensation: the employer’s own plan documents, because nothing public covers your specific package.
The discipline that matters is naming which source answered which question rather than blending them into one confident number. A figure that came from three places and admits to none of them is the least trustworthy kind.
Common questions
Does the wage survey include bonuses?
It includes production bonuses, commissions and cost-of-living allowances as part of straight-time pay. It excludes annual discretionary bonuses, overtime premiums and all employer benefit contributions.
Why is my occupation not listed?
Either it falls inside a broader occupation code, or it is too small or too geographically concentrated for the survey to publish without disclosing individual employers. Both are common and neither means the work does not exist.
Does it cover part-time workers?
Yes, and their wages are reported on an hourly basis so part-time work does not drag the annual figures down. Annual figures for occupations that are predominantly part-time should still be read carefully.
Is there a better source for the self-employed?
Not really. Tax filing data and industry association surveys are the nearest options and both have worse coverage and worse comparability. It is a genuine gap rather than an oversight.
How far behind is the data?
Generally about a year between the reference period and publication. The reference period is what the figure describes, and it is the one to compare against.
Does BLS wage data include years of experience?
No. It is not collected or published, so the tenth percentile mixes beginners with part-time-equivalent workers and low-paying regions.
Why are some wage figures missing?
Suppression — where publishing would identify an employer or the estimate fails quality standards. A blank usually means the occupation is small there.
Does it cover self-employed workers?
No. Self-employed people and owners of unincorporated firms are excluded, which matters for occupations where many work for themselves.