How to read public pay statistics

What the numbers actually say

Labor statistics — for example the U.S. Bureau of Labor Statistics Occupational Employment and Wage Statistics — line up pay for an occupation, a place, and a year. They answer what people in the sample earned then. They do not answer what you should earn, or what one company will offer tomorrow.

An offer still depends on budget, how scarce the skill is, and whether you can describe results. None of that is in the table. Ranges on this site prepare a conversation. They are not a promise of hire or pay.

Read the footnote first: hourly wages converted to annual, or annual already; bonus in or out; national sample or one state. Two rows with different footnotes should not be compared sideways.

How to use P25, the median, and P75

The median (P50) splits the sample in half. It is steadier than an average because a few very high or very low wages do not drag it as hard. In a talk, use it as an anchor, then move up or down for years and results.

P25 to P75 roughly frames the middle half. Thin evidence or a career change: stay between P25 and the median. Scarce skills and results someone can check: then consider median to P75. Asking for the top with no proof stalls as soon as they ask for an example.

Illustration, not a real offer: suppose similar work in a region shows pre-tax annual pay of $70,000 at P25, $95,000 at the median, and $120,000 at P75. Someone with three years who owned a full project can aim mid-to-upper — about $95,000 to $110,000 — not “I must have $120,000.”

• Median: anchor, then adjust for proof
• P25 to median: thin evidence or a switch
• Median to P75: results someone can check
• Say a range and name the basis

Four things to align before you compare

Occupation. Statistical jobs are often wider than a posting title. “Software developer” mixes junior and lead work, so the dollars spread out. Pick the row closest to the posting. Do not apply a manager table to an individual contributor role.

Place. Prices and industry clusters move pay. For remote work, ask whether pay follows the employer’s location, your home, or one national band. Those three rules can differ by a lot.

Year. Surveys often lag one or two years. Use the old table as a base map, then check five to ten recent postings that state a range. Hot roles look low if you only read the old table.

Unit. Convert everything to pre-tax annual pay. If they quote monthly, ask whether the year has 12, 13, or 14 pays, then multiply. Mixing monthly and annual either scares you or undersells you.

Bonus and equity are not the wage table

Many wage surveys record wages. Bonus may be in a footnote or absent. When they say “package,” ask how much is fixed and how much depends on company or individual results. A target bonus is not cash received.

Stock and options are usually outside occupational wage tables. Do not add an entire unvested grant into year-one wages and compare that to the median. Split it by vesting year so ongoing pay is visible.

Insurance, leave, and remote work belong on a separate line. They can be worth thousands of dollars a year to you, but folding them into one “versus the survey” number breaks the comparison.

Misreads that waste the conversation

Treating the median as “fair pay” or money the company owes you. It is only the middle of a sample. If your evidence sits below the middle, aiming above it needs a reason.

Using a national figure for a high-cost city, or a high-cost city figure for remote pay tied to where you live. If place is wrong, the band is not a reference.

Remembering one number. One number has no room. Write P25, median, and P75 so you can see whether their offer is low, middle, or high.

Ignoring year and currency. A five-year-old table next to this year’s offer, or treating dollars and another currency as the same unit, both produce a wrong conclusion.

How to say it in the room

Describe what you finished, then mention the range. The reverse sounds like a price tag, not a basis.

A usable line: “I looked at published annual pay for similar work in this region in [year]. The middle half sits roughly between P25 and P75. Given two projects I owned, I am aiming mid-to-upper.” Replace the band with the dollar figures you wrote down.

If they name a range first, confirm base versus total, and annual versus monthly. If the whole range sits under published P25 for the same scope, ask whether the level is lower or the budget is capped. Do not silently accept a number you have not decoded.

Using it with this site

The evaluator shows annual P25, median, and P75 for the role, country, and region you pick, with source and year. It is a compiled reference band, not a live scrape of every employer budget.

When you change country, read the local currency. English examples on this site use dollars. Japanese examples use yen. If you select the United Kingdom, read the result in pounds — do not treat it as dollars in your head.

Checking skills only nudges the marker up inside the band. That is an estimate, not a raise the company has agreed. Before a real talk, re-check the official table’s footnotes.

One page before the talk

One page is enough: role and place; P25, median, P75, year, and currency; your floor, target, and stretch in dollars; three results; two open questions (is bonus included, and which location sets pay).

This article and the figures on the site are for job-search prep only. They are not legal, tax, or financial advice. Pay-talk rules differ by country.