Why salary sites disagree

Four kinds of salary data

Employer surveys. The Bureau of Labor Statistics asks employers what they pay the people they actually employ. Its OEWS survey covers about 1.1 million workplaces over three years and every occupation, state and metro area. It is the source on this site. Its weaknesses: it is published about a year after the reference month, it leaves out the self-employed and it does not split pay by experience.

Job ads. Job boards estimate pay from the ranges in postings. Ads describe jobs that are open now, not the pay of everyone in the field. Open jobs skew toward high-turnover employers and entry-level roles in some fields and toward hard-to-fill senior roles in others, and posted ranges are often wide. Hourly ads are common, so annual figures are often hourly rates multiplied out.

Self-reported pay. Some sites collect salaries from their users. The answers come from people who chose to share them, often recently and often in large companies and big metros. Some sites include bonuses, stock or tips that employer surveys handle differently.

Models. Compensation-data companies blend surveys of employers with their own adjustments to produce a figure for a job title, often for a "typical" level of experience or a specific city. The method is usually proprietary, so the figure cannot be checked.

What each one measures

Beyond the source, check four things before comparing two figures:

• Median or average. A median is the middle wage; an average can be pulled up by a few very high earners. For all US jobs the median was $50,980 and the mean $69,770 in May 2025.
• Base pay or total pay. OEWS counts base pay, tips, commissions and production bonuses but not overtime premiums, shift differentials or year-end bonuses. Sites that report "total pay" can be much higher for jobs with large bonuses or overtime.
• Who is included. All workers, only full-time workers, only new hires or only people who answered a survey. Part-time workers lower annual figures; leaving out the self-employed changes pay for real estate agents, barbers or trainers.
• When and where. A figure from a single expensive city, or from an older year, will not match a national figure for the latest year.

An example: dental hygienists

BLS puts the median for dental hygienists at $98,100 a year (May 2025), with the middle half between $83,570 and $108,480 and state medians from $61,160 to $137,220. A job board that averages current ads, a site that reports what its users entered, and a model for one city will each land somewhere in or around that range, and none of them has to be wrong for the numbers to differ. Many hygienists work part time or for several practices, which pulls down annual figures built from hours worked and pushes up figures built from hourly ad rates.

How to read any salary figure

• Look for the source and the date. A figure with neither cannot be checked.
• Prefer ranges to single numbers. The spread between the 25th and 75th percentiles tells you more than any one figure.
• Compare like with like: national with national, base pay with base pay, the same year.
• Adjust for where you live. The state and metro tables on our career pages, and the cost-of-living adjustment beside them, show how much location matters.
• For an offer, use local data: our state and metro figures, the employer's posted range and, where they exist, union or government pay scales.

Our figures are BLS employer-survey medians and percentiles throughout, so they are consistent from one page to the next. The methodology explains how we use them.