Journalist pay rose 30.7% in a year. Clinical psychologists rose 27.5%. Barristers and judges fell 37.6%.
Every one of those figures is published by the Office for National Statistics, and every one of them is a bad guide to what is happening to pay in those professions. This article is about why, because the same trap catches almost every "salaries surge" headline you will read this year.
The short answer
The biggest year-on-year movements in ASHE are concentrated almost entirely in small occupations. Grouping the occupations we track by how many people work in them, and averaging how far each moved:
- Over 300,000 employees: average movement 4.5%, largest 7.5%
- 100,000–300,000: average 5.4%, largest 10.7%
- 50,000–100,000: average 5.0%, largest 13.6%
- Under 50,000: average 8.0%, largest 30.7%
Occupations with fewer than 50,000 employees move nearly twice as far in a year as the largest ones, and every extreme reading in the survey sits in that bottom group.
Why small occupations swing
ASHE is a sample. It draws on roughly 1% of employee jobs through PAYE records, which is a very large sample by survey standards and still a small one once it is divided into several hundred occupations and then split by region, sex and working pattern.
For teaching professionals, with 930,000 full-time employees, the sample is deep enough that one year's estimate closely tracks the next. Teaching moved 6.1%. Nursing moved 4.2%. Financial managers moved 0.8%.
For an occupation with 9,000 employees, the sample is a few hundred people. Change which employers happen to fall into it, or which grades within it respond, and the median moves — without a single person getting a pay rise.
The barrister case
The largest fall in the data is barristers and judges, down 37.6% on the all-employee measure. It is worth taking apart, because it shows several effects stacking.
The occupation has 28,000 employees in this edition, of whom 10,000 work part-time. On the full-time measure the fall is 17.6% rather than 37.6% — so more than half of the headline drop is a change in the mix of full-time and part-time workers captured, not a change in anyone's rate.
The remaining fall sits in a full-time sample of 18,000, in an occupation that combines two quite different jobs. ONS groups barristers with judges, and self-employed barristers — a large part of the profession — do not appear in a survey of employees at all. What is left is a small, mixed group where a shift in composition produces a large apparent move.
Nothing here proves barrister pay did not fall. It shows that this figure cannot tell you whether it did.
The same applies to the rises
It would be convenient if the noise only affected bad news. It does not.
- Journalists and reporters, +30.7%: 14,000 employees
- Clinical psychologists, +27.5%: 9,000 employees
- Arts officers, producers and directors, +19.6%: 18,000 employees
- Train and tram drivers, +19.3%: 24,000 employees
Train drivers are the one case with a plausible mechanism behind the number: pay in that occupation moves through negotiated settlements that land in a single year rather than drifting upward annually. Even there, a one-year figure from a 24,000-person sample is a weak basis for a claim about the profession's direction.
What to trust instead
Three habits get you most of the way.
Look at the level, not the change. A median built from a small sample is far more stable than the difference between two such medians. Both estimates carry error; subtracting them adds the errors together.
Check the employee count before believing a percentage. It is published alongside every figure. Under 50,000 is where scepticism should start.
Prefer several years to one. A single edition cannot distinguish a trend from a wobble. Where an occupation moves the same direction across editions, that is worth something; a single 30% jump usually is not.
Why this matters more than it sounds
Salary content is built on exactly the figures most likely to be noise, because a 30% rise makes a better headline than a 4% one. The result is a steady supply of articles announcing that some profession is booming or collapsing, resting on a sample of a few hundred people.
The occupations where pay genuinely moved for a large number of people this year are unglamorous by comparison: teaching up 6.1% across 930,000 employees, warehouse work up 5.5% across 359,000, nursing up 4.2% across 683,000. Those numbers describe far more working lives than any of the headline movers, and you will see them quoted far less.
Check any occupation
Every page on this site shows the year-on-year change alongside the employee count, so you can weigh one against the other: journalists, clinical psychologists, barristers, teachers and nurses are good places to see the contrast. Our methodology page explains what ASHE does and does not measure.
Source
Source: Office for National Statistics, Annual Survey of Hours and Earnings, Table 14 (occupation, 4-digit SOC), 2025 provisional edition. The sampling and quality caveats described here are set out in the ONS guide to interpreting ASHE estimates.