Averages, graduate salaries, and the number that nobody actually earns
Imagine a bar with ten people in it. Nine of them are having a perfectly ordinary Tuesday – nurses, teachers, office workers, a couple of students. Between them they earn an average of around £30,000 a year.
Then Bill Gates walks in.
The average salary in that bar is now somewhere north of £10 billion. Statistically, everyone in that bar is a billionaire. In reality, nine people are nursing a pint and wondering why their round just got so complicated.
This is the problem with averages – or more specifically, with the mean, which is what most people mean when they say “average.” Add everything up, divide by the number of things, and you get a number that can be technically correct and yet tell you almost nothing useful about the reality it’s supposed to represent.
Now replace that bar with a university graduate cohort, and you start to see why this matters.
What happens with graduate salary data
When universities report graduate salaries – in prospectuses, on course pages, in league tables – they typically report a mean figure. The mean is pulled upward by the highest earners in a cohort. In most degree subjects, that means a small number of graduates in high-paying roles are doing significant work on behalf of the average.
Take a humanities degree cohort of one hundred graduates. Ninety of them are in roles paying between £22,000 and £32,000. Five are in roles paying £50,000 to £70,000. Five have gone into finance, law, or technology and are earning £90,000 or more. The mean salary for that cohort might be reported as £38,000. The vast majority of graduates from that course will never see £38,000 in their first few years of working life.
The median – the middle value, the salary that half the cohort earns more than and half earns less than – would tell a very different story. It would sit somewhere around £27,000. That is the number that better represents what a typical graduate from that course might expect to earn.
Both numbers come from the same data. Only one of them is useful for a prospective student making a decision about their future.
Why the mean keeps getting used
The mean is not wrong. It is a legitimate measure with genuine uses – particularly when a distribution is reasonably symmetrical and there are no extreme values pulling it in one direction. In those cases, the mean and median will be close to each other, and either will give you a reasonable picture.
The problem arises when the distribution is skewed – when there are a small number of very high or very low values that pull the mean away from where most of the data actually sits. Income and salary data is almost always skewed in this way. A small number of very high earners pull the mean upward, away from the experience of the majority.
This is why economists and statisticians tend to prefer median household income over mean household income when describing living standards. The mean is sensitive to the billionaire in the corner of the room. The median is not.
Graduate salary data has the same property. The mean is sensitive to the graduate who went into investment banking. The median reflects the graduate who went into the role most graduates from that course actually go into.
The mean gets used partly because it is familiar, partly because it is easy to calculate, and partly – it would be uncharitable but not entirely wrong to note – because it tends to produce a higher number.
What to ask when you see an average
When you encounter a salary figure, a performance metric, a satisfaction score, or any other number described as an “average,” a few questions are worth asking:
Is this a mean or a median? If the source doesn’t specify, it is almost certainly a mean. Ask which one it is, and whether both are available.
What does the distribution look like? A mean of £38,000 tells you very little without knowing whether most people cluster around that figure or whether it is being pulled up by a small number of outliers. A range, or better still a chart, tells you far more than a single number.
Who is included and who isn’t? Graduate salary data typically captures outcomes at a fixed point after graduation – fifteen months is the current standard in the UK. It may exclude graduates who are self-employed, studying further, working part-time, or not yet in the labour market. The average is only an average of the people who were counted.
What is it being compared to? A mean graduate salary is most useful when compared to a median – or to the salary data for the same subject at other institutions, or to regional salary norms. A number without a point of comparison is a number without context.
The point of this series
Numbers are not neutral. They are always a choice – a choice about what to measure, how to frame it, and what to leave out. That does not make data untrustworthy. It makes data literacy essential.
Each post in this series takes one common data concept that tends to get misread, misreported, or misunderstood – and tries to explain it in plain terms. Not to be cynical about data, but to help you use it with more confidence.
Because yes, the number might be technically correct. But what does it actually mean?