Big Five percentiles and norm groups explained

A percentile is not a grade or a percentage

A percentile tells you where you stand, not how much of something you have or how well you did. Three misunderstandings come up a lot:

• 70 is not an exam mark out of 100. There is no pass or fail. A 30 is just as much a result as a 70.
• 70 does not mean you are 70 percent extraverted. It only says that about 70 out of 100 people in the group come out lower.
• 70 does not mean you got 70 percent of the questions right. A personality test has no right or wrong answers.

Every score is a good score. A 20 on Conscientiousness is not a shortfall: it means you are more flexible than most people in the group and less tied to a plan. An 85 means you would rather work to a plan than most people would. Both sides have their own strengths.

Why is the middle at 50?

At 50, half the group scores lower and half scores higher. That is the middle by definition, for every trait and in every group.

Around that middle, a lot of people are packed closely together. Most people answer in a way that is not far from the average on any given trait; strongly pronounced answers are less common. That leads to something many people find surprising. Near the middle, a small shift in your answers can move you quite a few percentile points, because you pass a lot of people at once. At the edges it works the other way round. To get from 90 to 99, your pattern of answers has to differ much more.

So going from 50 to 60 is a small step in how you are, while going from 90 to 99 is a big one. Do not read the numbers as a ruler with equal spacing.

A percentile of 0 or 100 would mean that nobody, or everybody, scores lower than you, but in a large group there is always someone a little higher or lower. That is why DepthFive always shows a whole number from 1 to 99.

Why are 48 and 53 almost the same?

Because every measurement contains a little noise. Your answers also depend on your mood that day, on how you read a question and on chance choices between "agree" and "strongly agree". Take the same test again in a month and you will rarely land on exactly the same number. That noise is called measurement error.

Measurement error weighs most heavily around the middle, where so many people sit close together. A 48 and a 53 say practically the same thing: you are close to the middle and recognise something of both sides. That is why DepthFive describes scores from 40 to 60 as "just above the middle" or "just below the middle". On a new measurement you could easily end up on the other side of 50. Further from the middle, at 15 or 85, the direction is much clearer.

The number of questions matters too. Research on very short questionnaires, with one or two questions per trait, shows that they considerably underestimate how personality relates to behaviour, and that even a slightly longer list helps clearly ^[1]. The more questions per trait, the better chance answers average out. A test of 20 questions gives a first picture; with 120 questions that picture becomes more precise.

What is a norm group?

A norm group is the group of people you are compared with. They have answered the same questions, and it is known how their scores are spread. Without a norm group your score is a loose number. With one, it becomes a position: higher than so many out of 100.

Two things make a norm group useful.

Size. In a small group, chance shapes what the spread looks like. A handful of people with unusual answers can shift the whole picture. In a large group those flukes average out and every percentile is more settled. DepthFive compares you with a full reference group of more than 100,000 people.

Make-up. Who is in the group decides what counts as average. People change a little on average over the years: between their twenties and their forties, for instance, they tend to become somewhat more conscientious and more emotionally stable ^[2]. A group of only people over 50 therefore has a different middle from a group of all ages. Men and women also differ a little on average, although those differences are small compared with the differences between people within each group ^[3].

Why do the same answers give a different number?

Because a percentile depends not only on you but also on the group. Your answers stay the same; the yardstick moves.

Here is a made-up example. Say you come out at 50 on Conscientiousness compared with the full reference group. Older adults score higher on this trait on average, because people tend to become more conscientious as the years go by ^[2]. Compare the very same answers with people aged 50 and over only, and you stand lower, at 35 for example. Both numbers are correct. They simply answer different questions: where do you stand among everyone, and where do you stand among people your age?

It also explains why two different tests sometimes give you different numbers, even when you answer in roughly the same way. Different questions, a different norm group and a different way of calculating lead to a different percentile. When you compare results, look mainly at whether you are high, low or close to the middle, and less at whether it is 64 or 71. DepthFive deliberately shows two numbers for every trait; the reason is explained in two comparisons.

Percentiles at DepthFive

After the free test of 20 questions, DepthFive shows your percentile for each trait compared with more than 100,000 people, with a curve in which the coloured part stands for the people who score lower than you. If you fill in your age and gender, you also see where you stand among people like you. You can see it for yourself with the free test. What the five traits are is explained in what is the Big Five.