How accurately do people rate themselves? What the research says

People rate themselves with moderate accuracy: good enough to be useful, not good enough to be the whole story. Research suggests self-views of skill line up only loosely with measured performance, while self-views of personality are more accurate but contain real gaps. How accurate you are depends a lot on what you are rating.

That is not an insult to anyone's self-awareness. It is a normal feature of being a person looking out from inside your own head. Here is what the evidence does and does not show.

Skills are where self-ratings slip most

The clearest numbers come from research on ability. Ethan Zell and Zlatan Krizan pulled together 22 meta-analyses comparing how people rated their own abilities with objective measures such as test scores, grades and supervisor evaluations. The average correlation was .29.

To put that in context, a correlation of 0 means no relationship and 1 means a perfect match. So .29 is a real link, but a modest one. The authors describe it as moderate insight.

The same work offers some good news. Self-ratings matched performance better when:

• The rating was about a specific skill, not a broad one
• The task was objective
• The task was familiar
• The task was not very complex

In other words, "I am good at reading contracts" is likely to be closer to the truth than "I am smart". If you want a fairer self-rating, get specific.

We also tend to lean optimistic

A major review by David Dunning, Chip Heath and Jerry Suls looked at self-assessment across health, education and the workplace. They concluded that self-views often have only a tenuous to modest link with actual behaviour and performance.

They also describe a pattern of overrating. On average, people say they are above average, which cannot be true of everyone. People also tend to overestimate how likely they are to do desirable things, and to be too hopeful about how soon they will finish projects. If you have ever been sure a job would take an hour, you know this one already.

None of this means you should distrust everything you think about yourself. A bit of optimism can come with its own benefits, and a review by Timothy Wilson and Elizabeth Dunn makes the point that a perfectly matched self-view is not always an advantage. Still, they suggest that greater awareness is generally a good thing.

Personality is a different picture

Rating your personality is not the same as rating your skills, and the results look different. Your personality is partly made of things only you can observe, such as what runs through your mind before a big conversation.

Simine Vazire's self-other knowledge asymmetry model explains which traits you are likely to judge best. In her study of 165 people, participants rated themselves and were rated by friends and strangers, and then completed behavioural tests. Three patterns stood out:

• Traits that are hard to see from outside: people were the best judges of their own neuroticism-related traits. This is the area we call emotional stability.
• Traits that carry a lot of social value: friends were the best judges of intellect-related traits, probably because it is hard to be neutral about how clever you are.
• Traits that are easy to see: everyone was about equally good at judging extraversion-related traits.

So the answer to "how accurate am I?" is really "accurate about what?". You are often the expert on your inner world. You may be less reliable on the traits you most care about looking good on.

What about everyday behaviour?

Vazire and Matthias Mehl tested this in daily life. They compared predictions of everyday behaviour with objective records of what people actually did over four days. Close others were as accurate as the people themselves. Both perspectives also added something the other did not, which is the most useful part of the finding. There was no single viewpoint from which a person was known best.

If that interests you, our piece on why your friends might know you better than you do goes further.

Why does this happen?

Wilson and Dunn describe a basic limit: much of the mind is not open to introspection. You can't look inward and read off your own attitudes or habits the way you read a dashboard. Add a natural wish to think well of yourself, and some of the gaps make sense. They point to two ways of learning more: looking at yourself through other people's eyes, and watching what you actually do.

How to get a truer picture of yourself

You don't need to second-guess every thought. A few habits help:

1. Rate the specific, not the general. Ask "how do I do in a tense meeting?" before "am I a good communicator?".
2. Compare against a record. Note how long tasks really take, or what you actually did last week.
3. Ask people who see you often. A person who knows you well can offer a view you can't get from inside. Our guide on how to ask for honest feedback has some easy ways to start.
4. Treat gaps as information. If someone sees you differently, that does not settle who is right. It tells you how you come across, and that is worth knowing.

That is the thinking behind Second Sight, where the people around you can rate you anonymously. Friends, family and co-workers are only ever shown to you as averages, combined with at least two other raters. The Partner circle is different: it is a named, one-to-one comparison, and your partner knows that before they rate.

Take your own measure

The research suggests that your own view is valuable and incomplete. A good next step is to put your self-view on the record, then compare it with how others see you. The Core Personality Expedition is free to take with no sign-up, and it uses public-domain IPIP items built on the Big-Five Factor Markers.