When it comes to judging people’s characters, you’d think that your family and friends would know you best, but a new study begs to differ. Researchers from the University of Cambridge and Stanford University claim that computers can actually be a better judge of character than the people who are supposed to know us best.

In a study involving thousands of volunteers, researchers found that computer models are capable of predicting a person’s key character traits. In most cases, the computer was more successful than siblings, parents and partners in identifying five important traits.

The model is created by analysing ‘likes’ on Facebook. Researchers from the two universities had already suggested that you can predict a huge amount, including information such as a person’s political beliefs and sexual orientation, by analysing their social media profile and this time, they wanted to delve deeper into the online persona to analyse the relationship between ‘likes’ and 5 key traits, including agreeableness, conscientiousness, extraversion, neuroticism and openness.

As part of the study, volunteers were asked to fill in a survey and they then asked people close to them, including family members and friends, to complete a questionnaire as a character witness.

Researchers then compared the results of the surveys with the computer model and found that the computer could outperform the individual’s colleagues by analysing just 10 ‘likes’ from their Facebook profile, their friends by analysing 70 ‘likes’, their family members by analysing 150 ‘likes’ and their partner by using 300 ‘likes’.

Dr Youyou Wu, lead researcher, said that the average Facebook user has 227 likes and this gives the computer plenty of information to go on.

Digital footprints are already used by advertising agencies and could be useful for employers looking to learn more about potential employees.