Ali presenting at Black Hat Abu Dhabi 2011

Alison from the Online Privacy Foundation presenting at Black Hat Abu Dhabi 2011

We recently published a Press Release regarding the results of our Facebook personality study, ‘The Big 5 Experiment’.  This post provides some further insight into those results.

By the way, if you Tweet, go check out our Twitter experiment at http://www.bigfive.me/

 

When we set out on the project we asked two simple questions.

1) Can you determine personality and privacy concerns purely from someone’s Facebook activity?

2) If you can, what are the implications?

 

1) Can you?

Well, sort of…Our study showed that there were some statistically significant correlations between Facebook usage and a user’s personality and privacy concerns. This tells us that there’s something going on.

Putting our results into practice

So, lets assume we want to guess one dimension of someone’s personality, for example are they more extrovert than introvert? Armed with the information in our study, how well would we do?

Ideally, we’d want to be 100% accurate, but that’s extremely rare. For example, the relationship of measurements feet to meters displays perfect accuracy, 100% (the green bar in the graph below). No observation of personality can ever be this good…at least not today.

Perfect Accuracy

 

If you were to toss a coin, for arguments sake, you’d be reasonably safe to assume a 50% hit rate (unless you’ve read this). See the green & red bar (below) evenly matched. 50% chance of getting in right and a 50% chance of getting it wrong.

Coin Toss

 

Now, our study shows that if you had access to an individual’s Facebook profile, you could use this information to predict whether someone is more extrovert than introvert, which would give you slightly better than 50/50 odds. Please note that the following diagram is for illustrative purposes only and provides a rough approximation of the accuracy of Facebook personality prediction for extraversion.

Facebook Accuracy (roughly)

Just for comparison, the accuracy of a structured interview is believed to be around the 80% mark.

 

And what about guessing all five personality traits correctly?

In the graphs above, we just looked at one personality dimension out of the five in our study. If anyone were to try and accurately predict all five dimensions, they’d have a very slim chance. This is because in order to predict all five personality dimensions correctly, you need to multiply the probabilities of correctly guessing each individual personality dimension (it’s a little bit more complicated than that in reality as the five personality dimensions in our study aren’t entirely independent, but that’s another story).

E.g.

60%

x

55%

x

60%

x

52%

x

53%

=

5%

P

(openness)

 

P

(conscientiousness)

 

P

(extroversion)

 

P

(agreeableness)

 

P

(neuroticism)

 

P

(guessing all 5
correctly)

 

Please note that these figures are purely for illustration.

For full results, please consult our paper here.

 

 

2) What are the implications?

Well….our biggest concern was the number of people making hiring decisions based on Facebook activity. For now, at least, these people probably shouldn’t, but

  • People are looking at your online presence before making hiring decisions.
  • What they see does influence a hiring decision.
  • People are looking into improved models of personality prediction, so the clock might be ticking on this one.

This is problematic because…

  • There is no evidence to support their decision. i.e. There have been no published studies which examine future job performance in relation to Facebook activity (or Twitter, or MySpace or Bebo).
  • There appears to be some credence to the argument that people can and do behave differently in different contexts.
  • The issues of discrimination are ill understood.
  • It’s almost certainly an unfair intrusion of an individuals right to privacy.

As mentioned in this recent article about privacy, do we really want to live in a world where “nothing is ignored, nothing is forgotten, nothing is forgiven” ?

 

What could Facebook personality prediction be used for (today)?

Employee screening and pre-screening isn’t the only use of personality prediction. The information could be used in advertising and online marketing. For example, an automobile manufacturer may target potential extroverts with more sporty cars than potential introverts. For advertising, the slight edge gained in accurately predicting personality likely outweighs the downside risk of an incorrect prediction.

There is another, more sinister, downside. Information about elements of someone’s personality could be exploited by fraudsters. If a fraudster can find the people in a group who are more likely to be trusting and gullible (and therefore more likely to fall for a scam), they’d obtain a small advantage over a random guess.

 

Conclusion

Our study re-affirms that there is a relationship between Facebook activity and personality types. However, it is clear that the strength of that relationship is not a strong enough basis on which to make critical decisions about individual users. The results also indicate that there may be considerable consequences in revealing personal information on Facebook. Further research is required in order to better understand the relationship between social network use and personality, the consequences, and how users might best manage the personal information they reveal through social network sites. Our study points to critical questions around the possible need for regulatory controls and/or raising awareness amongst users in order to prevent the misuse of information derived from Facebook and other online social network activity

 

Closing Comments

Our intent isn’t to scare people away from Social Media. In fact, as individuals, we rather enjoy its benefits, but we do want to examine its potential abuse and make people aware so that they can make informed choices about how much information to make public and in the case of employers, critically evaluate their hiring practices.

 

The Online Privacy Foundation are currently conducting a similar experiment with Twitter and are seeking volunteers. To find out more, please visit our page on the Twitter Big 5 Experiment or go directly to the personality application at http://www.bigfive.me/