top of page

Study Finds Facebook ‘Likes’ Can be Used to Determine Demographic and Personal Information

While many social media users have become more aware of privacy issues and have made their profiles private, a recent academic study has found that what people ‘Like’ on Facebook can tell a lot about them. Unlike profile information, Facebook ‘Likes’ are less likely to be made private and are then publicly visible to everyone.

It has been documented in the past that consumers with open social media profiles are more likely to experience fraud as those with public Facebook profiles were 50 percent more likely to be victims of identity fraud in 2011. While many consumers have made their social media profiles private and have become stricter on which ‘Friend’ or ‘Invitation to Connect’ requests they’ll accept, far fewer have taken steps to restrict or monitor what they ‘+1’ or ‘Like.’

Tactics still used by fraudsters today include building fake profiles and trying to get thousands of strangers to accept their Friend Request or Invitation to Connect. Once this is achieved their private profile information becomes visible, and this information can be used to help guess passwords, answer security questions and obtain personal information such as a city, phone number or email address. Fraudsters then supplement the information they have for a compromised identity with this updated personal information to aid in taking over accounts, passing knowledge based assessments or furthering their fraud scheme in one way or another.

Another way fraudsters could use social media information to aid in committing fraud is by way of Facebook ‘Likes.’ A recent study analyzed data from 58,000 Facebook users and was able to predict certain traits and information about the person behind the profile, even though this information was not explicitly provided. The study examined the 55,000 most common Facebook ‘Likes,’ of which surveyed users had a median number of 68. Using just information from what the Facebook profiles had listed as ‘Likes’ an algorithm was able to accurately predict the user’s gender 93 percent of time and whether the user was Caucasian or African American 95 percent of the time. The algorithm was also able to correctly predict the user’s political affiliation and religion more than 80 percent of the time. Other factors, such as approximate age and whether the user smokes cigarettes, were accurately predicted at least 60 percent of the time.

Facebook likes are public by default, and are then available to be viewed by any user of the site as well as third-party apps, ISPs and others. While this ‘Like’ information is used by Facebook for targeted advertising, it can also be used by fraudsters to take over accounts or to make phishing scams more targeted at each user or groups of users. With all the identifiable and personal information that consumers are sharing online, knowingly and unknowingly, it is more difficult than ever for organizations to build trust in an e-Identity.

For more information:


bottom of page