What a person "likes" on Facebook can be used to accurately predict many personal traits, including sexual orientation, ethnicity, religious and political views, personality traits, intelligence, happiness, drug abuse, age and gender, according to a new study from the University of Cambridge.
The study, titled "Private traits and attributes are predictable from digital records of human behaviour", is based on the analysis of the Facebook Likes (correlated with detailed demographic profiles and psychometric tests) of over 58 000 volunteers in the US.
The researchers say: "This study demonstrates the degree to which relatively basic digital records of human behaviour can be used to automatically and accurately estimate a wide range of personal attributes that people would typically assume to be private."
Facebook Likes are defined as "a mechanism used by Facebook users to express their positive association with (or 'Like') online content, such as photos, friends' status updates, Facebook pages of products, sports, musicians, books, restaurants, or popular Web sites".
"Likes represent a very generic class of digital records, similar to Web search queries, Web browsing histories, and credit card purchases. For example, observing users' Likes related to music provides similar information to observing records of songs listened to online, songs and artists searched for using a Web search engine, or subscriptions to related Twitter channels," says the study.
"In contrast to these other sources of information, Facebook Likes are unusual in that they are currently publicly available by default. However, those other digital records are still available to numerous parties (eg, governments, developers of Web browsers, search engines, or Facebook applications), and, hence, similar predictions are unlikely to be limited to the Facebook environment."
High accuracy
Through the analysis of "Likes", the researchers were able to accurately classify people as Caucasian or African-American with 95% accuracy. The gender of the volunteers was correctly classified in 93% of cases, while Christians and Muslims were correctly classified with 82% accuracy. Since the study was based in the US, it was also able to correctly classify Democrats and Republicans in 85% of cases.
"Sexual orientation was easier to distinguish among males (88%) than females (75%), which may suggest a wider behavioural divide (as observed from online behaviour) between hetero- and homosexual males," says the study, adding that good prediction accuracy was achieved for relationship status and substance use (between 65% and 73%).
The model's accuracy was lowest (60%) when inferring whether users' parents stayed together or separated before users were 21 years old. It was found that people whose parents had split, were more likely to like statements about relationships, such as "If I'm with you then I'm with you. I don't want anybody else" and "I'm sorry I love you".
The study found the best predictors of intelligence to be liking things like "Thunderstorms", "The Colbert Report", "Science" and "Morgan Freeman's Voice". On the other side, people who liked "Bret Michaels", "Harley Davidson", "Lady Antebellum" and the "I Love Being a Mom" page, generally had lower IQs.
Dangerous implications
According to the study, the findings can be used to improve a number of products and services - especially in terms of offering personalised content and advertising.
"For instance, digital systems and devices (such as online stores or cars) could be designed to adjust their behaviour to best fit each user's inferred profile. Also, the relevance of marketing and product recommendations could be improved by adding psychological dimensions to current user models. For example, online insurance advertisements might emphasise security when facing emotionally unstable (neurotic) users, but stress potential threats when dealing with emotionally stable ones."
It is also noted, however, that the predictability of individual attributes from digital records of behaviour could have serious negative implications.
"It can easily be applied to large numbers of people without obtaining their individual consent and without them noticing. Commercial companies, governmental institutions, or even one's Facebook friends could use software to infer attributes such as intelligence, sexual orientation, or political views that an individual may not have intended to share," says the study.
"It is our hope, however, that the trust and goodwill among parties interacting in the digital environment can be maintained by providing users with transparency and control over their information, leading to an individually controlled balance between the promises and perils of the Digital Age."
The researchers involved in the study have made the mechanism they developed to analyse Facebook Likes available to the general public at www.youarewhatyoulike.com. The site does an instant personality test and rates how open, stable, conscientious, extroverted and agreeable a person is - based solely on their Facebook Likes.
Share