- A 2015 study found computers can make more accurate personality judgments about a person than most friends and family with enough Facebook “likes.”
- The study’s authors said that kind of information about people’s personalities can improve lives, but they warned it can also “be used to manipulate and influence” people.
According to a 2015 study that examined thousands of Facebook users, what you do on the social network — specifically what you choose to “like” — might paint a better picture of you than even your friends can. It’s useful to advertisers, but frightening if you think about how the same data could be used to influence a person’s politics.
The authors of the study, in fact, understood this years ago — noting that at “knowledge of people’s personalities can also be used to manipulate and influence them.”
The company says it looks at the information you’ve provided on Facebook and “other activity” to build a profile that might be interesting to advertisers. It gives the following example of how this works: the ads of a given advertiser trying to reach nearby 18-35 year olds interested in bicycling would populate for a hypothetical 30-year old who is interested in, among other things, bicycling in the area.
The 2015 paper, published in the journal PNAS, suggests that Facebook “likes” can reveal more about people than just interest and music, movie, book, and sports preferences. Wu Youyou, Michal Kosinski, and David Stillwell found that computers using this information can make more accurate personality judgments about a person than most friends and family.
The researchers gathered responses from tens of thousands of volunteers who filled out a 100-item International Personality Item Pool (IPIP) Five-Factor Model personality questionnaire that measured the “Big Five” qualities psychologists have argued give a good sense of who a person is: openness, conscientiousness, extroversion, agreeableness, and neuroticism.
The researchers then got the “computer-based personality judgments” from those people’s Facebook “likes,” which had previously been shown to successfully predict personality and psychological traits. And finally, they got “human personality judgments” from the participants’ Facebook friends, who were asked to give descriptions of the participants using a shorter version of the personality test.
Comparing all those of assessments, the researchers found that the computer’s average accuracy across the five traits “steadily grows” with number of likes on a person’s profile. Computer models only needed about 100 likes to outperform an average human judge in their sample.
“Compared with the accuracy of various human judges reported in the meta-analysis, computer models need 10, 70, 150, and 300 Likes, respectively, to outperform an average work colleague, cohabitant or friend, family member, and spouse,” they wrote, which you can see in the below chart. (Note that it’s log-scaled.)
Altogether, researchers got a sample of 17,622 participants judged by one friend or family member, and a sample of 14,140 judged by two.
The researchers also said exploring which “likes” were the most predictive of given traits showed that they “represent activities, attitudes, and preferences highly aligned with the Big Five theory.”
A fun example they gave was that study participants with higher levels of “extroversion” tended to “like” partying, dancing, or Snooki, the star of reality show “Jersey Shore.”
One key takeaway from their results was, as the authors wrote in the discussion section of their paper, how information from quick personality assessment tools can be used: (emphasis ours):
“Automated, accurate, and cheap personality assessment tools could affect society in many ways: marketing messages could be tailored to users’ personalities; recruiters could better match candidates with jobs based on their personality; products and services could adjust their behavior to best match their users’ characters and changing moods; and scientists could collect personality data without burdening participants with lengthy questionnaires. […] It is possible that such data-driven decisions will improve people’s lives.
However, knowledge of people’s personalities can also be used to manipulate and influence them. Understandably, people might distrust or reject digital technologies after realizing that their government, internet provider, web browser, online social network, or search engine can infer their personal characteristics more accurately than their closest family members. We hope that consumers, technology developers, and policymakers will tackle those challenges by supporting privacy-protecting laws and technologies, and giving the users full control over their digital footprints.”
In other words, digital footprints like Facebook “likes” don’t just reflect your interests and preferences, but might also reveal psychological characteristics that marketers can (theoretically) build personality profiles around.
We should note that this paper was published back in 2015, and things have moved forward in Facebook in particular and in technology in general since then.
Furthermore, researchers only looked at “likes.” There are plenty of other digital footprints we leave such as websites we visit, the frequency with which we visit them, things we google, what apps we download, people and things we search on Facebook, how many internet friends we have, who we interact with often, where we go in real life (phones have a GPS system and motion sensors), etc., etc., etc.
Getting all of that data can theoretically help computers build an even more accurate psychological portrait of a person — and then fit him or her into a psychological consumer category.
And finally, although many have focused on Facebook, this isn’t just a Facebook thing. Earlier this month, Hearst Magazines, which publishes Cosmopolitan and Elle, dijo it has more powerful advertising data than Facebook in some cases as it has amassed unique profiles on 120 million digital consumers over the past year via various things like online quizzes and Snapchat polls.
los company told Business Insider’s Mike Shields in mid-September that it is touting this pool of data, which it says is rich with information on people’s shopping habits, in planned meetings with numerous agencies and marketers over the next few weeks.