Researchers determined that computer models need only 10, 70, 150, or 300 Likes, respectively, to generate more accurate judgements than an average work colleague, friend, family member, or spouse.

Researchers determined that computer models need only ten, 70, 150, or 300 Likes, respectively, to generate more accurate judgements than an average work colleague,
friend, family member, or spouse. (Image from Sean MacEntee, flickr.com)

If you have 100 or more Likes on Facebook then your computer believes it knows you better than the average person, and if you have 300 or more, then it believes it could know you better than even your spouse.

Researchers analysed both individuals’ responses to a personality questionnaire and obtained information on Facebook likes for over 70,000 participants. The participants filled in the International Personality Item Pool Five-Factor Model of personality questionnaire, which measures openness, conscientiousness, extraversion, agreeableness, and neuroticism.

The researchers conclude that computer-based personality judgements are more accurate to assess these specific traits than human beings. These personality assessment tools could improve people’s lives by helping choose better suited activities, careers and romantic partners.

The authors now hope that consumers, technology developers, and policy makers will support privacy-protecting laws and technologies to prevent distrust and reject of these digital technologies.

Original research paper published in PNAS on January 12, 2014.

Names and affiliations of selected authors

Wu Youyou, University of Cambridge, United Kingdom