“Like” clicks on Facebook show a lot about personality
Anyone who marks a post on Facebook with “I like it” always shows the public something about preferences and personality traits. Now US researchers have found that computers based on a sufficient number of “likes” can possibly even better judge character traits than friends, family and partners.
Users show preferences about sharing and liking posts Whether group membership, sharing posts or the famous “Like” button: Facebook users often reveal a lot about themselves to the public. But apparently, computers can determine much more specific character traits based on Facebook activity than friends or even their own partner. This is the result of a study by the Universities of Cambridge and Stanford, which was recently published in the journal "Proceedings of the National Academy of Sciences" (PNAS) and is based on a previous study by the University of Cambridge in 2013.
Researchers evaluate data from more than 86,000 test subjects. According to this, the study had shown that special software was able to determine the characteristics of Facebook users more precisely than related parties thanks to the so-called “likes”. The researchers had analyzed the data from more than 86,000 subjects who had previously answered questions about their personality and had access to their "Like" information on Facebook. The five characteristics of openness, conviviality, sense of duty, tolerance and impulsiveness were at the center of the evaluation, so that, for example, a favor by the surrealist painter Salvador Dali or the topic "meditation" could be regarded as a hint for a very open being, the scientists said.
Computer provides more precise information than colleagues and family on the basis of the likes The users' sympathies and interests were analyzed in the next step by a computer, while friends, relatives and partners were also able to answer questions about the character of the respective person via an app. The scientists then compared this information from around 32,000 test participants with that of computer analysis and came to an interesting result: “In the study, a computer was able to predict the personality of the test subjects more accurately than a work colleague by analyzing just ten likes; more than one friend or life partner (roommate) at 70, one family member (parents, siblings) at 150, and one spouse at 300 likes, ”said the University of Cambridge.
People tend to think non-rationallye
According to Dr. Michal Kosinski, co-author and researcher at Stanford University, said this result was possible because the machines had some decisive advantages over humans: the ability to store data, access to huge amounts of information, and the ability to use algorithmic technology analyze. "Big data and machine learning offer a level of accuracy that will be hard to achieve for the human mind because people tend to put too much emphasis on one or two examples or fall into non-rational mindsets," explains Kosinski further.
Computer analysis could help people make important decisions "This type of data analysis could help people to reinforce their own beliefs and judgments when making important life decisions," said study lead author Wu Youyou of the Center for Psychometry in Cambridge. "The ability to judge a personality is an integral part of social life - in everyday decisions and long-term plans, as well as who we marry, who we trust, whom we trust, or who we elect to be president," added Cambridge Co Author Dr. David Stillwell. (Nr)
Image: Peter Derrfuss / pixelio.de