The University of Cambridge and Microsoft Research have found that attributes can be predicted by analyzing digital behavior.
The University of Cambridge and Microsoft Research partnered on a study to analyze what our Facebook Likes say about us. Over 58,000 U.S. users’ Facebook Likes were analyzed by feeding them into algorithms and corroborating them with information from profiles and personality tests.
It turns out the Like carries more weight than we might think, with high accuracy rates in determining gender, race, and even political affiliation. The researchers created statistical models that could predict these personal details using Facebook Likes alone. These proved 88% accurate for determining male sexuality, 95% accurate distinguishing African-American from Caucasian American, and 85% accurate differentiating Republican from Democrat.
Accurate predictions relied on ‘inference’, which meant aggregating huge amounts of less informative Likes such as music and TV shows to produce incisive personal profiles. The researchers also tested for personality traits and attributes, with some Likes having strong but random links, such as liking ‘Curly Fries’ indicating a high IQ, or liking ‘That Spider is More Scared Than U Are’ indicating a non-smoker. One of the researchers, Michal Kosinski, said:
We believe that our results, while based on Facebook Likes, apply to a wider range of online behaviours. Similar predictions could be made from all manner of digital data, with this kind of secondary ‘inference’ made with remarkable accuracy – statistically predicting sensitive information people might not want revealed. Given the variety of digital traces people leave behind, it’s becoming increasingly difficult for individuals to control.