How much, exactly? On Monday, researchers affiliated with Verto Analytics and the Qatar Computing Research Institute published a paper in the online journal arXiv that attempts to nail that down. In a nutshell, they cross-referenced the app usage and demographics of more than 3,700 people and determined which apps and attributes correlate. Based on those models, they then found that they could predict a user’s gender, age, marital status and income with between 61- and 82-percent accuracy.
That’s all very abstract, of course — what we really want to know is how well these predictive models work on us personally. So we pulled some of the data from this research and made a quiz from the findings.
For each app, click “yes” if you have it and “no” if you don’t. If you don’t have one of the exact apps listed, but do have something very similar, you should still select “yes.” (For our purposes, there’s no real difference between Tinder and Hinge, for instance.)
There are 16 possible results, based on your gender (male/female), your age (over/under 32), your marital status (married/single) and your income (over/under $52,000). It’s not perfectly accurate, particularly if you don’t have a lot of apps on your phone … but it should help illustrate exactly how much we’re inadvertently letting advertisers and app-makers know.
As you can probably tell from the quiz, popular apps with no direct relation to your gender or income can be indicators of both. If you have Pinterest on your phone, you’re almost definitely a woman; if you have Uber, you’re probably single. Single adults listen to music on SoundCloud. Older adults prefer iHeartRadio. Meanwhile, people who earn more than $52,000 a year tend to get their restaurant reviews on Yelp — while people turn to Foursquare when they’re making less than that.
Of the four traits the researchers studied, gender was the easiest to determine, and income was most difficult. But when you add in the other sorts of personal data that many apps collect — think location, contacts and phone usage — it’s easy to see how targeted ads and other personalized features can feel so creepily accurate. (Add in all of the data sites like Facebook have, and it’s no wonder that the site “knows” you better than your family and friends.)
“Studying the predictability of demographics … points out privacy implications of users allowing apps to access their list of installed apps,” the researchers write. “Many users undoubtedly do not carefully review the permissions that the apps they install require, and even less, understand the scope of the information that can be inferred from the data accessible by the apps.”