You may keep your personal details to yourself, as well as your gender, location and political afflictions. But it doesn’t matter – the internet knows who you are. It may not be able to figure out your name, but when it comes to the other stuff, to what makes you tick, the internet indeed knows.
This was the conclusion when the Mitre Corporation conducted a little study. The researchers fed a host of anonymous tweets into a computer, which had been programmed with the knowledge of which words were most likely to be used by men, and which were likely to be spoken by women. The result? The computer could guess the person’s gender 75.8% of the time. This was when the computer could read a person’s entire Twitter stream – when constricted to reading only one tweet, the accuracy dropped to 65.9%.
Obviously, this is not a hard and fast rule. Women don’t tweet about puppies and chocolate all day (and those who are letting the side down), but according to the art of socio-linguistics, there are definitely patters to how men and women speak. Did you know women tend to laugh more? And use more exclamation points! Computers can be programmed with this knowledge to help researchers find patterns.
Another feature of the study was when the researchers added the phrase “my” in front of certain words. This enlightening little chart shows how likely someone would be to be a Democrat or Republican based on this:
In spite of how it feels at times, Twitter is not a microcosm of life, so the findings are not entirely reflective of the real world. But advertisers may take an interest in this, as it will give them a decent shot at figuring out who we are and what we want in the murky waters of the web. So next time you trot around the internet in your anonymous disguise discussing the merits of Etsy versus Folksy, be aware: they are on to you.