Everybody Lies (2017) by Seth Stephens-Davidowitz is an interesting look at how big data gleaned from the internet can give us a better picture of what people are really like. Seth Stephens-Davidowitz (SSD) did a PhD in economics using search terms from the internet. After his PhD SSD worked at Google. He’s clearly seen a lot of data and thought about it and found a lot of interesting stuff.
The book emphasizes that what people say in public and put on Facebook isn’t what they are thinking. The dramatic use of racist search terms and their locations when President Obama was elected gave a great correlation for where Donald Trump would do well. People say how great their spouses are on Facebook, then they search for why they aren’t having sex with them on google. People outright lie about how many condoms they are using, with both sexes exaggerating substantially.
The book refers frequently to Freakonomics and Steven Levitt and the author states and shows his admiration for the work of Levitt and the way he looks for interesting data.
There are some other interesting observations about violent films and crime, which someone will hopefully soon extend to violent video games and some interesting notes on selective schools. There is even some things on the most unread books around.
The book makes the point that social science and literary studies have a fantastic new source of great data in big data. It’s an interesting and seemingly valid point.
It’s not a bad book and it would be hard to go through without learning something. SSD writes well and there are some fun facts on the way. It’s not great, but well worth a look.