Everybody Lies - Seth Stephens-Davidowitz [Review]

The results of a General Social Survey conducted in the United States, looking at heterosexual sex, suggested that men use approximately 1.1 billion condoms every year, compared to women who apparently use 1.6 billion. This is based on how much sex they claim to have, and how often they claim to use a condom. So, who is telling the truth? It turns out neither (according to the author) – fewer than 600 million condoms are sold each year.
 
This book looks to dispel the notion that surveys are particularly accurate and that they can generally be trusted, because ultimately people lie. People lie for all manner of reasons, particularly when it comes to rather personal questions or in ways that they believe will inflate people’s perceptions of them. Consider this for the former; are people in conservative states in the US less likely to be gay, or are people simply less likely to feel safe enough to admit this? An example of the latter may be someone acting rather boastful about how much sex they are having, and how safely they are practicing this…
 
Surveyors are aware of this issue and I’m not sure there is a great deal that can be done about it. There is, conversely, a few things surveyors can do to manipulate survey results in their favour, from keeping surveys online and anonymous (which is open to malicious attacks), to having surveys be conducted by attractive or powerful individuals to encourage boastful (and untruthful) responses in the surveyors’ favour. 
 
So, everybody lies, but there is something the author believes we speak to with absolute candour; the Internet. According to the author online data, primarily (but not exclusively) Google search data, yields much more specific, accurate, and telling data results, and therefore should be emphasised in many cases over traditional survey data.

This really reminded me of John Rawl’s theory of the veil of ignorance. The only way to have a fair society is if all its members voted under a veil of ignorance, where they do not know their age, sex, class, race, religion, or any other identifying features. The only way to have reliable data, then, is to take it from the only place people are not influenced by how they may be perceived. 
 
Some of the most interesting passages were regarding the 2008 US presidential election and the results of the EU membership referendum in 2016. He recounts the reaction of mainstream media at the time to the former event, how Obama’s inauguration was celebrated as a huge, positive step in race relations in the western world. The author contradicts this with google search data which shows a surge in extremely specific racist searches at the time. People may tell surveyors they are pleased by this landmark event, when they are in public, where they have to show their face amongst their peers, but when they go home they are searching for extremely racist jokes and engaging in abusive commentary in online forums.
 
The author does have a tendency to get distracted by some of the stories he’s telling that stray away slightly from the point of the chapter, but I can forgive that; I am all for data, knowledge, information, and nuggets of random information, and it didn’t require too much mental energy to manually link back to the purpose of the chapter.
 
The conclusion was similarly weak, which the author admits (though this doesn’t make it any less weak). He leads with the argument that data shows people do not generally reach conclusions when reading, so it doesn’t really matter if it is any good. It is a tongue in cheek comment and the conclusion circles back to make a wider point which I didn’t hate, and as I’m writing this now I’m realising that I actually thought it was quite clever so please ignore everything you just read 😊