Book Review: “Everybody Lies” by Seth Stephens-Davidowitz

I found this book because someone had put it on social media and to be honest, I read the blurb first and it goes over some very obvious talking points. I wasn’t fascinated by the argument to begin with and nor was I taken in by the writing style once I started reading. This book has its advantages including: accessibility, opening up conversation and the author has done a good job of keeping the argument logical. However, the downsides for me include: its readability lacks seriousness, it focuses too much on one or two things, it doesn’t really go into depth about anything and it’s obvious confirmation bias. Let’s see why this book is definitely great for ‘pop statistics’ but basically means nothing to those of us who have read Alex Edmans’ book “May Contain Lies”.

People often present an idealised version of themselves in surveys or social media, avoiding topics that might make them seem socially unacceptable. Stephens-Davidowitz demonstrates how search engine data, collected anonymously, reveals people’s real attitudes, beliefs, and concerns—things they wouldn’t typically admit in public.

This form of “data truth” uncovers genuine views on issues like health, relationships, and biases, offering a truer glimpse into human behaviour. The only problem I have with this argument is that it is basically the most common thing that everyone on social media knows. The theory that the writer has of this being an unknown concept is actually insulting to the reader. Just because people don’t really talk about it much doesn’t mean that it is unknown. Also the thinking of your own data as the ‘truth’ is a strange form of confirmation bias that I don’t think the writer intends to admit to but is doing so in the name he gives to it.

In their searches, people often reveal insecurities, questions, and desires they might not share even with close friends or partners. Stephens-Davidowitz uncovers patterns that suggest widespread dissatisfaction, curiosity about infidelity, and anxieties about sexual performance. These search patterns reveal a deeply private side to human relationships, uncovering what people genuinely wonder about but rarely speak of openly. The fact that this basically says ‘people ask Google questions they would like the answer to but feel stupid to tell others about‘ is not a discussion point. Is this not what Google was simply made for? It doesn’t matter what the question is, people using Google as a search engine is not research, it’s stating the obvious.

Search data often uncovers the real political sentiments people hold, which are sometimes hidden from pollsters. For example, Stephens-Davidowitz reveals how populist ideologies gain traction online, particularly in private search histories, which can serve as a gauge for measuring political polarisation. Analysing these patterns helps to understand the hidden drivers of political movements and voter behaviour that polls might miss. This is an important point because a lot of this goes on in the world. We will have one person who claims to be left and is actually right and vice versa. However, again this is not a research point. It is just obvious.

When it comes to a lot of the research, we can tell by reading that it has been cherry picked to represent the black and white instead of dealing with the reality of the issue which would no doubt hold a grey area. The confirmation bias comes from when the writer was discussing what he was looking at: how search data often shows us who people really are in comparison to their online lives. However, to only have data which massively confirms your point makes the people who regularly read research in the audience not only question how this is research but makes them question the entire point of the study. These facts have not only been cherry picked, but some of them are also (in research) able to be easily disproven.

All in all, I would refer to this book as ‘pop statistics’ or basically then McDonalds of data science. It’s a fun read and is often entertaining and you’re able to laugh along with the author. It is an accessible and easy read for those looking to get into the subject of reading about data science. However, it should not be taken seriously. This is a book with obvious talking points that have been done to death, shoddy research methods and on top of that, a huge amount of confirmation bias which offers no leg room to alternative view points. Fun does not always mean insightful.