Notes from Everybody Lies, by Seth Stephens-Davidowitz
Everybody Lies is the best airport book I’ve ever read. It is this generation’s Freakonomics.
Summary: A fun exploration of some interesting results from recent literature exploring the rich new datasets provided by internet search, and a thoughtful meditation on the opportunity these datasets offer to economists. The titillating search data is bait, a spoonful of sugar to go with the medicinal call to adventure: there’s an exciting new world of interesting, actionable data out there, and economists should be exploring it with the same vigour they approach the BLS.
- Traditional polling firms have not had a great year (Trump, Brexit, a few provincial elections here in Canada that defied expectations). Can internet search data do any better? Yes. When people search for candidate comparisons, they tend to list the candidates in their order of preference. So, Trump supporters will search “Trump Clinton” while Clinton supporters prefer “Clinton Trump.” Trump outperformed Clinton on this metric throughout the election.
- Clinton was hurt by low turnout in black communities. Her polling numbers were good, but actual searches for voting information (location and hours of voting, for example) were way down from 2008 and 2012.
- Do people turn to humour for relief from sadness? Counterintuitively, no. Searches for jokes are lowest on Mondays, the day when people most often report feeling unhappy.
- Can you guess what are the most common searches that begin with “My husband wants me to…” ? You’ll have to bu the book to find out, because this is a family blog.
I laughed out loud a few times while reading Everybody Lies. There probably aren’t more than ten attempts at humour in the entire book, but when they land, they land.
Final Review, 10/10
(A note on reviews: I assign points relative to genre. So, a perfect score means Everybody Lies is at the extreme right end of the quality distribution for popular economics/science aimed at a general audience.)