The book takes a data-driven approach to analyze the world.
- Both married and unmarried people exaggerate how often they have sex. Search for sexless marriage is 3.5 times more common than an unhappy marriage.
- On Obama’s elections results night, there were more searches for nigger president than the black president. Racist searches were no higher in Republican states than Democrats states.
- A Trump supporter would search for “trump Hillary …” While a Hillary supporter would put Hillary’s name first.
- Liberals search for “estate tax,” while conservatives search for “death tax”.
- Back pain does not correlate with pancreatic cancer. Back pain + yellowing skin does.
- Warm weather is twice as strong of antidepressants as the leading drug.
- The size of the heart, the left ventricle, predicts how good a horse is at racing.
- In the prediction business, you care about what works and not why it works.
- Women prefer saying tomorrow, so, “sooo,” etc. more than men.
- Social desirability bias leads one to lie in an in-person survey.
- “Regret not having children” more common search before children. “Regret having children” more common afterward.
- 9% more young girls are in gifted child programs, but parents search “is my son gifted” 2.5 times more. More sons are obese, but parents search for daughters being obese more often. No conservative or liberal difference in such a bias.
- Your chance of seeing an opposing viewpoint is higher online than offline.
- Google searches are digital truth, while social media posts are digital lies.
- Netflix used to ask the users for a wishlist, and people would give high brow aspirational movies they would never watch. It stopped asking and started predictions instead.
- Many of our random behaviors are formed because of certain events that happened during our childhood. A popular president during 14-24 years of your age will shape your views for life.
- Anxiety is higher in low education areas than in big cities.
- In the US, for the rich, the city of residence does not affect life expectancy. For the poor, it makes a huge difference. Having more rich people in the town is good for poor people’s life expectancy.
- College Town produces excellent outcomes because of the Gene pool and early exposure to innovation. Being in big cities helps too. It is great if the town has more immigrants as well.
- Violent movies reduce crimes as young men spent time in the theatre instead of drinking in the bar.
- Doppelganger searches are extremely powerful in predictions, including medical diagnosis. PatientsLikeMe.com
- Super bowl ads work because the revenues do go up.
- Keeping up with the Joneses – neighbors of lottery winners have an increased likelihood of bankruptcy due to overspending.
- Regression discontinuity analysis – do people barely missing a school entrance by a point worse off in like than those who make it? – No
- Getting into a great school like Harvard but not attending had a similar financial outcome to attending it.
- Curse of dimensionality – if you are testing too many dimensions, some will end up predicting just by chance.
- One fewer star on Yelp => 5-9 % drop in revenue