I am yet to be convinced that 4000 data points are sufficient to extrapolate how happy 2.8B people are in the world. (India and China) Especially when it deals with a complex topic as happiness without taking any cultural differences into account.

People on HN tend to argue it’s sufficient data to be statistically significant, but I don’t see how.