3 ms·
Which is why you need high-quality and large-scale data on the relationship between income and internet usage and the relationship between income and happiness
by saulrh 2y ago
Which is why you need high-quality and large-scale data on the relationship between income and internet usage and the relationship between income and happiness so you can chain those together and subtract the whole thing out. Yes, error bars stack up rapidly. That's why you need 2 million data points to get a reasonable conclusion, when you only need ~1000 to get high-quality information on things like consumer preferences or political affiliations. Good thing that we have plenty of people that are very interested in things like "how much happier does money make you" and full-scale datasets like that are produced on a regular basis!
Really, the math isn't any harder than what's necessary to land a rocket screaming along at mach 8 or reverse a couple terabytes of voltage readings into a nice clean volumetric medical image.
- laylower 2y agoExactly. 2.5 mil is not peanuts. Yes, you can always get more data. But let's not pretend 2.5 is not sufficient to establish statistical relationships.
- disgruntledphd2 2y agoSure, but it's a large sample based on self-report instruments. This is definitely evidence, but I do share the OP's concerns around controls and would regard this finding as somewhat likely to be biased by the kinds of people who both a) have internet and b) are willing to respond to a really, really long poll.