3 ms·
I think all of the test & measure stuff we talk about so much in the web context -- A/B testing, analytics, etc -- would revolutionize the world if it were appl
by patio11 17y ago
I think all of the test & measure stuff we talk about so much in the web context -- A/B testing, analytics, etc -- would revolutionize the world if it were applied with one tenth the dedication by human institutions which are not teeny little web startups.
This isn't to say that we have the patent on The Scientific Method. However, there are darn few organizations which seem to have it written into their DNA. Toyota does, famously. Imagine what would happen if your typical big-city school system did. ("First order of business: we had 37 A/B tests running in first grade classrooms last week. 31 of them did not lead to substantial results. Let's hear from our lead statistician about one promising innovation, which improved vocabulary retention among our lowest performing quartile of black students by 15%...")
- staunch 17y agoMultivariate testing is an extremely limited tool. It's great when you're doing everything else right and not much better than rearranging the deck chairs on the titanic in ever other case. I think it's dangerous (but very tempting) to try to apply it too generally. One passionate visionary thinking outside the box can do more than 1,000 statisticians ever could.
- teeja 17y ago"... what would happen if your typical big-city school system did." Education science never made it into US K-12 education. We Americans have a very fluffy concept of what education's about. We worry more about socialization than education (Neil Postman was right). For a looooong time there have been US educational research journals with data far ahead of current practice, but many US educators never get much chance to see them. And I've never heard of a US (K-12) school that's doing anything empirical (beyond, oh, statistics anyway).