Then the restless spirit of Paul Graham sat on my shoulder and told me to turn it into an amazing web service that would let people assign themselves into self-experimental cohorts, where they’re algorithmically assigned to balanced blocks so that effects of overlapping interventions can be teased apart.
So it’s a web service that would spit out a random Latin square and then run ANOVA on the results for you?
I don’t think I’ve heard of such a thing. (Most people who would follow the balanced design and understand the results are already able to do it for themselves in R/Stata/SPSS etc.) Statwing.com might have something useful, they seemed to be headed in that direction of ‘making statistics easy’.
I was imagining a site that would look at all the different things you’re trying at the moment, look at all the things other people are trying, and give you a macro-schedule for starting them that works towards establishing cyclicality across all users.
It could also manage your micro-schedule, (prompt you to take a pill, do twenty sit-ups, squirt cold water in your right ear, etc.), ask for metrics and let users log salient information and observations. Come to think of it, once that infrastructure is already in place, there’s no reason you couldn’t open it up as a platform for more legitimate and formal trials.
Mm. So not just scheduling your own interventions but try to balance across users too… No, I don’t know of anything like that. CureTogether actually got some research published, but I don’t think randomization or balancing was involved. (And trying to get nootropics or self-help geeks to collectively do something is like trying to herd deaf cats into pushing wet spaghetti...)
So it’s a web service that would spit out a random Latin square and then run ANOVA on the results for you?
I don’t think I’ve heard of such a thing. (Most people who would follow the balanced design and understand the results are already able to do it for themselves in R/Stata/SPSS etc.) Statwing.com might have something useful, they seemed to be headed in that direction of ‘making statistics easy’.
I was imagining a site that would look at all the different things you’re trying at the moment, look at all the things other people are trying, and give you a macro-schedule for starting them that works towards establishing cyclicality across all users.
It could also manage your micro-schedule, (prompt you to take a pill, do twenty sit-ups, squirt cold water in your right ear, etc.), ask for metrics and let users log salient information and observations. Come to think of it, once that infrastructure is already in place, there’s no reason you couldn’t open it up as a platform for more legitimate and formal trials.
Mm. So not just scheduling your own interventions but try to balance across users too… No, I don’t know of anything like that. CureTogether actually got some research published, but I don’t think randomization or balancing was involved. (And trying to get nootropics or self-help geeks to collectively do something is like trying to herd deaf cats into pushing wet spaghetti...)