- Joined
- Feb 28, 2007
- Messages
- 9,786
I almost want to say- get deeper into the stats. of course, a larger sample size would help with that, but even n=5 you can have some fun.
Christof, I think we could have a fun conversation on this stuff, but yeah, there are limitations. The rule of thumb is 5 times the replicates for every group and dependent variable which quickly becomes impossible. Still, in the realm of multivariate states, the real power is in its ability to reduce dimensionality and make broad inferences across a set of covariates. Does any of that really nail down which is the best knife? No. But it just might point out to Tonym that he has an unhealthy and unfounded fondness for scandi grinds