Is there a more comprehensive Statistics choosing framework than the following:
from Charlie Kufs from Stats with Cats Blog
Asked
Active
Viewed 5,285 times
1
-
7I am sorry, but this flowchart seems incredibly incomplete. It seems to pack a lot of information but it it actually very shallow in he advice t gives like "cluster analysis" (what type of clustering, how does data size factor into the decision, what type of that can this algorithim cluster well) ; or Regression (what type? Lm, GLMs, GAMs, LASSO,etc...). Every analysis can be used in many different contexts ( I mean, anova has a regression interpretation... ) and this flow chart dangerously simplifies the choices you have to make in order to analyse data. – Guilherme Marthe Jun 15 '17 at 03:59
-
3That's far from the worst one of those I've seen but suffers from many of the usual problems (ignoring great swathes of statistics, for one thing), and at least it considers one parametric-but-non-Gaussian analysis (though ignoring all the rest). The main problem, however, is that even a considerably more complete one still bases what's being done on the form of the data rather than the specifics of the question being asked of it (though the above flowchart at least considers that there's more to analyzing data than hypothesis tests). What you want to find out should be first and foremost. – Glen_b Jun 15 '17 at 04:45