I really want to learn statistics, experiment design, and how to avoid or measure or model bias; haven't got a chance to learn it in uni, didn't seem to have much success with the coursera course (sure, I did their assignments, I didn't get much understanding out of it). what do?
-
This Tweet is unavailable.
-
-
Replying to @whitequark @csirac2
How to avoid bias in statistics is like how to avoid having a bug in your code: it's lots of small things, not one big rule. The big rules to know in statistics are things like the central limit theorem.
2 replies 0 retweets 0 likes -
to extend your analogy, what i want is a tutorial, best practices and community conventions for statistics, then
2 replies 0 retweets 1 like -
Replying to @whitequark @csirac2
Different communities have different standards and practices. In the medical world, placebo controlled double-blind studies rule, and epidemiology drools. In paleoclimatology, almost everybody drools. Physicists lock it down tight (six-sigma; none of this "p<0.05" nonsense).
1 reply 0 retweets 1 like -
It's a matter of what sorts of data they typically have to work with, and how loosey-goosey they can get away with being. Both of those vary a lot. In medicine there are lots of experiments you can't do for ethical reasons, whereas physicists can torture electrons all they want.
1 reply 0 retweets 1 like
Anyway, if you know core statistics you can pick up the various methods used in different fields as you go. And that course actually might have taught you the required core, just not in such a way as to make you confident about it. Studying on one's own can be like that.
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.