I do a power analysis when I just have a vague idea for a new study
-
-
-
I think people need to get serious about power analysis. 80% power is way too low to aim for. 95% at least. This enables easy interpretation of null findings because the high precision means that it will be an informative null, just a "maybe sampling error, lols" null.
End of conversation
New conversation -
-
-
Well, n<100 doesn't always have to be that terrible. For a really strong effect it can do (even though it is obviously suboptimal).
-
It's a rule of thumb. Most effect sizes in psychology are too small to be seen in n < 100 datasets. Those that are large enough are probably fake or due to data/analysis error. E.g., those studies with d = 2 for stereotype threat.
End of conversation
New conversation -
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.