Measurement invariance in nationally representative samples of whites, blacks, hispanics -- n = 2000. KABC-II and KTEA-II batteries. https://www.gwern.net/docs/iq/2016-scheiber.pdf …pic.twitter.com/FLhvLoQup8
Show this thread
@JDHaltigan study has an example of what I was talking about: one can definitely find violation of MI, but it will be tiny in effect size.
https://www.gwern.net/docs/iq/2016-scheiber.pdf …pic.twitter.com/Jf8YFME46C
-
-
(With sufficient sample sizes, the null hypothesis will always be rejected, since 'everything is correlated', especially in psychology. The point estimates, however, will usually be trivial - 'bet on sparsity'!)
-
Is that a quote from somewhere or just the gwerny summary?
-
Positive manifold related but moreso sensitivity of chi-square to sample sizes. Hence change in RMSEA and CFI have come into a bit of favor
-
Tests which aren't based on our true decision-theoretic criteria will never leave people fully at ease, but I'll definitely take RMSEA/CFI any day over p-values, especially for n>1000s...
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.