|
Dale Barr
@
dalejbarr
Glasgow, UK
|
|
human social interaction and communication, data analysis in #rstats
|
|
|
4,567
Tweets
|
791
Following
|
1,705
Followers
|
| Tweets |
|
|
Dale Barr
@dalejbarr
|
13h |
|
apparently Bayesian SEM is a thing if you really want to torture them github.com/ecmerkle/blava…
|
||
|
||
|
|
Dale Barr
@dalejbarr
|
13h |
|
linear-mixed effects models: a special case of structural equation models. From Yves Rossell #useR2017 talk on SEM channel9.msdn.com/Events/useR-in… pic.twitter.com/6hfHP9XrGu
|
||
|
|
||
| Dale Barr retweeted | ||
|
The 100% CI
@the100ci
|
Jul 20 |
|
We're looking for input for a new blogpost: What are weird "rules of thumb" in psychology that you know of or have been taught?
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 20 |
|
but only test random slopes... *never* test whether to include random intercepts or fixed effects in the model pic.twitter.com/5Ax2SxWwe1
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 20 |
|
but you *can* do data-driven model speification with mixed effects models, and indeed you MUST! pic.twitter.com/RArSxfuJuK
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 20 |
|
you can't do data-driven model specification with ANOVA (whoops! you can!) pic.twitter.com/Nrgd93i3N6
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 20 |
|
ANOVA only "allows" one observation per subject per cell pic.twitter.com/jvr27LaIEu
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 20 |
|
alright then! myths about ANOVA/mixed models
1) mixed-effects models are more powerful than ANOVA (they are certainly easier to misuse!)
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 20 |
|
"truncate/discard RTs 2.5 SDs beyond the mean"
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 20 |
|
are you looking for "myths" or just rules of thumb?
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 14 |
|
yeah JP I don't know what you're on about
|
||
|
|
||
| Dale Barr retweeted | ||
|
Titus Brown
@ctitusbrown
|
Jul 14 |
|
osfclient: a Python library and command line client for @OSFramework / osf.io file storage interaction. osfclient.readthedocs.io/en/stable/
|
||
|
|
||
| Dale Barr retweeted | ||
|
Jake Westfall
@CookieSci
|
Jul 13 |
|
Periodic reminder to experimenters: in balanced factorials, high-order interactions not generally inherently harder to detect than low-order pic.twitter.com/DbkRVUXBcn
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 14 |
|
typo of the day:
demographic -> demongraphic
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 14 |
|
Yes, I'll weigh in: WTF?!
|
||
|
|
||
| Dale Barr retweeted | ||
|
Ben Jones
@Ben_C_J
|
Jul 12 |
|
This is the project I am most proud of. No evidence for links between hormonal status and masculinity preferences.
biorxiv.org/cgi/content/sh…
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 12 |
|
let's do it! open teaching FTW
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 12 |
|
Found this heartwrenching note today on Belmont Ave in Glasgow. Will return to its rightful owner, in case it's needed for legal purposes pic.twitter.com/NbF3cZq28o
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 11 |
|
on the bright side, use of ANOVA has greatly diminished in favor of LMEMs, but yeah, F1xF2 lives on
|
||
|
|
||
|
|
Dale Barr
@dalejbarr
|
Jul 11 |
|
thanks! actually I noticed the typo... right after I tweeted it, of course ¯\_(ツ)_/¯. Fixed in the preprint I posted earlier today
|
||
|
|
||