Logo

MB6: Neonatal Imitation workshop #1 - Shared screen with speaker view
Laurie Bayet
42:49
may be necessary to reshare your screen and check the “share audio" box?
Michael Frank
42:54
Sharing settings, advanced "share audio"
Michael Frank
43:01
Laurie beat me to it. :)
Heidi Baumgartner
45:35
https://docs.google.com/document/d/14K-4NLD2m8YgGZteJIh7YoSXyGMwAaJQcIs9Xpm-h14/edit
Heidi Baumgartner
45:53
everyone is welcome to add notes, questions, thoughts to the linked document above
Krista Byers-Heinlein
01:26:11
Heidi, you might want to re-post the link to the agenda, for those who joined late
Grace Zhou
01:26:48
Here it is: https://docs.google.com/document/d/14K-4NLD2m8YgGZteJIh7YoSXyGMwAaJQcIs9Xpm-h14/edit
Kiley Hamlin
02:08:10
My own experience running “newborns" after we got kicked out of hospital due to Covid is that parents of newborns revert to the oldest possible day they can come in -- as in, we'd get a lot of 6 week olds
Michael Frank
02:09:43
@laurie: Pooling across labs works OK - from a multi-level models perspective you just want random age slopes by lab as well as a fixed effect of age across labs.
Laurie Bayet
02:10:03
thanks!
Michael Frank
02:10:17
(in MB1 we did a meta-analytic approach that would mean that incomparable ages across labs would be hard to deal with, but we've moved away from that analytic strategy in subsequent projects, which all use mixed models)
Laurie Bayet
02:11:07
great point above also from Kiley Hamlin re: recruitment. we didn't see this online (there was a decent range within the 0-45 eligibility window) but in the lab it may be harder to recruit younger babies if the window is large?
Kiley Hamlin
02:25:53
+1 on this being something foundations might be interested in!
Kiley Hamlin
02:26:58
(or, maybe that's a related point to what Mike said ;)).
Michael Frank
02:27:41
@kiley - this feels like it connects well with Templeton priorities.
Kiley Hamlin
02:28:30
Sorry all I didn't realize I was messaging a select group -- just noting this in particular feels like a very foundation-friendly topic
Michael Frank
02:35:59
@Jonathan - note that MB2 is doing exactly what you are describing, that is doing some pre-registered analyses on trial 1 but then analyzing and modeling carryover for multiple trials.