One motivation for this is that I’ve recently moved to Hamilton, Ontario and started a new position as an assistant professor at McMaster University. I’m aiming to do a better job publicizing the research I do in this new position.
Stoked to do a thread on results from a recently accepted paper! Our team searched high and low for meaningful links between personality and morphometry (i.e., brain structure) using mixed effects models and machine learning....and didn't find much. (1/n)
https://t.co/oLvOsJ1rUB
@RochePhD @JMill_er Depends on the site obviously, but definitely $1mil plus. Combined with how long it would take to collect 2,000 participants at one site, this suggests more consortium-based efforts are necessary for replicable findings in neuroimaging.
@JMill_er @lpsatchell@CourtlandHyatt @drl4567 Yes, I do believe that’s the only ffm trait they have. But probably would be worthwhile to look if it has robust structural correlates in that dataset at some point (if someone hasn’t done that yet)
Vaping helps people quit smoking, is much less harmful than smoking, and getting people to change from smoking to vaping could save millions of lives in the next few years. So why is the WHO trying to ban e-cigarettes and hinting they're as bad as tobacco? https://t.co/6Qq0QfgSUV
Out now from our soon-to-be-finished spec issue of @PNeuroscience! fMRI investigation of relations btwn Extraversion, internalizing psychopathology, & neural activation following reward receipt. This one by @CourtlandHyatt @JMill_er et al. @HiTOP_system https://t.co/9HbnO2ii2D
@PhilSpechler Thanks! That's a great thought about subgroups- in its simplest formulation it would be easy to check if effect sizes go up if you split by sex (I'd bet they would a tiny bit). They would probably go up more if your made groups based on cluster analysis or something like that.
New pre-print describing effect size (ES) in the ABCD study. We examine correlations among 161 psychological & demographic variables, representing most instruments in ABCD. Most ES are “small” using Cohen’s heuristic - median ES: r=.03, 90th %ile: r=.14. https://t.co/6hA8OJZ2zs
In short, for associational psychology research using univariate methods, it’s probably not realistic to use Cohen’s heuristics for small/medium/large ES.
We report ES benchmarks, showing “meaningful” correlations that are r<.2 – e.g., age x pubertal development (r=.17), sleep problems x overall cognitive function (r=-.06), family psychiatric history x child psychiatric problems (r=.15), impulsivity x detention frequency (r=.08).
Exaggerating Harmful Drug Effects on the Brain Is Killing Black People https://t.co/OA7gK62fLw
This is my open letter to @NIDAnews, drug researchers, practitioners of the politics of respectability, among others. You've got to do better b/c you're killing us.
New post-print: in N=11,052 children (ABCD data), impulsive traits (UPPS-P) associated w/ regional cortical surface area, thickness + WM tract FA. Effect sizes very small (R^2<.02). Results converge in univariate analysis & cross-val elastic net regression https://t.co/uLZ8muliGZ