| Lab website | https://motivationsciencelab.com/ |
| ORCID | https://orcid.org/0000-0003-2902-9600 |
| Google Scholar | https://scholar.google.com/citations?user=r0AJphIAAAAJ&hl=en |
RT @JPubEcon
Teachers take classroom behavior into account when grading achievement exams.
This is an inefficient way of assessing students’ knowledge, explains part of grading discrimination against boys, and reveals a causal effect of noncognitive skills that is unrelated to proficiency.
🤔 Which is causally correct?
RQ: Studying math influences history grade?
A: Increasing math study time by 1 hr has no influence on history grade.
B. Wait, if you study math more, you lose time to study history. So causal effect is negative.
One week before the deadline for this data science junior research group leader/postdoc position (5 years)!! I especially look for those who have good knowledge in #ML (not maximum-likelihood, just to be sure :-).
This is really a nice paper for behavioral scientisits on the critical discussion of interpretable machine learning. Very easy to follow. One concrete take --- many feature importance metric represent marginal, rather than partial effects.
By Mirka Henninger, Rudolf Debelak, Yannick Rothacher, & Carolin Strobl
The German word "hören" means "to hear". But it has funny relatives:
anhören = to listen
zuhören = to listen (a bit more active than anhören)
umhören = to ask around
abhören = to wiretap, to auscultate
überhören = to not hear
erhören = to hear and react to (eg. a plea or a prayer)
gehören = to belong to
angehören = to be a member of
satthören = to get enough of hearing
Should I rather aufhören*? OK.
Enjoy our language!
___________________________
*aufhören = to stop
#Causalinference meets #measurementmodels. Must read by VanderWeele. All measures are formative, all measures are reflective, but both are fallacy.
He claimed the importance of examining item-specific causal effects. Sorry for the self-promotion, but this is indeed what we also recently argue for ...
Please use our 'Naturalistic Neuroimaging Database' (NNDb v2). It consists of N=86 people who did a battery of behavioural tests and watched one of 10 full-length #movies during #fMRI. It has 46,249,600,000 voxels or 6.69 days worth of data to play with. 🥳
Data: https://openneuro.org/datasets/ds002837/versions/2.0.0
It is also available on Neuroscout: https://neuroscout.org
Read the paper: https://www.nature.com/articles/s41597-020-00680-2
We are now collecting a 3T NNDb with adolescents and adults, and before/after #psychedelics.
Pietro Biroli
Jens Clasen
The LAB Lab 🏳️🌈🏳️⚧️
Patrick Mineault