Very excited about this new paper PNASNews:

pnas.org/doi/10.1073/pn…

Key Q: what predicts how much young kids (👶)talk?

How much 🗣 kids heard predicted how much 👶talked, but other factors, e.g. mom’s education, didn’t. #PsychSci #DevPsy 🗣💬

INCOMING SUMMARY🧵ALERT 1/14

W/Alex Cristia +11 more great coauth’s , we asked what predicts how much 👶🗣️in daily life.

we answered w/LENA data:

A machine learning algo (🤖) took in kid-centered audio, & spit out estimates of who’s talking, when, & for how long.

There r legit prob's w/automagic processing but real benefits too: we pooled 18 datasets & identically processed >40k hrs audio (43 langs!) 12 countries, clin & non-clinical pop’s, etc.

Data curation was 😓, but by-hand would’ve taken ~500 people a yr of fulltime work 🤯3/14

(What’s that? You wanna play with speech tagging? Cool, Alex Cristia's team led by Marvin Lavechin created a new speech tech algorithm that gave us the same results as the LENA one, and it’s OPENLY available here: arxiv.org/abs/2005.12656 . Use it! #OpenScience ) 4/14