Twitter | Search | |
Justin Reich
MIT Teaching Systems Lab | Fellow | Founder , Blog …. CV:
10,104
Tweets
1,221
Following
6,278
Followers
Tweets
Justin Reich Apr 29
we won at but also had a fine showing at ... :) two conferences
Reply Retweet Like
Justin Reich Apr 29
The Paradox of Routine: very nicely illustrated by DuoLingo
Reply Retweet Like
Justin Reich Apr 29
if I were in rva I would gladly help you determine
Reply Retweet Like
Justin Reich retweeted
Engagement Lab Apr 29
Watch discuss "The Civic Mission of Schools" at the April event
Reply Retweet Like
Justin Reich retweeted
ScratchEd Team Apr 29
Join us at Harvard for four days of , August 8-11.
Reply Retweet Like
Justin Reich Apr 29
Thank you!
Reply Retweet Like
Justin Reich Apr 29
Can someone tweet that coding manual link?
Reply Retweet Like
Justin Reich retweeted
Greg Kulowiec Apr 29
Reply Retweet Like
Justin Reich retweeted
Doug Clow Apr 29
Liveblog notes from Robert Mislevy keynote : A dispatch from the psychometric front
Reply Retweet Like
Justin Reich Apr 29
Here is Andrew Gelman and Eric Loken on the Garden of Forking Paths
Reply Retweet Like
Justin Reich Apr 29
Sorry for broken OSF link. One of two is now working: Other needs colleagues in US.
Reply Retweet Like
Justin Reich Apr 29
JK, here's one . Sorry about that. Need to wait til Mike wakes for the other one.
Reply Retweet Like
Justin Reich Apr 29
. I can't until the IRS finishes auditing me. But at that time I will release all documents.
Reply Retweet Like
Justin Reich Apr 29
When a psychometrician says "that's a terrible idea", that's an excellent predictor of the next likely policy push.
Reply Retweet Like
Justin Reich Apr 29
Saying "Make better assessments" is really hard, and means massive, complex, interdisciplinary projects, and Mislevy walks the talk.
Reply Retweet Like
Justin Reich Apr 29
Or, if you want to improve learning analytics, improve the quality of the activities and assessments rather than algorithms
Reply Retweet Like
Justin Reich Apr 29
If you want to improve pyschometrics, improve the quality of situations to demonstrate evidence rather the quality of models.
Reply Retweet Like
Justin Reich Apr 29
Presumably some subset of lower level actions is more predictive than any particular strategic sequence? If you never build solar...
Reply Retweet Like
Justin Reich Apr 29
In a very deep way, Twitter's recommender algorithm's understand me...
Reply Retweet Like
Justin Reich Apr 29
Assessment is fair if it's different for every person but gives us same evidence for a construct. Mislevy
Reply Retweet Like