Twitter | Search | |
Richard McElreath
Studies human evolutionary ecology & wrote an applied Bayes stats book: Occasional blog:
2,771
Tweets
539
Following
6,813
Followers
Tweets
Richard McElreath 23h
Replying to @Voovarb @bjjbarrett
Right, let's you add arbitrary constraints to the parameter definition.
Reply Retweet Like
Richard McElreath Nov 13
Ecology seems extra hard when sampling needs to be part of model. e.g. undercounting low N spp.
Reply Retweet Like
Richard McElreath Nov 13
looking at definition of the outcome, I am tempted to a count model on the underlying frequency vectors. That might be quite a project though, unless someone has already developed such a model.
Reply Retweet Like
Richard McElreath retweeted
Julia Rohrer Nov 13
Curious about Open Science? On Thursday, nobody less than will give a talk at . Everybody is welcome!
Reply Retweet Like
Richard McElreath Nov 13
Replying to @fusaroli @polesasunder
My instinct is that visualization will depend upon meaning of the zeros (and ones in your case).
Reply Retweet Like
Richard McElreath Nov 13
Replying to @fusaroli @polesasunder
Seen it as a zero-inflated hurdle (detection) model. e.g.
Reply Retweet Like
Richard McElreath Nov 12
Replying to @GoatsThatStare
I was reading about this service that simplifies Amtsdeutsch, because so many native speakers have trouble with it.
Reply Retweet Like
Richard McElreath Nov 12
Reply Retweet Like
Richard McElreath Nov 12
For those near Cambridge, I will speak on Wednesday this week in the bio anthro seminar series. Usual topics: human evolution and model-based inference. Details and abstract:
Reply Retweet Like
Richard McElreath Nov 12
Replying to @DavidPoe223
Reply Retweet Like
Richard McElreath Nov 12
Replying to @DocEd
Reply Retweet Like
Richard McElreath Nov 12
Replying to @swamids1
For the 2nd edition:
Reply Retweet Like
Richard McElreath Nov 11
Working chapter title.
Reply Retweet Like
Richard McElreath Nov 11
Working up a teaching module for Bayesian instrumental variable regression, and I stumble across this tutorial by , who has of course already done my job for me.
Reply Retweet Like
Richard McElreath Nov 10
Replying to @WhisperinJay
Would be useful, since most of stats in psych and bio is generalized linear models, and these are very similar in Bayes and non-Bayes. But since my book is Bayesian, it won't describe majority practice that includes lots of p-values and ad hoc tests.
Reply Retweet Like
Richard McElreath Nov 9
It's , so in Germany there is always a chance something history making will happen today.
Reply Retweet Like
Richard McElreath retweeted
Dan Simpson Nov 9
This popped up in a hydrology application this week: given that no one believes that the climate isn’t changing, what should the null be?
Reply Retweet Like
Richard McElreath Nov 9
Replying to @rlmcelreath
Here's a thoughtful article from 2010 on the philosophical and empirical issues. Begins with a classic Zen koan.
Reply Retweet Like
Richard McElreath Nov 9
Replying to @rlmcelreath
A broader issue in inference in biology: These phenomena are too complex to address with "null" models. There are no unique null models in population dynamics. Require explicit alternatives.
Reply Retweet Like
Richard McElreath Nov 9
This article is pretty good, but the title should have been "Neutral Theory Isn't Even Wrong"
Reply Retweet Like