This came in the email one day:
We have used the great software Stan to estimate bycatch levels of common dolphins (Delphinus delphis) in the Bay of Biscay from stranding data. We found that official estimates are underestimated by a full order of magnitude. We conducted both a prior and likelihood sensitivity analyses : the former contrasted several priors for estimating a covariance matrix and the latter contrasted results from a Negative Binomial and a Discrete Weibull likelihood. The article is available at:
https://www.sciencedirect.com/science/article/pii/S1462901116301514
Unfortunately (and I know that this is not truly an excuse), given the journal scope and space constraints most of the modelling with Stan is actually described in two appendices. Data and R scripts are available on github (https://github.com/mauthier/bycatch).
Thanking you again for the amazing Stan,
Yours sincerely,
Matthieu Authier
Reference:
Peltier, H. and Authier, M. and Deaville, R. and Dabin, W. and Jepson, P.D. and {van Canneyt}, O. and Daniel, P. and Ridoux, V. (2016) Small Cetacean Bycatch as Estimated from Stranding Schemes: the Common Dolphin Case in the Northeast Atlantic. Environmental Science & Policy, 63: 7–18, doi:10.1016/j.envsci.2016.05.004
That’s what it’s all about.
I don’t get the picture reference
That’s what it’s all about.
I had no idea it was called something different in the UK: https://en.wikipedia.org/wiki/Hokey_cokey (and the “boogie woogie” in Denmark!) Or that the US rights are owned by Sony/ATV Music Publishing … Andrew, did you check with their permissions department?
Now I got it, thanks for the clarification
Mmmmm … hokey pokey!
https://en.wikipedia.org/wiki/Hokey_pokey_(ice_cream)
It’s a bizarre world when the data and the code used to run the analysis is all available for free for anyone to scrutinize, but the paper for it is behind a paywall.
You might want to comment on this at some point (hint: it’s completely out of social psychologists control!)
https://www.nytimes.com/2016/05/29/opinion/sunday/why-do-so-many-studies-fail-to-replicate.html
Numeric:
I remain frustrated that so many people don’t want to even consider the possibility that the original published results in these studies are nothing but noise.
If we start thinking that, what’s left for us to do? We (people using stats to make a point) have to believe what we publish isn’t noise.
The whole idea behind significance was to distinguish random noise from actual effects. Andrew has been very good (as have many others) about pointing out that significance can be misapplied to represent random noise as an actual effect. And, quite frankly, no one is saying that a p-value of .0001 is typically noise or garden of forking paths (and a statistician often finds this in the physical sciences). It’s running through 30 variables and finding two or three significant results at the .05 level that is almost automatically suspect. I remember an article by Friedman in the 80’s (and I can’t find the reference but maybe someone can provide it) where he simulated 50 “independent” variables and consistently found two to three significant relationships for each simulation. So it’s been known for a long time.
” I remember an article by Friedman in the 80’s …”
Look at https://www.jerrydallal.com/LHSP/multtest.htm for an online demo, handy to use in class.
And if you like xkcd’s jelly beans cartoon, follow up with https://www.jerrydallal.com/LHSP/jellybean.htm