Michael Lee

@AtomicCross @Askwhyisit blog.id.com.au/2011/id-news/f…

@Askwhyisit Was that a Jedi at the end?

A multi denominational interfaith dinner….

In Psych Rev with @jesus_mvu and @AlxEtz. A predata measure of model falsifiability and complexity. Examples for hierarchical and nested models challenge common intuitions, and a final example shows the usefulness of incorporating informative data priors. psycnet.apa.org/doi/10.1037/re…

We're hosting MathPsych / ICCM /EMPG! We will welcome abstract contributions related to research into mathematical and computational theories of human cognition. The call for abstracts will be in February 2023. Come join us in Amsterdam this summer!

@JimGrange The book was largely written 15 years ago and both jags and my understanding of it have improved. A new version of the book is finally underway. Not a second edition or a second volume but a do-over. Im sure there will be a better GCM chapter in it.

@JimGrange w[k, 1:4] ~ ddirch(rep(1, 4)) and then w[k, 3]. I think you have to be explicit about the 1:4 on the LHS, and can’t just use w[k, ] ~ … The GCM code in the book should work, but (like a lot of the code in the book) could be cleaned up quite a bit.

I'm trying to declare prior, w, (in #jags or #stan), which is a vector (length 4) of attention weights, and must sum to 1. When I have just 2 attention weights, I use w ~ dbeta(1, 1), and I refer to "w" and "1 - w". How do I declare the vector, w, of length 4 summing to 1?

@drewhalbailey Such Libra indecision.

This is good advice. For some reason, Twitter was sending me tweets about 10 cryptocurrencies and 4 astrological signs. Not sure which I’m more upset about.

@mhtessler @gershbrain I’d suggest reading about “parameter space partitioning” and “landscaping” from the 2000s (jay myung, mark pitt, Danielle navarro, and others) for what has been done. And then wolf Vanpaemel’s recent data prior paper for a vision of the future.

@gershbrain I came into this issue a lot when doing Bayesian model comparison, that qualitative patterns were not part of the model selection criteria. Seems like the prior predictive check would be the way to go, but maybe @mdlBayes has thoughts too

I've come across an interesting phenomenon, which I'm sure has been discussed before but I don't know the right references.

The Department of Cognitive Sciences at the University of California, Irvine (UCI) invites applications for an assistant professor (tenure-track) position with an anticipated start date of July 1, 2023. Details here: Assistant Professor in Human Vision recruit.ap.uci.edu/JPF07912

I will be recruiting PhD students this upcoming cycle, so if you have motivated students interested in modeling memory and the brain, please send them my way (email: qiong.z@rutgers.edu)! Lab information: sites.rutgers.edu/memory-optimiz…

@dggoldst I think I remember @david_lagnado casually saying at a Dutch royal academy workshop “the last time I was in Amsterdam I was playing bass for van Morrison”. Not sure about permanent band membership or whether van Morrison is “rock” but it’s close.

What academics used to be in rock bands (besides Brian May and Carolyn Bertozzi)?

@JCSkewesDK *bear* with me, perhaps?

@JCSkewesDK @AndyPerfors Parenthetically, “Lee” is very hard to sportify. My brother got “Leesy” a bit. I mostly got “Bruce” then “General” until I joined a cricket club that already had Dave Patton, so “General” was taken, and I ended up with “Tiger Chung”.

@JCSkewesDK @AndyPerfors Perf and Perfs sound about the same and equally (moderately) good to my Adelaidean ear. In any sporting context, though, (and that is the point of all abbreviations) it needs to be Perfo.

And, yes, some journals (and action editors) do not do this, either providing detailed assessment, or at least highlighting the key content that underpins their decision. Which brings me back to the original question.

No, nothing of mine has recently been rejected to prompt this. Yes, these are journals for which I have either action edited in the past or been on the editorial board for a long time, so I think I understand some of the relevant work-flow and work-load context.

Is it reasonable to expect an action letter to demonstrate concretely that the editor has read the paper and reviews? I see a lot of action letters that are free of any submission-specific content. Basically one generic template action letter for each decision.

@NoContextBrits I had only heard the pronunciation on South Park, and thought it was a running gag. When I moved to the US, a new American friend used the pronunciation introducing her husband. So I roared laughing at both of them appreciating the edgy pop culture reference. Good times.

The way Americans pronounce the name ‘Craig’ should be a crime.

And people probably construct rosters that are too representative of what is needed in an active roster each week. It looks like not drafting kickers and defenses, in order to draft more running backs and wide receivers, might be a good under-used strategy.

The distributions of when different types of players are picked shows all sorts of interpretable regularities

Just published in Judgment and Decision Making with Siqi (Susie) Liu (not on twitter): an empirical analysis of how people draft their rosters in fantasy football. Real world data with interesting insights into competitive sequential decision making. journal.sjdm.org/22/220318/jdm2…