Charles Driver

Corner zones are the most dangerous though.

Ok indeed. Why I went back to optimisation :)

If these things really do keep you awake but the abbreviated form above is too abbreviated, see osf.io/preprints/ps... for a start.

Your dynamic systems models aren't complicated enough? Skewed distributions keep you awake? Your dynamics dynamic differently at different dynamic levels? Here's a not-very-beginner friendly outline for addressing the issue with ctsem cdriver.netlify.app/post/ctsemas...

The distribution of the individual parameters involved might be kind of poor, but the marginal result should be fine, particularly in a Bayesian sampling context.

I doubt it honestly, even just optimizing these parameters in a tricky modelling context (dynamic systems, generating data model doesn't match fitting data model), ctsem is able to approximately recover the overall distribution just fine. Looked into it last week: cdriver.netlify.app/post/ctsemas...

When I briefly looked into this just the other day I decided the sinhh-arcsinh transformation was for me, parameters at least seemed more natural. Curious if you or anyone reading has pro/con thoughts though!

I'm happy as long as there's a narrative.

Fair, but they're already working on that...

a concrete example would really help, because to me it sounds like you're talking about typical mediation / moderation, but with some kind of arbitrary discretization.

I thought people were generally happy to claim some cause as soon as they can explain any variance and imagine they have controlled for other explanations... do you have any examples of this apparent need to explain most / all variance? Have never seen such a thing.

Currently wishing someone had written a guide to chasing all the reasons embed-resources fails to work for html output, so sympathizing with thread! It would be less frustrating if it wasn't generally so good.

when 'academic freedom' looks nice at the individual level and strangles science at the herd level...

You seem to suggest the burden is on the authors... why not shift it higher up the chain where there aren't the personal incentives? If state and other major funding agencies mandated specific journals / journal criteria, we'd be at least halfway to fixed as far as I can naively see...

a single cause as an amalgam of multiple variables? can you give an example of that?

Have long thought this would be fun, if only to see which approaches at least attempt to throw their hands up and say 'I don't really know' with appropriate regularity. Ping me if you want to chat on it sometime.

Tw/bleet summaries are so much better than abstracts generally :/

Well yeah, actually it wasn't related to x (I dislike that story), just, *gestures broadly*, the universe.

Learning now but still curious about the toaster.

I'm not usually called Marcus, but fair point nevertheless.

Are we really hoping to change their behaviour though? I would assume the public shaming stuff is about convincing others.

In case I can reach the @quarto.org and #rstats and windows users driven mad by the inability to render files when stored on dropbox (there must be billions of you right?), here's an Rstudio add-on that uses temp files instead:

github.com/cdriveraus/Q...

So... a rant shouldn't be a rant?

To be fair, the approach to building new housing in many countries kind of was.

I tried but it didn't help the young men...

Australia. Panadol (paracetamol) would have been closest for me...

Not in my English...

UZH: LECTURER RESEARCH IN DIGITAL HEALTH COMMUNICATION (PERMANENT POSTDOCTORAL POSITION) AT IKMZ – DEPT. FOR COMMUNICATION & MEDIA RESEARCH The lecturer position is a permanent postdoctoral position (100%, paid according to cantonal salary scheme; 123K CHF p.a.) with optional promotion to senior lecturer Workplace is Zurich

Excellent job opportunity
@ikmz.bsky.social!

- Work: Digital Health Communication
- Contract: Permanent
- Salary: 123k
- Location: Zurich

jobs.uzh.ch/offene-stell...