Lightnews — Scholar-powered news

Ben Bolker @bbolker.bsky.social · 8h

More seriously, I think much of the difference is between quantifying uncertainty by conditioning (profile likelihood curves etc.) and marginalizing (marginal posterior densities etc.)? (If anyone has pointers to nice discussions of this I'd appreciate it ...) [prior \approx regularization ...]

1 4

Ben Bolker @bbolker.bsky.social · 2d

Check the vroom package? (Also possibility of converting to parquet/arrow?)

1

Ben Bolker @bbolker.bsky.social · 6d

A polite query after 3 months sounds reasonable.

3

Ben Bolker @bbolker.bsky.social · 6d

Interesting. Earlier thread with pointers to similar refs (possibly all in your literature review/discussion sections, haven't checked yet ...) bsky.app/profile/bbol...

Ben Bolker @bbolker.bsky.social · Jul 14

Some possibly interesting/useful refs on socioeconomic barriers to success in academia ... comments welcome.

McClain 2025 doi.org/10.1371/jour...;
Morgan et al. 2022 doi.org/10.1038/s415...; Arbuckle et al 2025 doi.org/10.1111/jzo....; Wanelik et al 2020 doi.org/10.1002/ece3....

Figure 4 from Morgan et al 2022. Average income distribution estimated using faculty members’ childhood zip codes (green), compared with the income distribution across the 1998 US population (black). Two densities overlaid: both are skewed with a long/fat right tail. Black (US pop) peaks around $40K, median around $60K. Green (faculty) peaks at $50K, median around $70K.

2

Ben Bolker @bbolker.bsky.social · 8d

404 link? This one works for me, sort of: www.pnas.org/doi/abs/10.1... (but paywalled and so far can't find a way to it via my institution ...)

Meta-analysis finds large variation but no general patterns in the relationship between climate and parasitism in terrestrial animals | PNAS

Climate can vary spatially and temporally and is becoming increasingly unpredictable due to climate change. It can have a large impact on host–para...

www.pnas.org

2

Ben Bolker @bbolker.bsky.social · 8d

This gets discussed a lot (of course) ... scifi.stackexchange.com/questions/47... (plus reddit, www.cbr.com/why-sauron-c... , ...)

Why is Sauron called "the Necromancer"?

During the time period of The Hobbit, a character is mentioned called the Necromancer. Radagast even (in the movie) specifically states that the Necromancer can "summon the spirits of the dead." ...

scifi.stackexchange.com

1

Ben Bolker @bbolker.bsky.social · 8d

maybe fct_rev_infreq <- function(x) {
x |> factor() |> fct_infreq() |> fct_rev()
}
mutate(across(x, fct_rev_infreq))
?

2

Ben Bolker @bbolker.bsky.social · 8d

I think `mutate(across(x, ~forcats::fct_infreq(factor(x)))` might work? (although that puts most frequent grp first: not sure if there's an easy way to reverse). (It annoys me slightly that the forcats::fct_* functions don't have an option to convert their argument to factor automatically ...)

1 3

Ben Bolker @bbolker.bsky.social · 9d

Can't sign myself (not affected) but posting for visibility ...

3

Ben Bolker @bbolker.bsky.social · 10d

Now I want a map annotated with answers to this question: scifi.stackexchange.com/questions/27...

Origin of quote along the lines of "Q: Where do you get your ideas from? A: I get them from Saskatchewan."

Looking for the author and exact words of a quote. A science fiction writer was asked too many times "Where do you get your ideas from?" He came up with a sarcastic reply along the lines...

scifi.stackexchange.com

2

Ben Bolker @bbolker.bsky.social · 10d

Hate to ask, but are you sure that the gamlss actually does *survey* weighting? Docs say only "weighted likelihood analysis where the contribution of the observations to the likelihood differs according to ‘weights’". Can't find anything informative online and don't want to dig into the code ...

2

Ben Bolker @bbolker.bsky.social · 11d

I think I joked once about people spreading Fear, Uncertainty and Doubt en.wikipedia.org/wiki/Fear,_u... about `lmer` -- lmer FUD ...

Fear, uncertainty, and doubt - Wikipedia

en.wikipedia.org

1 5

Ben Bolker @bbolker.bsky.social · 12d

"In view of what Professor and Mrs. Peckham have discovered regarding the habits of Ammophila in a bean patch, it may be that these wasps were busy during the day in the adjoining gardens and repaired at night to this patch of grass to rest and dream of big caterpillars ..."

1 10

Ben Bolker @bbolker.bsky.social · 13d

It helps that I'm 50% in the math&stats department (and hence regularly reminded of my inadequacy in, say, proving anything halfway rigorously)

5

Ben Bolker @bbolker.bsky.social · 14d

nominating this data set github.com/nmouquet/NOU... for #tidytuesday ...

NOUNOURSUSvsBEARS/data at master · nmouquet/NOUNOURSUSvsBEARS

Contribute to nmouquet/NOUNOURSUSvsBEARS development by creating an account on GitHub.

github.com

1

Reposted by Ben Bolker

Nicolas Mouquet @nmouquet.bsky.social · 14d

Very proud of this paper in BioScience :

« Too cute to be wild: what teddy bears reveal about our disconnection from nature »

🧸 are more than toys, they shape among our first emotional connections to nature. But their design may also distort how we see wildlife !

👉 doi.org/10.1093/bios...

🧪🦤

6 25 69

Ben Bolker @bbolker.bsky.social · 16d

github.com/grantmcdermo... is a great addition to this tinyverse/base-R ecosystem. Only thing I miss so far is automatic handling of/spacing adjustments for axis label widths ...

GitHub - grantmcdermott/tinyplot: Lightweight extension of the base R graphics system

Lightweight extension of the base R graphics system - grantmcdermott/tinyplot

github.com

1 4

Ben Bolker @bbolker.bsky.social · 19d

bbolker.github.io/mixedmodels-....

GLMM FAQ

bbolker.github.io

1 7

Ben Bolker @bbolker.bsky.social · 21d

I'm not really sure; in the frequentist world I think this would be called a penalty or regularization term. However, penalization/regularization is usually justified on grounds of reducing bias (e.g. Firth logistic reg) *or* mean squared error (e.g. ridge regression), not avoiding boundaries ...

3

Ben Bolker @bbolker.bsky.social · 25d

is there/will there be a preprint? Very interested in these methods ...

1

Ben Bolker @bbolker.bsky.social · 26d

questionablecontent.net/view.php?com...

Questionable Content

questionablecontent.net

1 2

Ben Bolker @bbolker.bsky.social · 27d

haven't *all* propagated (in particular arm64 versions)

Ben Bolker @bbolker.bsky.social · 27d

MacOS binaries appear to have been built on CRAN cran.r-project.org/web/packages... but haven't propagated to Posit's mirror yet: cran.rstudio.com/web/packages...

ggplot2: Create Elegant Data Visualisations Using the Grammar of Graphics

A system for 'declaratively' creating graphics, based on "The Grammar of Graphics". You provide the data, tell 'ggplot2' how to map variables to aesthetics, what graphical primitives to use, and it ta...

cran.r-project.org

2

Ben Bolker @bbolker.bsky.social · 28d

Is it on the link/linear predictor scale or the response/data scale?

1 1

Reposted by Ben Bolker

Marie Auger-Methe @marieaugermethe.bsky.social · Sep 5

Like to develop spatio-temporal models? Interested in Arctic marine mammals 🦭🐳 and their susceptibility to oil spills? You may be the postdoctoral fellow Dr. Mariane Marcoux and I are looking for! See the details in the job ad: tinyurl.com/spatiostatamm

11 11