Matthew Russell
matteomics.bsky.social
Matthew Russell
@matteomics.bsky.social
Measurements want to be accurate;
Experiments want to be elegant;
Data wants to be beautiful and Data wants to be free
#proteomics
#rstats

Take n pairs of twins; separate them from their parents and each other at birth; randomly allocate one twin to a bilingual couple the other to a monolingual couple; follow up over 90 years. Costs include bribes for ethics comittee.
November 11, 2025 at 2:25 PM
I mean, there may be other issues with the paper! Psyc research being what it is. It's just the problem wasn't power.
November 6, 2025 at 1:39 PM
Also, doesn't kin selection really only apply (in the sense of increasing population frequency over time) to the actual genes that confer kin selecting altruism behaviours. All other genes are just along for the ride.
November 5, 2025 at 1:59 PM
Not power though. The original paper made a type-I error, false rejection of null hypothesis. Power is sensitivity of experiment to correctly reject null hypothesis, or avoidance of type-2 error. Type-I errors will occure 5% of the time at p-value 0.05. Independed replication provides protection.
November 5, 2025 at 11:01 AM
Since yoy mention excel...
For ggplot2 graphs you can get data out with ggplot_build(), and then output the actual data for the plot to excel with openxlsx.
ggplot2.tidyverse.org/reference/gg...
cran.r-project.org/web/packages...
Build ggplot for rendering. — ggplot_build
build_ggplot() takes the plot object, and performs all steps necessary to produce an object that can be rendered. This function outputs two pieces: a list of data frames (one for each layer), and a p...
ggplot2.tidyverse.org
November 4, 2025 at 9:23 PM
Arguably both trends are driven by some weird stuff going on 1985-95. That's when computers became available, but researchers maybe didn't know how to use them yet, data entry was manual and error prone and software wasn't well developed.
November 4, 2025 at 3:57 PM
A premium product in which you get the right answer slowly, rather than the wrong answer quickly.
October 30, 2025 at 4:12 PM
So far so good. Original file starts all sentences on a newline, setting up original qmd and docx derived md helps alignment work.
Also, working in sections makes the process feel more managable.
October 30, 2025 at 4:07 PM
I think there may be a potentially very lucrative gap for for a "how to actually learn stuff" influencer. Somehow hook into the manosphere - "real men actually learn stuff" and $$$.
October 29, 2025 at 11:55 AM
I mean zoologically I think a panther is an all black leopard or jaguar. But maybe curators of ancient pots use "panther" to mean this cat figure however it appears. I'm just intersted in what the ancient artist intended to convey vs how the modern accadmic describes its.
October 24, 2025 at 1:38 PM
It would depend on the social occasion.
October 24, 2025 at 10:35 AM
Genuin question. The cat looks more like a cheetah with it's small head and grey hound like hind legs. Did the people who made these pots distinguish big cat species the way we do?
October 24, 2025 at 10:31 AM
There's always a Douglas Adams quote.
In this case:
“We demand rigidly defined areas of doubt and uncertainty!”
October 23, 2025 at 9:22 PM
I'm sure I remember the British National Party and Nick Griffen advocating for this kind of policy 20 years ago.
October 22, 2025 at 1:07 PM
Love it!
The AI doesn't pick out the dark humor of the Turkey celebrating a festival at which it will be dinner, referencing "turkeys voting for christmas" though.
October 22, 2025 at 1:04 PM
These images are great.
If you'll take shredded as "muscular and lean" how about Myosin II?
October 13, 2025 at 1:06 PM