Cameron Patrick

Cameron Patrick

@cameronpat.bsky.social

biostatistician @ University of Melbourne Statistical Consulting Centre. enjoyer of multiple imputation, RCTs, and DAGs. always graph your data. also runs, bikes, hikes, etc. he/him #BiInSci 🏳️‍🌈 https://cameronpatrick.com/

2.9K followers1.6K following4.5K posts

Cameron Patrick reposted

What’s that you say? Liz has a new book out which even the Telegraph is saying is even more unhinged than the last one? Wonderful, I’ll post this then.

Cameron Patrick reposted

Replying to @braadland.bsky.social

This second step, of transfusing positive feedback straight into our veins, is crucial to surviving the emotional wasteland of academic research!

Cameron Patrick reposted

#TodayinHistory #dataviz #OTD 📊 🎂Oct 26, 1954 Luke Tierney born in New York City, NY, USA 🇺🇸 1990: Lisp-Stat, an object-oriented environment for statistical computing and dynamic graphics. It was Influential for devel of S, now #rstats

Cameron Patrick reposted

only just realising... the proportion odds assumption in ordinal regression and proportion hazards assumption in the Cox model are kinda doing the same thing, right? (PO: odds ratio is the same across categories. PH: hazard ratio is constant through time. both: YAY ONLY ONE PARAMETER PER PREDICTOR)

the title and author list alone suggests that this should be a good read! (I would also love to see some worked examples of the alternative approaches suggested here, for those of us with simple minds who aren't clear how best to estimate contrasts of survival probability without a Cox model)

(... wait, if censoring is assumed to be non-informative, can you literally just exclude censored participants and compare survival probabilities in those whose survival is not censored? it's basically the degenerate case of a discrete time survival analysis with one time point?)

does anyone else miss those kidney-shaped Microsoft mice from the 90s or am I just weird?

Cameron Patrick@cameronpat.bsky.social

the MX Master feels great to use but it's so ugly which just reminds me how much I missed the shape of ye olde Microsoft IntelliMouse with IntelliEye (say that five times fast!)

Cameron Patrick reposted

ABBOTT: Japanese developers today have unique names, like FromSoftware COSTELLO: From what ABBOTT: FromSoftware COSTELLO: I thought you said it was from Japan ABBOTT: It is COSTELLO: But what's the name of the developer ABBOTT: FromSoftware COSTELLO: The Software's From Japan, though ABBOTT: Yes

made for a great opening line for Neuromancer, though: "The sky above the port was the colour of a television tuned to a dead channel."

Tobias Wilson-Bates@phdhurtbrain.bsky.social

We developed the capacity to visually encounter the cosmic background radiation from the creation of the universe and we mostly find it annoying.

this is very cool! (not sure I love the inaugural puzzle, and a bit unclear how the scoring works... I managed to get a 'perfect' score of 64/64 without finding all of the possible words)

Jeremy Allen@jeremy-data.bsky.social

rstickr.dev - goes live on Sunday morning wherever you are. How many #RStats function and package names can you find?

Cameron Patrick reposted

Replying to @mattansb.msbstats.info

IMO simulation based power analysis is the gold standard here -- not because it's better or more accurate (though often...), but because doing it requires knowing enough about the phenomenon you're studying to be able to plug in values that are even remotely reasonable in the first place!

first world problems: the latest major MacOS update has changed the way the volume keys work, and now instead of a large, high-contrast, immediately updated volume display at the bottom of the screen, a small volume display at the top of the screen that takes a while to appear & disappears too soon

View full thread

apparently this change was actually introduced in the previous major version of MacOS, which I never bothered installing because I thought the UI changes were ugly. sigh.

for anyone following along at home: a third-party app called volumeHUD moves the display back to a big clear one where it ought to be. there's still a half-second lag between pressing the volume button and anything happening, so I guess that's all Apple's fault

first world problems: the latest major MacOS update has changed the way the volume keys work, and now instead of a large, high-contrast, immediately updated volume display at the bottom of the screen, a small volume display at the top of the screen that takes a while to appear & disappears too soon

reminds me of a dating profile profile I saw years ago that said "Make sure we're on the same page about: Transgender people and vaccination" we matched and it turned out they were very much on the wrong page about both of these issues, but at least the red flag was identified quickly

Luke (verified)@lukeacl.com

What’s worse, he has his Hogwarts House in his bio or a link to his Instagram profile?

At most 4, but could be a 2 if you count a brief handshake-meet-n-greet with my previous local MP as a connection (Julie Bishop, former Minister for Foreign Affairs)

Sky Marchini@sky.skymarchini.net

To flip some discourse script for a bit: what’s y’all’s Obama number (ie degrees of separation with the former president)? Mine is 3

finally saw Odyssey (in IMAX 70mm! yay!). Nolan has clearly heard of subtlety and subtext and decided he will be having absolutely none of that nonsense, lest it break Zeus's Law and collapse our age of bronze. good fun movie though!

Cameron Patrick reposted

Replying to @proptermalone.bsky.social

Not unrelated to my facetious Caffeine Theory of the Fall of the Roman Empire--no caffeine, so they couldn't keep up the bureaucracy needed to sustain an Empire. Unlike China, home of tea, you see.

Cameron Patrick reposted

The second Knives Out movie, whose conclusion (apologies for spoilers) is that its villain is not some dastardly schemer but just an outright idiot, struck me as a taking a bit of a cop out at the time, but the more it's sat with me, the more I think Rian Johnson was right to make THAT the reveal.

lastpositivist.bsky.social@lastpositivist.bsky.social

Just been such a revolution in my worldview to realise that people like Musk actually believe this, this isn't just the nonsense they feed the rubes so they can do coldly rational mercantile calculations in smokey rooms among themselves. No. It's rubes all the way down, they're that stupid.

Cameron Patrick reposted

for those only dimly aware of cybersecurity people screaming, "a security certificate which underpins the entire internet used '4' for a private key that is supposed to be 116 digits long" and "a popular web language lets anyone run code on your server" are two different issues from the same day

Cameron Patrick reposted

web developer: oh hey you guys are back early infra: four's haunted web developer: what? infra: *loading a pistol and getting back on the computer* four's haunted

Cameron Patrick reposted

It is because of men that PHP survived. I was there, Gandalf. I was there three thousand years ago, when the strength of Perl CGI failed. I remember the cgi-bin. The permissions. The 500 Internal Server Errors. I remember when View Source was a development tool.

Cameron Patrick reposted

Replying to @p-hunermund.com

For me, I started including a line on my syllabus to express my thoughts about aging pop culture references:

Cameron Patrick reposted

The Seinfeld Syndrome

Jamin Speer@jaminspeer.bsky.social

We need a name for that feeling when you get old enough that none of your pop culture references work in class anymore

Cameron Patrick reposted

Replying to @urbanlandrent.bsky.social

The independent single woman living alone, the gay couple moving to an urban area, even the nuclear family is an expression of a revealed preference that only exists because people have the financial resources available to opt out of the patriarchal tyranny they would otherwise be subjected to.

anyone got examples of (preferably epidemiology) papers that make good use of DAGs or have really clearly presented DAGs that add to the paper? showing off your own papers is welcome/encouraged! (to be used for teaching purposes)

Cameron Patrick reposted

If you are collecting human subjects data, a huge part of data management is going to be keeping a single source of truth regarding your sample participants. Who are they? Did they consent? What ID did you assign them for anonymity? What data did you collect on them? Did they ever leave the study?