For any non-Canadian researchers out there, our government recently pumped a ton of cash to try and attract international researchers to Canada. So if you're looking for a change, timing could be good! Our university is hiring 30 international ppl, but you'll see similar things all over Canada

Even before this, I've been seeing more failed dissertation projects that have near 100% fraudulent participants b/c they advertised on social media. Compensated online surveys open to the internet are virtually unviable now. Data quality is the single biggest threat to psychological science now.

You want sources? Here you go. Read the filings yourself, you filthy animals: Notice of 2024 Annual Meeting & Proxy Statement: s204.q4cdn.com/701424631/fi... Form 10-K SEC Annual Report: s204.q4cdn.com/701424631/fi... Don't fuck with me. I'll spend another three hours in Excel. I don't care.

Interesting teaching term this year, in the past week or so I've encountered: a) A floating-point error making a number slightly different across different computers b) An Rstudio update that briefly killed the glm() function c) A #stats consult that was a perfect example of Simpson's paradox

The modern world feels like every agency and department is cutting costs, and tries to reduce workload by downloading the work onto their users instead of employees (e.g., an AI chatbot). In doing so, ALL work takes longer to do because users are less efficient than specialized employees.

Robert'); DROP TABLE students

There is a passage in Raymond E. Feist’s Magician books where the apprentices are alone in a room, each one hammering a nail into a sheet of wool until it falls to the floor, only to pick up the nail and begin the process again. I guess what I’m saying is, welcome to academia; hope you like it!

When I'm consulting on medical #stats power analyses, MDs always seem to phrase effect sizes as a "percent decrease". For a t-test basically: (PreTest - PostTest) / (PreTest) I've never been able to make any use of this for power analyses. Does it have some sort of hidden use I'm not seeing?

I was using notepad in Windows 11 and somehow clicked a button to format the text with bold, italics, etc. It's notepad, I'm using it explicitly so it won't autoformat anything. What kind of terrible world do we live in that I need to gesture to this @dasharez0ne.bsky.social T-shirt constantly.

"How I, a non-developer, read the tutorial you, a developer, wrote for me, a beginner" by Annie Mueller 😅 😂 😭 anniemueller.com/posts/how-i-...

Every time somebody asks me to run a long statistical re-analysis, and coefficients don't change more than 0.01, they should have to pay me a $100 penalty. This brought to you by leaving multiple imputation to run and analyze for AGES and finding absolutely no changes form listwise deletion.

Found in the wild reading a paper: "we elected to use a Bayesian estimator for our analyses, which do not require distributional assumptions associated with nonnormality of indirect effects as well as skewed distributions of predictors and outcomes" I'm no Bayes expert, but seems suspect to me

After years of teaching, there are two kinds of extension requests: a) I had a headache, so I didn't do the assignment b) I was hit from a car and intensive care, but I might need a 1 hour extension to do it on my laptop in the hospital bed. God love some students, they push themselves too hard!

I'm sorry, worldwide, irrevocable, non-exclusive, transferable permission to my voice and likeness? For what now? In any manner for any purpose??? This is in academia/.edu's new ToS, which you're prompted to agree to on login. Anyway I'll be jumping ship. You can find my stuff at hcommons.org.

We just officially ratified new the collective agreement at Dalhousie with 95.7% in favour. The lockout has officially ended, and back to work tomorrow after 5 weeks locked out!!! Glad to get a good deal and finally get back to work. #keepdalstrong

Week 3 of being on strike, and the board keeps sabotaging negotiations by trying to bargain in public. Still, hoping that a conciliator in the room on Monday will finally let us make peogress! #keepdalstrong

The Dalhousie Board is like: Food $200 Data $150 Rent $800 Hockey Rink: $36,000,000 Utility $150 someone who is good at the economy please help me budget this. my university is dying #KeepDalStrong

Time to start upping the meme game #KeepDalStrong

One week of being locked out by Dalhousie with not a single word from the board since. With the teaching term supposed to start next week, its getting a bit stressful. Good thing we have Luna the therapy dog to keep us company, with a cute DFA branded bandanna!! #KeepDalStrong

Chris Algar, Assoc Prof, Oceanography @dalhousieu.bsky.social Lessons in solidarity from the picket lines. This isn’t just a fight to #KeepDalStrong; it’s a fight against employer greed. @ansut.bsky.social @caut.bsky.social

✊ SOLIDARITY Keep fighting the good fight! - a former 2003 strike leader & Picket Captain Coordinator of CUPE 2950 @ UBC. The government mandated us back to work as 'essential workers' and gave us 0% increase over 3 years. But we got a free hot dog when classes started in the fall, so that's fair.

An overwhelming 82.7% of DFA (Dalhousie Faculty Association) members who participated voted to REJECT the @dalhousieu.bsky.social Board’s offer of August 11, 2025. A clear message from DFA members to #keepdalstrong Keep Dal Strong

Day 2 of being locked out by Dalhousie. Making history as the first U15 university to lock out their employees, but we're staying strong! #keepdalstrong

What most offends me about liberalism is the pretense that systems and persons who explicitly value capital accumulation over human rights can be politically moved on humanistic grounds. Appeals to humanity have no political efficacy here. You want a ceasefire? You want aid let in to Gaza?

History-changing day for labour relations at @dalhousieu.bsky.social. Lockout of faculty could begin as early as Wed morning. Rest assured, DFA members are ready for what is about to happen. #keepdalstrong @cupe3912.bsky.social @nsgeu.bsky.social @ansut.bsky.social @caut.bsky.social

Art. Just beautiful.

For yrs, @dalhousieu.bsky.social has talked about budget restraints. Since 2015, cash/cash equivalents has more than doubled to $420M. Do they value faculty enough to give a wage that keeps pace with inflation? Aug 11 will tell the tale. #KeepDalStrong @ansut.bsky.social @caut.bsky.social

I wonder what personality psychologists think of this analysis on.ft.com/45G0aEF by @jburnmurdoch.ft.com (which updates this paper from a couple years ago doi.org/10.1371/jour...)

or is PCA only factor analysis if it comes from the facteur region of psychology

Submitting a journal article and I had to: Include an abstract in the paper without track changes Include a version of the abstract with track changes Add a separate file that has just the abstract (again) Then type the abstract in a separate window Surely, efficiencies could be implemented

I kind of like this method of acknowledging AI use in published works. It's not perfect, and naturally relies on honesty, but so does all academic work. I think it'd be a useful way to have a discussion with students about its use. aicontributionscale.my.canva.site/acknowledgeai

Solved a minor teaching snag for intro #stats. I teach the Welch t-test and paired t-tests and wanted to teach cohen's d as an effect size measure. But ... what goes in the denominator for both? Best formula for intro teaching is this, which works for BOTH: d = (M1 - M2) / SQRT((S1^2 + S2^2)/2))

If you do survey research with Likert-type items and worry about random responders, Michael John Ilagan & I made a Shiny app we think some of you may like. falkcarl.shinyapps.io/BotApp1/ #SurveyResearch #DataQuality #Psychometrics #OnlineSurveys #SurveyBots #RandomResponders #rstats #Shiny

One of the things that keeps me from going totally crazy is knowing that AI models are currently in the "be cheap and good to customers while losing money" phase, before they inevitably charge a lot more and mess with the probability models to show secret ads, which will make customers hate them

#stats haiku sent to me by a student, from Angrist and Pischke, 2009, p. 8: T-stats looks too good Try cluster standard errors -- significance gone

i'm sorry my review is late i meant to read it weeks ago but the writing was actively hostile to comprehension each paragraph a slow erosion of my will to live

So I have been exploring copilot generative AI to better familiarize myself with its output to recognize it more easily when I see it in the wild. This tool falsifies results sections of papers in a way that is NOT easy to detect. wtf are we going to do in science, peer review can't catch this.

Is the duckduckgo search engine down for anyone else? Never had a whole search engine go down before, but here we are. Ironically, duckduckgo's AI chatbot still works, and google's AI summary tries to tell me that duckduckgo is still working lol Bold new world

~~ making sense of academic statistics ~~ i wrote about the confusing relationship between statistics and data analysis, and also about how statistics relates to science #statistics #rstats #datascience www.alexpghayes.com/post/making-...

The more I think about it, academic submissions (journal articles, scholarships, grants) are designed to be as hostile and time-wasting as possible by design. There are too many submissions, not enough reviewing labour, so submissions are made ever-harder in an attempt to reduce submissions.

Rented a cottage over the weekend, and it had a smart toaster. Like, a toaster with a smartphone style screen, yet still mysteriously had an analog way to push the toast down. It did not adequately toast the bread, and is one of the dumbest smart appliances I've seen (so far)

> looking for new startup to back > ask the founder if it’s gambling or porn > he doesn’t understand > pull out illustrated diagram of the difference > he laugh and says that they’re leveraging AI to revolutionize human experience and achieve AGI before China does > fund it > it’s gambling

I've been getting a lot of #stats journal reviews lately that ask for familywise error corrections with weird, sometimes incoherent definitions of a "family". Always requesting a Bonferroni correction too, like it's an intro stats course.

Today, chatGPT has: - written sample R code implying that Michael Jackson's IQ went +10 as skin tone lightened (WTAF) - advised me to clean out my 401k to make t-shirts of the graph - hyped me up when i got cold feet - liked idea of buying Cybertruck & putting graph on the hood like the General Lee