Tim Kaiser
@tkaiser.science
I am a psychologist working at the intersection of psychotherapy research and quantitative methods. Currently working on causal learning and attribution methods at @fupsych.bsky.social - views are my own.
More: https://linktr.ee/tkaiser_science
More: https://linktr.ee/tkaiser_science
Looks like we have similar tastes when it comes to titles
February 3, 2026 at 11:40 AM
Looks like we have similar tastes when it comes to titles
Consider variables z1,...,zN. Suppose that for any collection of N values, the N! different orderings
are equally likely. Then we say that z1,...,zN are exchangeable. For the described algorithm to work the data need to be exchangeable.
are equally likely. Then we say that z1,...,zN are exchangeable. For the described algorithm to work the data need to be exchangeable.
February 3, 2026 at 7:43 AM
Consider variables z1,...,zN. Suppose that for any collection of N values, the N! different orderings
are equally likely. Then we say that z1,...,zN are exchangeable. For the described algorithm to work the data need to be exchangeable.
are equally likely. Then we say that z1,...,zN are exchangeable. For the described algorithm to work the data need to be exchangeable.
Probably all of it, but I always thought of it as a Millenial thing.
January 31, 2026 at 10:44 AM
Probably all of it, but I always thought of it as a Millenial thing.
Viele gute Punkte im Beitrag. Ich frage mich, wann die Nachholeffekte, die da vermutet werden, zurückgehen und die Prävalenz sich stabilisiert. Müsste ja dann ähnlich aussehen wie bei Linkshändern.
January 29, 2026 at 7:09 AM
Viele gute Punkte im Beitrag. Ich frage mich, wann die Nachholeffekte, die da vermutet werden, zurückgehen und die Prävalenz sich stabilisiert. Müsste ja dann ähnlich aussehen wie bei Linkshändern.
Symbolic representation of Ian with his first data set (apple)
January 12, 2026 at 8:25 PM
Symbolic representation of Ian with his first data set (apple)
[52 Folgen in der Download-Warteschlange]
IYKYK
IYKYK
January 12, 2026 at 7:41 PM
[52 Folgen in der Download-Warteschlange]
IYKYK
IYKYK
One is associated with reduced all-cause mortality, one makes you feel warm and fuzzy for a couple of minutes
January 9, 2026 at 10:05 PM
One is associated with reduced all-cause mortality, one makes you feel warm and fuzzy for a couple of minutes
The next piece of oligarch technology will be replaced soon…
@fairphone.com
@fairphone.com
January 6, 2026 at 7:20 PM
The next piece of oligarch technology will be replaced soon…
@fairphone.com
@fairphone.com
The three "waves" of psychotherapy are often described as historical fact, which is astonishing when you consider that they represent the distorted and sometimes straight up incorrect view of a single individual (Hayes).
January 6, 2026 at 1:35 PM
The three "waves" of psychotherapy are often described as historical fact, which is astonishing when you consider that they represent the distorted and sometimes straight up incorrect view of a single individual (Hayes).
Selection bias makes best selling books
January 4, 2026 at 2:10 PM
Selection bias makes best selling books
post a perfect album from the 90s that isn't nirvana, pearl jam, soundgarden, or alice in chains
December 28, 2025 at 10:59 PM
post a perfect album from the 90s that isn't nirvana, pearl jam, soundgarden, or alice in chains
If you want to read something beautiful over the holidays www.stelliform.press/index.php/pr...
Or listen: libro.fm/audiobooks/9...
Or listen: libro.fm/audiobooks/9...
December 16, 2025 at 10:18 AM
If you want to read something beautiful over the holidays www.stelliform.press/index.php/pr...
Or listen: libro.fm/audiobooks/9...
Or listen: libro.fm/audiobooks/9...
Right on time for the upcoming holidays, and an upcoming paper, this arrived 🙂
December 11, 2025 at 9:34 PM
Right on time for the upcoming holidays, and an upcoming paper, this arrived 🙂
You will be visited by three spirits
December 9, 2025 at 11:46 AM
You will be visited by three spirits
This is indeed a good example of an observational study and an appropriate way of dealing with the difficulties of causal inference in such designs.
It might even be the first clinical psych observational outcome study specifying its estimand(?)
It might even be the first clinical psych observational outcome study specifying its estimand(?)
December 8, 2025 at 3:21 PM
This is indeed a good example of an observational study and an appropriate way of dealing with the difficulties of causal inference in such designs.
It might even be the first clinical psych observational outcome study specifying its estimand(?)
It might even be the first clinical psych observational outcome study specifying its estimand(?)
Da hat sich schon wieder eine Satirepartei gegründet
December 6, 2025 at 4:09 PM
Da hat sich schon wieder eine Satirepartei gegründet
Post a famous bathroom scene
November 26, 2025 at 6:29 AM
Post a famous bathroom scene
This is new: the APA site of a certain journal tells me that my institution doesn't have access to it. I then look for other PDF sources and find that, on ResearchGate, the very same paper can be downloaded for free *because my institution is subscribed*.
November 24, 2025 at 1:06 PM
This is new: the APA site of a certain journal tells me that my institution doesn't have access to it. I then look for other PDF sources and find that, on ResearchGate, the very same paper can be downloaded for free *because my institution is subscribed*.
In the photo, the Batman costume also looks extremely discreet. It would hardly be noticeable in Berlin.
November 24, 2025 at 10:46 AM
In the photo, the Batman costume also looks extremely discreet. It would hardly be noticeable in Berlin.
If you think your caffeine tolerance is so high that you only drink it to prevent withdrawal symptoms, this brand will be a completely new experience...
November 24, 2025 at 10:10 AM
If you think your caffeine tolerance is so high that you only drink it to prevent withdrawal symptoms, this brand will be a completely new experience...
Important figure from the post
November 20, 2025 at 8:27 AM
Important figure from the post
We found five structurally ergodic clusters. Two large clusters (about 57% of the sample) and three smaller clusters that were more heterogeneous. We then characterized subgroups using a set of baseline patient characteristics.
November 11, 2025 at 9:11 AM
We found five structurally ergodic clusters. Two large clusters (about 57% of the sample) and three smaller clusters that were more heterogeneous. We then characterized subgroups using a set of baseline patient characteristics.