Doug Parry
@dougaparry.bsky.social
Assistant Prof of Media Psychology in the Department of Communication Science at VU Amsterdam (🇳🇱) researching media effects, digital well-being, attention, mental health. Open Science, Computational Methods
🖥️ https://dougaparry.com
🇿🇦 From South Africa
🖥️ https://dougaparry.com
🇿🇦 From South Africa
For sure! I use 2x quite a lot, but 4x feels really extreme (at least on typical YouTube content)
January 28, 2026 at 2:04 PM
For sure! I use 2x quite a lot, but 4x feels really extreme (at least on typical YouTube content)
Ok, that makes sense... *Industry* being the operant term.
January 19, 2026 at 4:01 PM
Ok, that makes sense... *Industry* being the operant term.
Just to check my understanding: if an author previously collaborated with Company X (e.g., Microsoft) and 2-years later publishes a paper on Platform Y (e.g., Instagram) without disclosing that prior collaboration, this would be coded as an undisclosed disclosable tie under their approach?
January 19, 2026 at 3:53 PM
Just to check my understanding: if an author previously collaborated with Company X (e.g., Microsoft) and 2-years later publishes a paper on Platform Y (e.g., Instagram) without disclosing that prior collaboration, this would be coded as an undisclosed disclosable tie under their approach?
Yes, time spent on platform Z = still no idea what someone is doing or what content they are engaging with.
The holy grail data to actually understand what is going on here..
The holy grail data to actually understand what is going on here..
January 7, 2026 at 3:39 PM
Yes, time spent on platform Z = still no idea what someone is doing or what content they are engaging with.
The holy grail data to actually understand what is going on here..
The holy grail data to actually understand what is going on here..
We're talking correlations here though...
"In each model, we used the social media use variable measured at the same time point as the outcome variable (i.e., from the same evening)."
"In each model, we used the social media use variable measured at the same time point as the outcome variable (i.e., from the same evening)."
January 6, 2026 at 9:28 PM
We're talking correlations here though...
"In each model, we used the social media use variable measured at the same time point as the outcome variable (i.e., from the same evening)."
"In each model, we used the social media use variable measured at the same time point as the outcome variable (i.e., from the same evening)."
💯 This was something that @cjsewall9.bsky.social and I tried here: psycnet.apa.org/doi/10.1037/... (preprint: osf.io/preprints/ps...) to limited (?) success...
APA PsycNet
psycnet.apa.org
December 15, 2025 at 7:35 AM
💯 This was something that @cjsewall9.bsky.social and I tried here: psycnet.apa.org/doi/10.1037/... (preprint: osf.io/preprints/ps...) to limited (?) success...
Definitely!
Though in this instance a 2 second scan of the reference list gives away the game - not a single paper that you'd expect to see in a review/meta-analysis on this topic.
While this is partly a story about AI, I see this case more as a story about fake journals polluting the literature.
Though in this instance a 2 second scan of the reference list gives away the game - not a single paper that you'd expect to see in a review/meta-analysis on this topic.
While this is partly a story about AI, I see this case more as a story about fake journals polluting the literature.
December 9, 2025 at 7:47 AM
Definitely!
Though in this instance a 2 second scan of the reference list gives away the game - not a single paper that you'd expect to see in a review/meta-analysis on this topic.
While this is partly a story about AI, I see this case more as a story about fake journals polluting the literature.
Though in this instance a 2 second scan of the reference list gives away the game - not a single paper that you'd expect to see in a review/meta-analysis on this topic.
While this is partly a story about AI, I see this case more as a story about fake journals polluting the literature.
💯! Definitely. When I was a bachelor's student I would actually race with some of my friends to finish first as we all realised that we did better with our intuitive judgements than when we overthought questions (at least for social science subjects...different story for maths/stats/coding subjects)
December 3, 2025 at 6:09 PM
💯! Definitely. When I was a bachelor's student I would actually race with some of my friends to finish first as we all realised that we did better with our intuitive judgements than when we overthought questions (at least for social science subjects...different story for maths/stats/coding subjects)
Yes, I think time management itself is an important outcome, though for me I see this more at the level of assignments than tests.
bsky.app/profile/dann...
bsky.app/profile/dann...
In other words: I sort of don’t care.
I do worry about students who use accommodations for flexible deadlines on assignments though. I worry that some will struggle in the workplace. College is where you can learn how to deliver on deadline and I hate the idea they’re not getting that practice.
I do worry about students who use accommodations for flexible deadlines on assignments though. I worry that some will struggle in the workplace. College is where you can learn how to deliver on deadline and I hate the idea they’re not getting that practice.
December 3, 2025 at 5:39 PM
Yes, I think time management itself is an important outcome, though for me I see this more at the level of assignments than tests.
bsky.app/profile/dann...
bsky.app/profile/dann...
I create an archive of the page with the Wayback Machine (there's a browser extension) from the Internet Archive and then use that URL when needed. So far seems to have worked
November 25, 2025 at 8:49 AM
I create an archive of the page with the Wayback Machine (there's a browser extension) from the Internet Archive and then use that URL when needed. So far seems to have worked
Ok, interesting - a good start!
Tons of ML work in there, which is to be expected... It would be fairly straightforward to filter this set to 'social science'/non-ML papers
Tons of ML work in there, which is to be expected... It would be fairly straightforward to filter this set to 'social science'/non-ML papers
November 21, 2025 at 2:30 PM
Ok, interesting - a good start!
Tons of ML work in there, which is to be expected... It would be fairly straightforward to filter this set to 'social science'/non-ML papers
Tons of ML work in there, which is to be expected... It would be fairly straightforward to filter this set to 'social science'/non-ML papers
Yes, I'm sure.
But that will be the case for meta-funded research.
It should be possible to identify papers authored by researchers listing Meta/Facebook as their affiliation though?
But that will be the case for meta-funded research.
It should be possible to identify papers authored by researchers listing Meta/Facebook as their affiliation though?
November 21, 2025 at 1:54 PM
Yes, I'm sure.
But that will be the case for meta-funded research.
It should be possible to identify papers authored by researchers listing Meta/Facebook as their affiliation though?
But that will be the case for meta-funded research.
It should be possible to identify papers authored by researchers listing Meta/Facebook as their affiliation though?