Toots for tomstafford@mastodon.online account

Written by Tom Stafford on 2025-01-28 at 08:24

Bodywork as cure, cause or coping strategy https://www.astralcodexten.com/p/are-woo-non-responders-defective

"hyperfocus on bodily sensations is a likely risk factor for psychosomatic disease."

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-27 at 07:57

Covid-19 Pandemic as a Natural Experiment: The Case of Home Advantage in Sports

https://journals.sagepub.com/doi/10.1177/09637214241301300

"The COVID-19 pandemic, with its unparalleled disruptions, offers a unique opportunity to isolate causal effects and test previously impossible hypotheses. Here, we examine the home advantage (HA) in sports—a phenomenon in which teams generally perform better in front of their home fans—and how the pandemic-induced absence of fans offered... natural experiment. "

[#]CausalInference

=> View attached media

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-23 at 11:20

Just read an article proposal and felt like the guy in this cartoon

(side note: Does anyone know original artist credit?)

[#]InOurTime

=> View attached media

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-23 at 08:07

People ask me why I am on mastodon and I tell them that I just want to be friends with all the linux users and all the activist librarians of the world. Is that so hard to get??

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-22 at 13:01

Lost in the Middle: How Language Models Use Long Contexts

https://arxiv.org/pdf/2307.03172

You can provide long contexts (prompts) to language models, but the probability that the model uses the provided information varies with the positions of that information. Just like human memory has a primacy and recency bias

=> View attached media

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-22 at 08:57

Lots of new sign-ups to the METASERV email list recently. Don't know where everyone came from, but welcome. Stay tuned for #metascience announcements and discussion

Join: https://jiscmail.ac.uk/METASERV

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-22 at 08:51

We Should Put Fact-Checking Tools In the Core Browser

https://hapgood.us/2018/04/21/we-should-put-fact-checking-tools-in-the-core-browser/

@mikecaulfield post from 2018 arguing that browsers could and should have provenance and context tools built in, to help users navigate the uncertain information seas.

"my model of how this impacts society is not that everyone makes use of these tools, but that the five percent of people who do create a herd immunity that helps protect others from the worst nonsense."

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-22 at 08:15

New preprint!

Comparing Human-Only, AI-Assisted, and AI-Led Teams on Assessing Research Reproducibility in Quantitative Social Science

https://econpapers.repec.org/paper/zbwi4rdps/195.htm

Brodeur et al (including me and 261 others)

Thread on findings on 🦋 https://bsky.app/profile/i4replication.bsky.social/post/3lgccq376fk2y

[#]Preprint #TeamScience #Replication #Reproducibility

=> View attached media

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-22 at 07:57

Trust in scientists and their role in society across 68 countries

https://www.nature.com/articles/s41562-024-02090-5

Trust in scientists is high, globally. Two stand outs:

• trust in the scientific method is predictor of higher trust -> we should communicate, defend and improve the steps to take to produce reliable results

• high levels of belief that scientists should communicate about science with the general public -> need for transparency, including around open access publishing

Data: https://www.nature.com/articles/s41597-024-04100-7

=> View attached media

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-21 at 12:01

Call for evidence: "Disinformation diplomacy: How malign actors are seeking to undermine democracy"

https://committees.parliament.uk/work/8818/disinformation-diplomacy-how-malign-actors-are-seeking-to-undermine-democracy

UK Parliament's Foreign Affairs committee is accepting written submissions until Feb 24

=> View attached media

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-21 at 09:57

TODAY: Cambridge, England

And around for a few days if anyone on the life of the mind - cognitive science - metascience axis is available for chats

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-21 at 08:54

Catch up on part 1 - what we know about Community Notes https://tomstafford.substack.com/p/the-algorithmic-heart-of-community

And part 2 - what fact checkers said about the Meta decision https://tomstafford.substack.com/p/unhinged-unfortunate-unproven-inadequate

[#]FactChecking #CommunityNotes #Meta

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-21 at 08:16

Newsletter!

The Making of Community Notes

https://open.substack.com/pub/tomstafford/p/the-making-of-community-notes

RP#60 A third (and final?) comment on the Community Notes system recently picked up by Meta, following a great article about the original team which designed it.

[#]FactChecking

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-20 at 17:29

Pre-announcement: UKRI Metascience AI early career fellowships

https://www.ukri.org/opportunity/ukri-metascience-ai-early-career-fellowships/

"The UK Metascience Unit will fund a cohort of early career fellows to build our understanding of:

• how the growing adoption of AI is changing the research landscape, including what metascientific, ethical, epistemic, and socioeconomic questions this raises

• how governments, industry, and funding organisations should respond"

Closes 10 April,

[#]MetaScience #MetaResearch #Funding

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-19 at 10:16

Made me realise that our paper on Drift Diffusion models could have been pitched more directly as demonstrating the parameter recovery power of the DDM

Quantifying the benefits of using decision models with response time and accuracy data

https://link.springer.com/article/10.3758/s13428-020-01372-w

=> View attached media

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-19 at 10:12

BJKS Podcast

Robert Wilson: 10 simple rules for computational modelling, phishing, and reproducibility

https://bjks.buzzsprout.com/1390924/episodes/16120296-108-robert-wilson-10-simple-rules-for-computational-modelling-phishing-and-reproducibility

Wide ranging discussion based on this paper

Ten simple rules for the computational modeling of behavioral data. https://elifesciences.org/articles/49547 (Wilson & Collins, 2019)

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-18 at 09:22

Podcast on algorithmic price fixing and price gouging:

Lately

https://pca.st/episode/9257e731-5330-46b4-bdc9-118f112904eb

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-17 at 08:18

"Using ChatGPT is not bad for the environment"

https://andymasley.substack.com/p/individual-ai-use-is-not-bad-for

Andy Masley claims water and energy costs of chatGPT queries are irrelevant compared to other typical consumption choices

=> View attached media

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-17 at 08:05

The Making of Community Notes

The team that built X’s Community Notes talks about their design process and the philosophy behind their approach to combatting false information on the platform.

https://asteriskmag.com/issues/08/the-making-of-community-notes

interview from Nov 2024 with Jay Baxter Keith Coleman Lucas Neumann Emily Thai reveals the Community Notes design process to be thoughtful and seriously engaged with trust and quality issues. Essential a pre-rebuttal to many recent criticisms of Community Notes

[#]FactChecking

=> More informations about this toot | View the thread

Written by Tom Stafford on 2025-01-17 at 06:44

The Effects of Diagnosing a Young Adult with a Mental Illness: Evidence from Randomly Assigned Doctors

https://gwern.net/doc/psychiatry/2023-bos.pdf

via ACX, who notes that conscript sample is is very different from that of people who voluntarily request a diagnosis

Good causal inference doesn't get you away from possible context effects. Comment here notes that observed effects could be explained by military discrimination against those with diagnosis https://www.reddit.com/r/slatestarcodex/comments/1htp5qh/the_effects_of_diagnosing_a_young_adult_with_a/

=> More informations about this toot | View the thread

=> This profile with reblog | Go to tomstafford@mastodon.online account