Prof Gaëlle Vallée-Tourangeau

So I think we need both. But we can do without Freud ☺️

But doing so, are we losing our ability to see people behind the data? We get quality data we do quality science but how much have we really understood? Latour calls for understanding through inquests. It's not science as we know it but sometimes understanding singular events can change the world.

We are teaching our students to build psychological knowledge based on mouse clicks, key presses, blood flows, words. We say we aim to be objective but as soon as we operationalise we make a subjective decision to "carve" reality up.

Freud is pseudoscience. Latour isn't science it's theory, an invitation to reflect on how scientific facts come into existence. I learnt nothing from Freud but Latour challenged my thinking.

If you read STS I would say you need to read Latour, start with Pandora's hope. It's not a criticism of science (he was often wrongly accused of this), it's an invitation to step outside, and reflect on science as a praxis.

Here we go again. The implication here is that the way AI generates its output is with human-like cognition. Which is untrue. "Cognitive capabilities humans have" is not the same is "outputs comparable to those of humans."
www.theguardian.com/technology/2...

So true! We are not even in our own bubble anymore, I feel stuck in Plato's cave watching the shadows of a sinister movie unfolding, too comfortable, afraid, busy or clueless to take action...

hello @cos.io here's an idea: how about adding a bluesky option so users can showcase their handle on their osf profile?

Great stuff as usual - now on my course reading list 👌

Hello Susan - I would like to be added, thank you ☺️

Same - but interesting that the better control was the LLM alone, no different from Dawes. Paraphrasing the author interviewed: doctors with LLMs disregarded LLM responses that disputed their initial (incorrect) hypothesis.

They were discussing the fact that the third group (LLM alone) outperformed the doctors or the LLM "augmented" doctors. I don't know enough about medical diagnosis to comment but would be curious to see if that replicates in other domains.

What does interacting with our personal data via LLMs mean for the future of human behaviour? Can they become our companions or alter egos? Learn more through this new BBC radio documentary www.bbc.co.uk/sounds/play/... and, if you are inspired, we are editing a call for papers 👇👇👇

jamanetwork.com/journals/jam... @olgakos.bsky.social just heard about this on the hard fork podcast and thought of your work on #diagnosticreasoning

Home - Obsidian Publish Request If you use our template (.QSF) to set up your research, we would appreciate it if you cite our paper when describing your method: Durably reducing conspiracy beliefs through dialogues with AI…

Tom Costello ( @tomcostello.bsky.social ) has made freely available tools that let you build LLM interaction into Qualtrics

publish.obsidian.md/qualtrics-do...

Super pertinent as usual @sdpbht.bsky.social do you have a reference for this? I've seen it done but in a rather ad-hoc way from some of your Penn neighbours arxiv.org/pdf/2309.10771. I am looking for a rigorous approach to prompt for this, including validation prompts.

A screenshot of slides drawn from 7 slide decks, showing the range of topics addressed on the webpage

Calling univ profs & lecturers! If you want to bring concepts from Doughnut Economics into your teaching, we've just launched a webpage bursting with resources for you: 7 slide decks, reading lists, videos & activities - all open access. Dive in & pls reshare! doughnuteconomics.org/university-c...

Workshop outline (with names and affiliations in case that helps you find and cite the experts).

Ian Krajbich and visual examples of online eye-tracking, mouse-tracking, and text one may want to analyze (e.g., online reviews)

Nitisha Desai showing an example of mouse tracking patterns can reveal about decisions and judgments.

Ada Aka explaining how think-aloud protocols and other forms of text process tracing matter: they get us inside the opaque box of people's reasoning! Did you know think-aloud protocols can also be done online? See Byrd et al 2023 below: Byrd, N., Joseph, B., Gongora, G., & Sirota, M. (2023). Tell Us What You Really Think: A Think Aloud Protocol Analysis of the Verbal Cognitive Reflection Test. Journal of Intelligence, 11(4). DOI: 10.3390/jintelligence11040076

My final session at #SJDM 2024: Online #ProcessTracing!

Ian Krajbich gave the rationale.

Xiaozhi Yang demoed webcam #eyeTracking.

Nitisha Desai showed #mouseTracking code.

@adaaka.bsky.social advised on #coding scalable #textAnalysis with #LLMs.

Files on github.com/krajbichlab/...

#python #AI

I meant ask Santa 🎅🏼 and the 🧝 🧝‍♀️ 🧝‍♂️ sorry 😅

Thinking 🤔 about the circular economy as the winter holidays approaches🌲. We need to change our gift giving culture - buying all new can't be a measure of how much we love each other. How about we all shop and ask for gifts via vinted depop Etsy thredup 🌱? #decision-making #behaviourchange ♻️

Jordan Suchow and Vahid Ashrafi's approach

The huge datasets mean graphs were comically oversized, even when text size was minimized. This is the first part of a very (very!) wide plot of mean decision relevancy per cognitive bias.

The graph showing biases' relevance to advice-seeking scenarios is only weakly correlated with the biases' coverage in research.

Jordan Suchow & @vahidash.bsky.social are compiling and analyzing MASSIVE datasets of biases, decision scenarios, and #research #abstracts

- Biases clustered
- Some biases relevant to *most* decisions (e.g. #attention)
- Relevance and #science coverage unrelated?

scholar.google.com/citations?us...

Designing more sustainable products: 18 categories with high potential A new report published by the JRC identifies 18 product groups with significant potential for sustainable design improvements.

European Commission researchers say clothes, shoes, detergents, toys, iron&steel, glass... in total 18 product groups should be prioritised to set binding sustainability criteria in the EU. Particularly durability, recyclability, repairability. 1/3
joint-research-centre.ec.europa.eu/jrc-news-and...

Agreed. Also peer-reviewing is not just a skill, it's a social practice, we should teach students to have a critical reflection on biases and processes www.nature.com/articles/s41...

Great initiative, I'd be happy to be added, thank you ☺️

I'm here for this.

Yes 👍 that's where the fine (philosophical) line is! But if you then report a significant p-value, you are saying that the evidence support the [B>A] hypothesis. That's implicitly harking. If you report the Bayes factor though you are fine. You are following the evidence.

I also would add my point only stands for NHST and theory testing - and agree that Bayes factors make the point moot because you are then making an inductive inference. Ie you are not testing a theoretical prediction, only looking at which hypothesis the evidence favours.

There's a study of human judgement in there. I wonder... is it a case of the inverse fallacy?! Just a hunch, I may be wrong but I feel there is some fallacious reasoning involved (hopefully it isn't mine!)

But if you then conclude you have found evidence for B > A, when your theory was A > B, you are harking, imo.

Happy to be disagreed with 😊 it's more of a meta issue anyway. I'm not arguing for one sided pre-regs by default. I'm just saying that if your theory predicts A > B and you find B > A then you can be pretty sure that A is not > B. 1/2