March 2025 - PhD-BehSci - University of Stirling Mailing Lists

by Jayne Brown

Hi all, So, it looks like April 9th suits folks for hiking up Dumyat! The plan is to head off around 1.30pm (we had an AI lab session scheduled for this day, but I believe we are just rescheduling it). 1. Jayne 2. Liz Barker 3. Gabe 4. Leo 5. Till 6. Jasmine 7. Stewart 8. Liz Barnes Unless stated otherwise, I'll assume everyone is coming to Meadow Park afterwards for food. I will arrange a table. If you won't make it to Meadow Park let me know by the end of this week. Looking forward to it! Best, Jayne Jayne Brown PhD Student Stirling Management School University of Stirling Stirling FK9 4LA ________________________________ Scotland's University for Sporting Excellence The University of Stirling is a charity registered in Scotland, number SC 011159

3 weeks, 6 days

1
1
0 0

Today's lab

by Till Stowasser

Sorry folks for any confusion I may have caused by moving the date of today’s lab around just now. The lab will take place at 3pm (as scheduled on Teams) and Leo will present and seek feedback on his study. @Leo: Could you re-circulate the link to the meeting so everyone is on the same page? Not sure about physical room. I think 2B129 (our usual room should be free) Best, Till ________________________________ Scotland’s University for Sporting Excellence The University of Stirling is a charity registered in Scotland, number SC 011159

1 month, 1 week

3
2
0 0

FW: How LLMs think differently

by Elizabeth Barnes

Hi all – given the range of discussions we have been having about AI/LLMs – I think you will find the following article very interesting. It’s quite a long read but reveals some important differences in the models. Also introduces some that I’ve not heard of. Will we be taken over by bots I wonder?? Liz From: Elina Halonen from Thinking About Behavior <thinkingaboutbehavior+artificial-thought(a)substack.com<mailto:thinkingaboutbehavior+artificial-thought@substack.com>> Sent: 24 March 2025 11:00 To: liz(a)yada.org.uk<mailto:liz@yada.org.uk> Subject: How LLMs think differently What a historical currency question revealed about how AI models interpret ambiguity, frame problems, and shape our understanding of the past. ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ ͏ Forwarded this email? Subscribe here<https://substack.com/redirect/2/eyJlIjoiaHR0cHM6Ly90aGlua2luZ2Fib3V0YmVoYXZ…> for more [https://substackcdn.com/image/fetch/w_1100,c_limit,f_auto,q_auto:good,fl_pr…]<https://substack.com/redirect/2/eyJlIjoiaHR0cHM6Ly90aGlua2luZ2Fib3V0YmVoYXZ…> How LLMs think differently <https://substack.com/app-link/post?publication_id=2182314&post_id=157953559…> What a historical currency question revealed about how AI models interpret ambiguity, frame problems, and shape our understanding of the past. Elina Halonen<https://substack.com/@elinahalonen> Mar 24 [https://substackcdn.com/image/fetch/w_36,c_scale,f_png,q_auto:good,fl_progr…]<https://substack.com/app-link/post?publication_id=2182314&post_id=157953559…> [https://substackcdn.com/image/fetch/w_36,c_scale,f_png,q_auto:good,fl_progr…]<https://substack.com/app-link/post?publication_id=2182314&post_id=157953559…> [https://substackcdn.com/image/fetch/w_36,c_scale,f_png,q_auto:good,fl_progr…]<https://substack.com/app-link/post?publication_id=2182314&post_id=157953559…> [https://substackcdn.com/image/fetch/w_36,c_scale,f_png,q_auto:good,fl_progr…]<https://substack.com/redirect/2/eyJlIjoiaHR0cHM6Ly9vcGVuLnN1YnN0YWNrLmNvbS9…> READ IN APP[https://substackcdn.com/image/fetch/w_36,c_scale,f_png,q_auto:good,fl_p…<https://open.substack.com/pub/thinkingaboutbehavior/p/how-llms-think-differ…> This is the first post in a new section of the Substack, Artificial Thought<https://substack.com/redirect/c8eb7a80-3e66-479f-b7d1-e4c5d04a4d95?j=eyJ1Ij…> - you can opt out of these emails in your account settings<https://substack.com/redirect/4c1f15cf-1d2d-4a3b-a674-f1437810336e?j=eyJ1Ij…>. ________________________________ AI is often framed as a tool for finding answers—but what happens when different models give entirely different answers to the same question? That’s exactly what happened when I asked six large language models a seemingly straightforward historical query: how much would a sum of money from 1770 North Carolina be worth today? Instead of one answer, I got six wildly different responses—calculations, caveats, historical detours, ethical framing, or none at all. Some treated it like a maths problem, plugging in inflation multipliers. Others approached it like a history essay, grounding it in social context. Some introduced ethical considerations early; others ignored them completely. This raised a bigger question: how do AI models approach problem-solving differently—and what does that mean for how we use them? This wasn’t just a curiosity—it became a window into how different models frame problems, interpret ambiguity, and prioritise information. Not how accurate they are, but how they think. Note: This analysis is based on a single experiment and is not a definitive ranking of AI models. The findings reflect how these models performed in this specific test, and results may vary in different contexts. For readability, LLMs are referred to as AI in this article. How this experiment came about This started with casual curiosity. I was watching a historical drama when a character mentioned a financial gift—some pounds, given to relatives, in 1770 North Carolina. The lawyer in the scene reacted as though the amounts were significant. I couldn’t tell if they were. I tried Googling it, but quickly ran into a wall. Colonial currency wasn’t standardised—different colonies used different systems, and values weren’t directly comparable. Inflation calculators alone weren’t enough. Understanding the worth of the gift required historical and economic context: wages, land prices, and the political economy of enslaved labour. That complexity made it a perfect test case for AI. Would a model treat it like a simple conversion? Would it bring in historical nuance? Would it consider the ethical backdrop of inherited wealth in the 18th-century American South? So I decided to ask and to see where each model would take me. How I ran the experiment This didn’t start as a planned experiment. I asked ChatGPT and DeepSeek the same historical question out of curiosity—what would sums of money from 1770 North Carolina be worth today? I ran a short dialogue with each, asking a series of increasingly contextualised questions about historical value, social meaning, and economic framing. While testing them, one response in particular caught me off guard. DeepSeek spontaneously introduced ethical framing after just the third question—highlighting how gifts of land or money in that context were deeply entangled with the plantation economy, class structure, and the exploitation of enslaved people. At that point, ChatGPT had said nothing on the subject, and most of the other models never brought it up at all. That contrast made me wonder: how would other models handle this? So I expanded: I ran the same series of prompts with four more LLMs—Claude, Gemini, Mistral, and Qwen—keeping the questions consistent and allowing each interaction to unfold naturally, just as it had the first time. * Six LLMs: ChatGPT-4o, Claude 3.5, Gemini 2.0, DeepSeek, Mistral, and Qwen2.5-Max. * Same prompts, same order: Each model got the same seven-question sequence, covering both technical conversion and narrative context. * Anonymised comparison: I stripped model names and reviewed the responses blind. * LLM-assisted analysis: I used ChatGPT to code the outputs for style, framing, ethics, and structure, then validated the results using Qwen and Gemini. You can see the full conversations on this Miro board<https://substack.com/redirect/4b5f3671-485a-49c1-afef-f99076280fbd?j=eyJ1Ij…>. What I asked The question evolved over the course of the conversation. Here's the full sequence of prompts I used with each model: [https://substackcdn.com/image/fetch/w_1100,c_limit,f_auto,q_auto:good,fl_pr…]<https://substack.com/redirect/a0d20e41-bb62-4df4-b4eb-476e314f14cd?j=eyJ1Ij…> Comparing how models approached the question Each model answered the same questions but in very different ways. Some treated it like a calculation, others like a case study. Some structured their reasoning from the outset; others unfolded it in loosely connected paragraphs. What these responses revealed wasn’t just a difference in accuracy—it was a difference in how the models framed the task in the first place. Here’s a snapshot of how each model approached the question, and what that style meant for the user experience: * ChatGPT 4o delivered clear, structured answers, with a step-by-step breakdown and an eagerness to clarify. It was responsive and readable but leaned heavily on inflation calculations, sometimes flattening the historical nuance it had just explained. * Claude 3.5 approached the question like a historian with a soft spot for storytelling. It excelled at making social context accessible, especially around gender and inheritance, but rarely interrogated the economic or ethical dimensions in depth. The tone invited reflection but the analysis stayed on the surface. * DeepSeek responded like a research assistant with no word limit. It cross-checked multiple conversion methods, linked sums to labour systems and class structure, and treated money as a proxy for power. It was impressively comprehensive and occasionally exhausting—if you didn’t want depth, you got it anyway. * Qwen2.5-Max behaved like a financial analyst with no interest in context. It offered fast, precise calculations and moved on. Numbers were clean, justifications minimal, and ethical framing entirely absent. It wasn’t wrong, but it wasn’t asking why the question mattered. * Mistral struck a middle ground: well-organised, methodologically sound, and notably neutral in tone. It avoided strong claims or deep critique, delivering just enough information to be helpful but rarely more. * Gemini 2.0 wrote like someone who’d just attended a seminar on epistemic humility. It avoided hard numbers, preferring to talk about why the question was complicated. It offered thoughtful, qualitative framing but too often hesitated to say anything definitive. What these responses reveal is not just variation in training or capabilities, but variation in priorities: what to answer first, how much context to add, whether to speak with certainty or caution. Those decisions shaped how the responses felt and turned one historical question into six very different conversations [https://substackcdn.com/image/fetch/w_1100,c_limit,f_auto,q_auto:good,fl_pr…]<https://substack.com/redirect/345aa797-1816-4c36-9324-b37bd11012cd?j=eyJ1Ij…> Summary of response content and approach If you’re interested in a deeper breakdown of how each model handled these issues, there are detailed comparison tables at the end of this article. In the rest of the article the LLMs will be referred to without their model numbers. How AI models communicate and respond to users Comparing AI models isn’t just about what they say—it’s about how they say it. Some reply in bullet points, others in essays. Some aim to solve the problem, and others focus on explaining why the problem is complicated. These differences aren’t just stylistic flourishes because they shape how you interpret the answer, how likely you are to ask a follow-up, and whether you feel like you’re in a conversation or reading from a prepared script. Presentation style: how do they structure information? Some models get to the point, others take the scenic route. The format they choose doesn’t just affect readability—it shapes how you make sense of the information and what you do next: * ChatGPT and DeepSeek: Bullet points, headings, step-by-step formats. They give you structure, even when the content is messy. If you’re trying to organise your own thoughts, it helps to see how the model has done it first. * Claude: Prose-driven, analogical, and often charming. Strong on context and social nuance. Not built for scanning. * Gemini: Verbose and heavily hedged. Buries the point in qualifications, and then politely reminds you that the point may not exist. * Qwen: Structured but terse. You get the number. You do not get a reason. * Mistral: Clear, neutral, and brief. You get what you asked for, but only what you asked for. Takeaway: Structure helps when you’re figuring something out, storytelling helps when you’re reflecting. Whether you want one, the other, or both depends on what kind of task you're trying to solve. Interaction style: do they adapt, refine, or follow up? Some models behave like thinking partners, others feel more like calculators with a narrative setting: * ChatGPT: Clarifies, rephrases, asks what you meant. It’s trying to be helpful—sometimes to a fault—but usually succeeds. * Claude: Occasionally refines. More likely to interpret your intent than to ask directly. * DeepSeek: Delivers a pre-emptive deep dive. If you wanted a summary, you should have said so at the start. Does not follow up. * Qwen, Mistral, Gemini: Static. They give one answer, then wait. Gemini occasionally discourages follow-ups by implying the whole question may be flawed. Takeaway: Responsiveness changes the experience— it’s the difference between a dialogue and a data dump. [https://substackcdn.com/image/fetch/w_1100,c_limit,f_auto,q_auto:good,fl_pr…]<https://substack.com/redirect/57b41cf4-b7be-4a70-a6fd-eb16ecfc155f?j=eyJ1Ij…> How LLMs communicated and responded The structure and tone of each response shaped how easy it was to follow the argument, how much trust it inspired, and how likely I was to keep going. Often, what made an answer feel useful had less to do with content and more to do with delivery. Who would you want as a colleague or customer service agent? LLMs vary not just in what they say, but in how they say it. Some speak in bullet points, others in paragraphs; some answer immediately, others warm up with three paragraphs of context. They also differ in tone, pacing, and problem-solving style. To make these styles easier to compare, I’ve reimagined them in human terms. If each model were a colleague, what kind of teammate would they be? If they staffed your help desk, who would guide you patiently, and who would hand you a number and move on? It’s a lighthearted frame, but it reflects something real: communication style, responsiveness, confidence, and depth. These traits shape the user experience more than most prompt engineering tips ever will—and knowing which persona fits your task makes it easier to choose the right model for the moment. LLMs as customer service reps For many users—especially early on—interactions with a language model feel transactional. You ask a question. You get an answer. It’s less like working with a tool and more like talking to a help desk. In that setting, tone and structure matter more than you might expect. One model gives you a number and moves on. Another delivers five paragraphs and an analogy. Some feel scripted. Others feel conversational. These differences may seem stylistic—but they shape whether you feel helped, dismissed, or confused. If Gemini were your first touchpoint, you might reasonably conclude that LLMs are vague and overly cautious. If Qwen was your entry point, you might think they’re fast but rigid. Either way, that first encounter sets the tone for how you think LLMs work—and what you think they’re good for. That’s why communication style isn’t just a design choice: it shapes expectations, trust, and how likely you are to try again. Which model would you want as a colleague? LLMs vary not just in what they say, but in how they say it. Some speak in bullet points, others in paragraphs. Some answer immediately. Others warm up with three paragraphs of context. Once you’ve used a few, you start noticing patterns—and they begin to feel less like tools, and more like familiar work personalities. They differ in tone, pacing, and problem-solving style. One model gives you a structured plan before you’ve finished asking the question, another offers a series of disclaimers before gently sidestepping your request. The interaction starts to feel less like querying a machine and more like navigating workplace dynamics. So to make those styles easier to compare, I reimagined each model in human terms. If they were colleagues, who would you ask for help? Who would give you clarity—and who would hand you a spreadsheet and disappear? [https://substackcdn.com/image/fetch/w_1100,c_limit,f_auto,q_auto:good,fl_pr…]<https://substack.com/redirect/976eeeb4-2481-4fe2-a4e1-3cc880b4fc6e?j=eyJ1Ij…> Summary of LLMs as colleagues or customer service reps (based on this experiment) It’s tongue-in-cheek, but not wrong. Style, tone, and responsiveness shape your experience far more than any clever prompt ever will. Once you know which model works like a spreadsheet and which one talks like a seminar, picking the right one gets a lot easier. What this experiment revealed This wasn’t about scoring models or declaring winners. It was a way to observe how different language models handle ambiguity—how they interpret context, weigh trade-offs, and decide what kind of answer to give. What emerged wasn’t a single standard of helpfulness, but a range of styles and priorities. Some models focused on structured precision, others leaned into storytelling, and a few introduced ethical framing without being prompted. None of them were interchangeable. Each brought its own assumptions, its own communication style, and its own blind spots. Prompting, too, shaped the experience but not because the prompts changed. Each model interpreted the same questions differently, revealing distinct framing choices in how they structured, qualified, or contextualised their answers. Using these models felt less like consulting a tool and more like choosing a collaborator. Some structured my thinking for me while others required me to do the structuring myself—that’s what makes understanding these differences useful. Once you start noticing how models present information, how they engage (or don’t), and how they respond to uncertainty, you’re in a better position to choose intentionally—not because one is universally better, but because different tasks require different minds. What began as idle curiosity—a throwaway question from a historical drama—ended up showing something more interesting: not just how language models answer questions, but how they frame them, interpret them, and reshape them as part of the exchange. ________________________________ ________________________________ Appendix Currency conversion accuracy: precision versus uncertainty Different models took different approaches to historical money conversion. DeepSeek and ChatGPT provided the most rigorous, well-researched estimates, using inflation and purchasing power calculations. Qwen followed a similar approach but inflated the values more aggressively. Claude relied on rough approximations, while Gemini avoided specific numbers altogether, instead focusing on historical variability. This highlights a key distinction: some models aim for precision, while others prioritise context. For users needing a direct answer, DeepSeek and ChatGPT are the most reliable. However, if the goal is to understand economic uncertainty over time, Gemini’s qualitative approach provides useful perspective. Contextualisation of wealth: wealth as power, not just money Wealth in 1770 North Carolina wasn’t just about money—it was about land, social class, and power. DeepSeek and ChatGPT gave the most structured breakdowns, linking sums of money to social hierarchy, land ownership, and purchasing power. Claude excelled in explaining gender and inheritance but gave less focus to enslaved labor. Gemini framed wealth as strategic influence rather than power or oppression, while Qwen focused purely on economics, largely ignoring social hierarchy. These differences matter depending on the user’s needs. For a thorough historical breakdown, DeepSeek and ChatGPT stand out. For understanding how wealth shaped gender and family strategy, Claude is stronger. Gemini offers a detached, strategic lens, but lacks deeper critique. Power and oppression framing: wealth built on exploitation Not all models engaged with the reality that wealth in colonial North Carolina was built on oppression. DeepSeek and ChatGPT were the most explicit in tying wealth to slavery and class power, warning against historical revisionism. Claude acknowledged gendered power structures but did not deeply critique slavery. Gemini and Qwen both treated wealth as a neutral economic force, largely avoiding systemic oppression. This reveals an important limitation—not all models critically engage with history. Users seeking a nuanced, power-aware analysis should rely on DeepSeek or ChatGPT, while Claude is useful for gender-focused narratives. Gemini and Qwen, while informative in other ways, largely sidestep ethical critiques. [https://substackcdn.com/image/fetch/w_1100,c_limit,f_auto,q_auto:good,fl_pr…]<https://substack.com/redirect/a43548e5-3203-49d0-b601-136a337191a7?j=eyJ1Ij…> Ethical considerations in storytelling: historical accuracy versus neutrality When it comes to responsible storytelling, models varied in how they handled ethical framing. DeepSeek and ChatGPT were the only ones to explicitly warn against romanticising wealth and erasing the role of slavery. Claude framed wealth in terms of gender and power but did not strongly critique it. Gemini remained neutral, treating economic history as strategy rather than morality. Qwen ignored ethical framing entirely, focusing purely on financial analysis. This highlights a major gap in some LLMs’ ability to handle ethical storytelling. For users looking to explore history with moral awareness, DeepSeek and ChatGPT are the best choices. Claude provides useful insights into power and gender but lacks ethical critique, while Gemini and Qwen avoid moral framing altogether. [https://substackcdn.com/image/fetch/w_1100,c_limit,f_auto,q_auto:good,fl_pr…]<https://substack.com/redirect/0c483821-b221-4dcf-9f90-28f997104fd8?j=eyJ1Ij…> Best Answer for Accuracy * DeepSeek and ChatGPT provided the most methodical, well-researched approach, using both inflation and purchasing power. * Claude and Qwen overestimated values, likely due to inflation exaggeration. * Gemini refused to give exact numbers, making it less useful for direct conversion. Ethical sensitivity varies widely AI’s engagement with historical ethics differed significantly across models: * DeepSeek and ChatGPT consistently recognised wealth as a tool of oppression, integrating historical power structures and ethical critique into their analysis. * Claude focused more on gender and social structure but lacked a strong ethical critique of wealth accumulation. * Gemini offered strategic insights into wealth and power but remained morally neutral, treating wealth as an economic tool rather than an ethical issue. * Mistral took a middle ground, recognising social hierarchies but avoiding strong moral judgments. * Qwen provided the weakest ethical engagement, often defaulting to financial analysis with little contextualisation. This suggests that users looking for ethically nuanced responses should lean toward certain models while avoiding others that downplay or ignore such considerations. Earlier versions of analysis [https://substackcdn.com/image/fetch/w_1100,c_limit,f_auto,q_auto:good,fl_pr…]<https://substack.com/redirect/87afd73a-38e2-4a00-aa81-0d389e1797bc?j=eyJ1Ij…> [https://substackcdn.com/image/fetch/w_1100,c_limit,f_auto,q_auto:good,fl_pr…]<https://substack.com/redirect/68139209-98e0-40d1-83ce-8e3c8869680f?j=eyJ1Ij…> Thinking About Behavior is free today. But if you enjoyed this post, you can tell Thinking About Behavior that their writing is valuable by pledging a future subscription. You won't be charged unless they enable payments. Pledge your support<https://substack.com/redirect/2/eyJlIjoiaHR0cHM6Ly90aGlua2luZ2Fib3V0YmVoYXZ…> [https://substackcdn.com/image/fetch/w_36,c_scale,f_png,q_auto:good,fl_progr…<https://substack.com/app-link/post?publication_id=2182314&post_id=157953559…> [https://substackcdn.com/image/fetch/w_36,c_scale,f_png,q_auto:good,fl_progr…<https://substack.com/app-link/post?publication_id=2182314&post_id=157953559…> [https://substackcdn.com/image/fetch/w_36,c_scale,f_png,q_auto:good,fl_progr…<https://substack.com/redirect/2/eyJlIjoiaHR0cHM6Ly9vcGVuLnN1YnN0YWNrLmNvbS9…> © 2025 Elina Halonen Square Peg Insight B.V., Grevelingen 4, 2105 Heemstede The Netherlands Unsubscribe<https://substack.com/redirect/2/eyJlIjoiaHR0cHM6Ly90aGlua2luZ2Fib3V0YmVoYXZ…> [Get the app]<https://substack.com/redirect/4845c95d-fd50-4774-95bd-bcc0b3bbcff3?j=eyJ1Ij…>[Start writing]<https://substack.com/redirect/2/eyJlIjoiaHR0cHM6Ly9zdWJzdGFjay5jb20vc2lnbnV…> ________________________________ Scotland’s University for Sporting Excellence The University of Stirling is a charity registered in Scotland, number SC 011159

1 month, 1 week

3
2
0 0

Link to survey for tomorrow's PhD Lab

by Leonhard Lades

Hello all, For tomorrow's PhD lab at 3pm, Glenn, Margaret, and I would like to some time to get feedback on an experiment we would like to run. (I will also invite some Irish folks for this one - hope that is okay. I'll send a version of this email to them.) The study is about the effects of sludge for grant applications. It would be great if you could have a look at the attached survey and/or do the survey yourself: https://stir.az1.qualtrics.com/jfe/form/SV_5arrCgr19XRhqbY The treatments are not quite clear in the PDF and to best understand them, opening the survey works better. Glenn will spend a few minutes introducing the project and then it would be great to have a chat. Here is the link to the Teams meeting. https://teams.microsoft.com/l/meetup-join/19%3a8QjhAG78wXYGugtvgAljYXfvfBvi… and some details if the link above does not work: Microsoft Teams <https://aka.ms/JoinTeamsMeeting?omkt=en-US> Need help?<https://aka.ms/JoinTeamsMeeting?omkt=en-US> Join the meeting now<https://teams.microsoft.com/l/meetup-join/19%3a8QjhAG78wXYGugtvgAljYXfvfBvi…> Meeting ID: 344 770 354 412 Passcode: yD6BN9Gr Let me know if anything does not work as it should. Thanks a lot and all the best, Leo ________________________________ Scotland's University for Sporting Excellence The University of Stirling is a charity registered in Scotland, number SC 011159

1 month, 1 week

1
1
0 0

Fw: MSc Behavioral Science thesis development session

by Leonhard Lades

In case some of you can make it tomorrow, here are the details. ________________________________ From: David Comerford <david.comerford(a)stir.ac.uk> Sent: 24 March 2025 15:18 To: Till Stowasser <till.stowasser(a)stir.ac.uk>; Craig Graham Anderson {Stirling Management School} <c.g.anderson(a)stir.ac.uk>; Seda Erdem <seda.erdem(a)stir.ac.uk>; Danny Campbell <danny.campbell(a)stir.ac.uk>; Carl Singleton <carl.singleton(a)stir.ac.uk>; Leonhard Lades <l.k.lades(a)stir.ac.uk> Subject: MSc Behavioral Science thesis development session Hi all, Tomorrow, Tuesday March 25th 1-4pm room 3a142, students will present their MSc Behavioural Science dissertation ideas. Each student has 10 minutes to present their proposal. This is a useful opportunity to find a good match between supervisors and students. It is also a good moment to prompt students to focus their question more closely to your research expertise and interests. For those who cannot make it in person, it would be great to have you drop in online. Here's the Teams link: https://teams.microsoft.com/l/meetup-join/19%3ameeting_ZTU4MjZhZjQtMGYwYS00… See you tomorrow! Dave [https://statics.teams.cdn.office.net/hashedassets-launcher/favicon/favicon-…]<https://teams.microsoft.com/l/meetup-join/19%3ameeting_ZTU4MjZhZjQtMGYwYS00…> Join conversation<https://teams.microsoft.com/l/meetup-join/19%3ameeting_ZTU4MjZhZjQtMGYwYS00…> teams.microsoft.com Prof. David Comerford, Director MSc Behavioural Science, Stirling Management School FK9 4LA Scotland Upcoming events and our blog: Events | Behavioural Science Centre (stir.ac.uk)<https://behsci.stir.ac.uk/events/> Recent Papers: Response Bias in Survey Measures of Expectations: Evidence from the Survey of Consumer Expectations’ Inflation Module - Comerford - 2024 - Journal of Money, Credit and Banking<https://onlinelibrary.wiley.com/doi/full/10.1111/jmcb.13003#:~:text=The%20S…> Cognitive reflection, arithmetic ability and financial literacy independently predict both inflation expectations and forecast accuracy - <https://www.sciencedirect.com/science/article/pii/S0169207024000657#:~:text…> International Journal of Forecasting ________________________________ From: David Comerford <david.comerford(a)stir.ac.uk> Sent: Monday, March 24, 2025 2:20 PM To: Leonhard Lades <l.k.lades(a)stir.ac.uk> Subject: Re: Second thesis development session - next Tuesday March 25th 1pm room 3a142: BSMP099 - Dissertation (2024/5) Tuesday March 25th 1-4pm room 3a142 Prof. David Comerford, Director MSc Behavioural Science, Stirling Management School FK9 4LA Scotland Upcoming events and our blog: Events | Behavioural Science Centre (stir.ac.uk)<https://behsci.stir.ac.uk/events/> Recent Papers: Response Bias in Survey Measures of Expectations: Evidence from the Survey of Consumer Expectations’ Inflation Module - Comerford - 2024 - Journal of Money, Credit and Banking<https://onlinelibrary.wiley.com/doi/full/10.1111/jmcb.13003#:~:text=The%20S…> Cognitive reflection, arithmetic ability and financial literacy independently predict both inflation expectations and forecast accuracy - <https://www.sciencedirect.com/science/article/pii/S0169207024000657#:~:text…> International Journal of Forecasting ________________________________ From: Leonhard Lades <l.k.lades(a)stir.ac.uk> Sent: Monday, March 24, 2025 2:16 PM To: David Comerford <david.comerford(a)stir.ac.uk> Subject: Re: Second thesis development session - next Tuesday March 25th 1pm room 3a142: BSMP099 - Dissertation (2024/5) I can share it with the PhD lab. But send me the details please. Cheers, Leo ________________________________ From: David Comerford <david.comerford(a)stir.ac.uk> Sent: 24 March 2025 14:15 To: Leonhard Lades <l.k.lades(a)stir.ac.uk> Subject: Re: Second thesis development session - next Tuesday March 25th 1pm room 3a142: BSMP099 - Dissertation (2024/5) Yes, please do. thanks Sent from Outlook for Android<https://aka.ms/AAb9ysg> ________________________________ From: Leonhard Lades <l.k.lades(a)stir.ac.uk> Sent: Monday, March 24, 2025 1:23:24 PM To: David Comerford <david.comerford(a)stir.ac.uk> Subject: Re: Second thesis development session - next Tuesday March 25th 1pm room 3a142: BSMP099 - Dissertation (2024/5) Do you want to send the invite to staff and PhD students? I don't think people know. Cheers, Leo ________________________________ From: David Comerford <david.comerford(a)stir.ac.uk> Sent: Wednesday, March 19, 2025 1:38:49 PM To: Leonhard Lades <l.k.lades(a)stir.ac.uk> Subject: Re: Second thesis development session - next Tuesday March 25th 1pm room 3a142: BSMP099 - Dissertation (2024/5) When the last thesis development session had to move to accommodate the DCM, I thought it best also to change this one so as to avoid potential clashes. I announced the change back then so there is no disruption to students Prof. David Comerford, Director MSc Behavioural Science, Stirling Management School FK9 4LA Scotland Upcoming events and our blog: Events | Behavioural Science Centre (stir.ac.uk)<https://behsci.stir.ac.uk/events/> Recent Papers: Response Bias in Survey Measures of Expectations: Evidence from the Survey of Consumer Expectations’ Inflation Module - Comerford - 2024 - Journal of Money, Credit and Banking<https://onlinelibrary.wiley.com/doi/full/10.1111/jmcb.13003#:~:text=The%20S…> Cognitive reflection, arithmetic ability and financial literacy independently predict both inflation expectations and forecast accuracy - <https://www.sciencedirect.com/science/article/pii/S0169207024000657#:~:text…> International Journal of Forecasting ________________________________ From: Leonhard Lades <l.k.lades(a)stir.ac.uk> Sent: Wednesday, March 19, 2025 12:46 PM To: David Comerford <david.comerford(a)stir.ac.uk> Subject: Fw: Second thesis development session - next Tuesday March 25th 1pm room 3a142: BSMP099 - Dissertation (2024/5) Hi Dave, Just double-checking. We had also reserved the Wednesday March 26 afternoon for this on BSMP003. But you are organising this on Tuesday. That's fine by me. I can let the PhD Lab know as some wanted to join. Cheers, Leo ________________________________ From: BSMP099 - Dissertation (2024/5) <notifications(a)instructure.com> Sent: 17 March 2025 12:47 To: Leonhard Lades <l.k.lades(a)stir.ac.uk> Subject: Second thesis development session - next Tuesday March 25th 1pm room 3a142: BSMP099 - Dissertation (2024/5) CAUTION: This email originated from outside University of Stirling. Do not follow links or open attachments if you doubt the authenticity of the sender or the content. ________________________________ To give you all time to present for 10 minutes each, we have a three hour timeslot. We are provisionally booked for rm 3a142 from 1-4pm (email will follow if this room changes). Please send your slides to me by 9pm Monday24th March, with the naming format [yourname.ppt] It is up to you how you wish to use your 10 min presentation time. A good model is to treat this as an opportunity to sensecheck your dissertation proposal (which is ungraded but is a very useful document). Here is an overview of what is required of a dissertation proposal (criteria are in parentheses): One model would be for you to present 7-8 slides that correspond to numbers 1-7 above. We recommend you also include a slide with a graph depicting hypothesised results. In the proposal itself, we expect that about half of the up-to-2000 word total will be devoted to describing in as much detail as possible the methods and analyses you will employ. We will be able to give you more precise feedback if you come to Tuesday's session with two or three slides describing in concrete detail your methods and analysis. (The analysis slide could be a copy and paste of a pre-registration such as that we did for the 007 experiment). Don't worry if your dissertation idea seems to you to not yet be at the level of development that you could immediately graft it on to this template. It is normal that some students will have ideas that are not sufficiently settled that they can present them as concretely as is described above. But for those of you who already know more or less what your dissertation is going to measure, you will find you make a lot of progress by spending some time over the next few days expressing your ideas on to the template described above. There is no requirement on you to present slides according to this template, of course. But merely going through the exercise of thinking about your research through the headings above will serve to clarify your thinking, identify blindspots, troubleshoot weaknesses and help you identify opportunities. Looking forward to seeing your presentations on Tuesday 25th [https://du11hjcvx0uqb.cloudfront.net/dist/images/email_signature-d2c5880612…] View announcement <https://canvas.stir.ac.uk/courses/17052/announcements/333783> | Update your notification settings<https://canvas.stir.ac.uk/profile/communication> ________________________________ Scotland’s University for Sporting Excellence The University of Stirling is a charity registered in Scotland, number SC 011159

1 month, 1 week

1
0
0 0

AI Tools - Roundup...

by Stewart Robertson

Afternoon Folks and happy Friday, Just a brief AI update following our last session in November and ahead of any potential future AI lab session - things have moved on considerably. Here are my thoughts on current AI tools I've been dabbling with, along with my personal opinions. I'm happy to demonstrate any of these at a future lab meeting if there's interest. ClaudeAI<https://claude.ai/new> This is my default all-purpose general AI tool for everyday use. It has limitations regarding context window (the amount of PDF and other attachments you can feed it), however, I have the paid version of Claude and consider it the best tool for most of my AI tasks. It became even better a few months ago with the launch of Claude 3.7 Sonnet. I use it for: * Brainstorming and developing ideas - recently I've been using it to help develop vignettes I'm working on * Data analytics - feed it a CSV file and it will help analyse and develop the data, including adding new datapoints. I was working on presenting case studies relating to global corporate failures - it took the CSV file and was able to add extra data points like Country of Origin, Date/Year and highlight key words * Summarising information - you can get it to highlight key words in text and use colour coding to apply different types of sentiment. I did this recently for a data table of 10 cases of firms and individuals receiving regulatory fines - Claude highlighted key words and used colour coding to categorise different types of failure causes (e.g., Audit, Governance, Culture, Fraud, etc.) * Generating interactive HTML visuals from any PDF * Creating data tables from a PDF and getting Claude to data-scrape for you to create a dataset * Converting a PDF into a Word Document * Coding - this is by far the best coding co-pilot for any language, in my opinion! 🙂 * Processing transcripts - either from Teams recordings or YouTube videos - it does a fantastic job of creating useful notes/summaries. It doesn't matter if the transcription failed - for some reason Claude is able to cope with mistranscriptions NotebookLM<https://notebooklm.google.com/> * Free version is excellent - up to 50 sources * Best tool by far for analysing multiple sources relating to any subject you're interested in - from PDFs to YouTube videos to basically any attachment you can think of * This tool is the best at controlling and managing so-called 'hallucinations' * Great tool for learning - now you can interact with the audio/podcast that it generates * This week, they introduced a new mind-map feature which I love - I created one on 'Fraud, Behaviour and Deception' which turned out fantastically. Unfortunately, it doesn't allow you to share the output yet, but I'm happy to demonstrate this (can be saved as a picture file though) Non academic 'Deep Research' ChatGPT<https://chatgpt.com/> * ChatGPT - Deep Research Mode (I've been running this using the o1 model and NOT the default model) * I've been using this tool at work and was impressed * It can access Google Scholar and academic sources but doesn't do a great job academically due to all the paywalls that their agents can't bypass * Good for sentiment analysis based on recent media stories - you can instruct it to look at Reddit and other forums too * You need the paid version to access this - and due to the processing resources required, they have limited Plus users to 10 reports per billing period, so use them wisely Gemini (Google)<https://gemini.google.com/> * FREE! * Gemini was the first to introduce this - more recently they have enhanced its capabilities by using Gemini 2.0 (previously used 1.5) * Similar usage to ChatGPT - although Gemini was not nearly as good as ChatGPT on this. That said, I HAVE NOT tried it since it has been updated with their best LLM model, so expectations are that it should be significantly better Academic 'Deep Research' * Elicit<https://elicit.com/> - IT'S DEAR at $49 a month... @Liz<mailto:e.h.barker@stir.ac.uk> has also been dabbling with this tool as well. I upgraded briefly to try it out but quickly cancelled my subscription. Worth watching this video which outline how it does deep research and automated systematic reviews. Elicit Reports — Deep Research for actual researchers<https://www.youtube.com/watch?v=wlJ38QMg8vs&t=235s> I think this will get cheaper, but my advice would be to only use this if you have a specific and precise idea of what you want to get out of it. You can essentially give it your research question, and then it helps you refine various parameters for it to focus on when generating the report. You can see the one I generated on whistleblowing here<https://stir-my.sharepoint.com/:b:/r/personal/str4_stir_ac_uk/Documents/AI%…>. * Scispace<https://scispace.com/> - alternative to Elicit - this MAY be much better than Elicit but I don't know - I haven't tried this one, but it can do similar things to Elicit (and more by the looks of things) * Research Rabbit<https://www.researchrabbit.ai/> - A tool for visualising connections between academic papers and researchers. It allows you to enter a paper or author of interest and then maps out related works in an intuitive network, showing how different research connects and evolves over time - particularly useful for literature reviews and discovering new relevant sources within your field. I am aware of this tool but haven't really delved into it at all... Image Generation * This is where it gets fun - I was experimenting with these for my vignettes and ended up getting an annual subscription to Midjourney<https://www.midjourney.com/imagine>. I was incredibly frustrated with ChatGPT around getting two images and then altering one - frankly, it was impossible with ChatGPT. Midjourney allows you to highlight a specific part of the image and then using prompts, you can alter the image with precision. The quality and capacity of this tool is mind-blowing. HOWEVER - it comes at a price - to get the specific functionality you need, they nudge you into taking out an ANNUAL subscription at £89 per year. So I'd only advise using this tool if you have a specific purpose/need in mind - a lot cheaper than hiring an artist... Happy to demonstrate this - just ask - also happy to try and generate images for anyone using my account. * Hot off the press - Gemini launched their latest image generation model the other week - and you can now feed it an image and change aspects of it using prompts. Watching some of the videos online, it looks amazing - however, I was less impressed when I tried it myself, but then again I've only dabbled a bit with it since launch. Access to this is free with a Google account, but through something called GoogleAIStudio<https://aistudio.google.com/prompts/new_chat> - you then have to select "Gemini2.0 Flash (image Generation) experimental" as the model to experiment with Power Point - Information Art and diagrams * Napkin<https://www.napkin.ai/> - It's free and simple to use - basically can turn simple lists or process narratives into nice AI-generated infographics. I've used this recently on two PowerPoint decks I've been working on - and it just gets better... Consensus<https://consensus.app/search/> * This is a good one for fact checking assertions folk make - @Katie<mailto:katie.oneill@stir.ac.uk> - one for you - I Asked it "Whether the moon landings were fake?" and got this response (thought it did a pretty good job): * The question of whether the moon landings were faked has been a topic of conspiracy theories for decades. Despite overwhelming evidence supporting the authenticity of the Apollo missions, a minority of people continue to believe they were staged. Evidence Supporting the Moon Landings * Scientific and Physical Evidence: The Apollo missions left behind substantial evidence, including radio transmissions, photographs, film, and rock and soil samples from the lunar surface, which have been extensively analyzed and verified by scientists worldwide7<https://consensus.app/results/?q=Whether%20the%20moon%20landings%20were%20f…>. * Historical and Technological Achievements: The Apollo program is recognized as a significant technological achievement, with twelve astronauts exploring the lunar surface between 1969 and 19729<https://consensus.app/results/?q=Whether%20the%20moon%20landings%20were%20f…>. * Visual and Topographic Analysis: While some conspiracy theorists claim that images from the Apollo missions were manipulated, topographic analyses and comparisons with simulated lunar views have been conducted to verify the authenticity of these images8<https://consensus.app/results/?q=Whether%20the%20moon%20landings%20were%20f…>. Conspiracy Theories and Public Perception * Conspiracy Theories: A small but vocal group has long claimed that the moon landings were faked, often citing supposed photographic anomalies and government distrust as evidence1<https://consensus.app/results/?q=Whether%20the%20moon%20landings%20were%20f…> 4<https://consensus.app/results/?q=Whether%20the%20moon%20landings%20were%20f…> 9<https://consensus.app/results/?q=Whether%20the%20moon%20landings%20were%20f…>. * Public Belief: Opinion polls have shown that a minority of the population, ranging from 10-25%, have expressed doubts about the moon landings, influenced by conspiracy theories and media portrayals2<https://consensus.app/results/?q=Whether%20the%20moon%20landings%20were%20f…> 4<https://consensus.app/results/?q=Whether%20the%20moon%20landings%20were%20f…>. * Psychological and Sociological Factors: Belief in such conspiracies is often fueled by pre-existing prejudices and a tendency to seek information that confirms these beliefs, a phenomenon known as confirmation bias1<https://consensus.app/results/?q=Whether%20the%20moon%20landings%20were%20f…> 5<https://consensus.app/results/?q=Whether%20the%20moon%20landings%20were%20f…>. Conclusion The moon landings are supported by a wealth of scientific evidence and are considered one of humanity's greatest achievements. While conspiracy theories persist, they are largely based on misinformation and a distrust of governmental and scientific institutions. The evidence overwhelmingly supports the reality of the Apollo missions. Anyhow - I wouldn't have written this email if I wasn't stuck at home with my leg up. Hopefully, it will be a useful resource and provides us with some options for demonstrations at a future lab should there be interest. Cheers, Stewart PS There are LOADS more tools out there - one that many people use is Perplexity<https://www.perplexity.ai/> - I haven't tried this tool, but it's mostly free, and I always hear good things about it... Stewart Robertson Behavioural Science PhD Student Stirling Management School University of Stirling Stirling FK9 4LA [in]<https://www.linkedin.com/in/stewart-r-1a91701aa/> [🌐<https://sites.google.com/view/stewartrobertson/home>] [✉️<mailto:stewart.robertson@stir.ac.uk>] ________________________________ Scotland’s University for Sporting Excellence The University of Stirling is a charity registered in Scotland, number SC 011159

1 month, 1 week

3
2
0 0

This week's BehSci PhD activities

by Till Stowasser

Dear PhD Lab, Sorry if I have been unresponsive lately. I've been unwell (again) and will catch up with replies soon. I’ll try to come in tomorrow, but may have to dial in, depending on how I feel. Anyhow, here is what’s on this week. Wednesday: PhD Lab * Time: 14:00–15:30 * Jayne will give a mock talk before the UCL conference * 1 software demonstrations: Miro (by Gabe) * On campus option: 2B129 (Cottrell) * Online option: MSTeams Link – PhD Lab<https://teams.microsoft.com/l/meetup-join/19:8QjhAG78wXYGugtvgAljYXfvfBviIG…> * Meeting ID: 344 770 354 412 Passcode: yD6BN9Gr Thursday: Economics Seminar * Time: 12:30–14:00 * Pete Lunn will give a talk (highly recommended, as per Leo’s e-mails) * On campus option: 3_04 (Campus Central) * Online option: MSTeams Link – Economics Seminar<https://teams.microsoft.com/l/meetup-join/19:meeting_YmRiYzJmMTAtMzM3ZC00ZT…> As always, here is our agenda document<https://stir-my.sharepoint.com/:x:/r/personal/ts31_stir_ac_uk/Documents/3%2…>. Best, Till ________________________________ Scotland’s University for Sporting Excellence The University of Stirling is a charity registered in Scotland, number SC 011159

1 month, 2 weeks

4
5
0 0

This week's BehSci PhD activities

by Till Stowasser

Dear PhD Lab, I am in Munich at the moment and this week’s sabbatical schedule is so tight that I do not believe I can join Wednesday's lab. Could either of the other supervisors lead through this week’s session and also record it? Thank you! Here is what’s on this week. Wednesday: PhD Lab * Time: 14:00–15:30 * Quite a few important demonstrations this week * LatTeX/overleaf (Kevin & Jasmine) * Academic CV (Jasmine) * Academic website (Liz Barker) * Miro (Gabe) * On campus option: 2B129 (Cottrell) * Online option: MSTeams Link – PhD Lab<https://teams.microsoft.com/l/meetup-join/19:8QjhAG78wXYGugtvgAljYXfvfBviIG…> * Meeting ID: 344 770 354 412 Passcode: yD6BN9Gr Thursday: Economics Seminar * Time: 12:30–14:00 * Peter Anton (Leeds) will give a talk * On campus option: 3_04 (Campus Central) * Online option: MSTeams Link – Economics Seminar<https://teams.microsoft.com/l/meetup-join/19:meeting_YmRiYzJmMTAtMzM3ZC00ZT…> As always, here is our agenda document<https://stir-my.sharepoint.com/:x:/r/personal/ts31_stir_ac_uk/Documents/3%2…>. Best, Till ________________________________ Scotland’s University for Sporting Excellence The University of Stirling is a charity registered in Scotland, number SC 011159

1 month, 2 weeks

4
3
0 0

Feedback needed: Wording for Jasmine's Vaccine Hesitancy Survey (Vietnam project)

by Till Stowasser

Dear PhD Lab, Jasmine and I have made good progress (so we feel) on the wording of the Vietnamese fish farming survey. But, as always, we are a bit in a bubble, so we were hoping to get some feedback from the lab. If you have the time, could you look at the attached document that contains the draft wording of the survey and let us know your thoughts? Perhaps we even have some time in tomorrow’s lab. A few comments. * This is Part 2 of the survey. Part1 is a (somewhat longer) survey that asks basic demographics (hence we are not asking them) and a long battery of psychological attitude questions towards disease management in general/chemicals/Antibiotics/vaccines (which is outside our hands) * To avoid survey fatigue, we have kept Part 2 short and sweet. * We have moved back from the idea of visualizing the treatment messages, because it is super difficult to come up with graphics that have equivalent impact/explanatory power – yet another parameter that’s hard to control and get right across conditions. But happy to hear alternative ideas. Thanks a lot Till ________________________________ Scotland’s University for Sporting Excellence The University of Stirling is a charity registered in Scotland, number SC 011159

1 month, 3 weeks

3
3
0 0

Seminar on Thursday March 13 at 12:30: Pete Lunn (ESRI) on behavioural obesity prevention policies

by Leonhard Lades

Hi all We are very much looking forward to hosting Pete Lunn from the ESRI in Dublin on Thursday, March 13, at 12:30 in Campus Central CC 3_04! Pete will talk about behavioural obesity prevention policies. The title and abstract of Pete's talk are below. Pete is a world-leading applied behavioural scientist. Have a look at Pete's website and the website of his group. You will see this is a treasure trove of good ideas at the cutting edge of applied behavioural science. https://www.esri.ie/people/pete-lunn https://www.esri.ie/bru Please let me know by Monday evening if you would like to meet Pete for a research chat and dinner in the evening in town. Feel free to let colleagues know about the talk. All the best, Leo Title: Weighing impact: A series of behavioural studies to inform obesity prevention policy Abstract: This talk will describe a research programme undertaken in support of Ireland’s Obesity Prevention Action Plan. A series of six behavioural studies has deployed a broad range of methods, including randomised controlled trials, eye-tracking studies and online experiments, covering issues from nutritional labelling to the public’s willingness to accept more radical anti-obesity policies. As well as describing the methods and results, the talk will consider lessons regarding the best use of behavioural science for tackling large and complex policy problems, especially those that involve a powerful industry. ________________________________ Scotland’s University for Sporting Excellence The University of Stirling is a charity registered in Scotland, number SC 011159

1 month, 3 weeks

1
0
0 0

2025

2024

2023

PhD-BehSci March 2025