«Philip Pärnamets DOCTORAL DISSERTATION by due permission of the Faculty of Humanities, Lund University, Sweden. To be defended at LUX, room C121, ...»
Observing and influencing
preferences in real time
Gaze, morality and dynamic decision-making
by due permission of the Faculty of Humanities, Lund University, Sweden.
To be defended at LUX, room C121, Lund, April 18, 2015, at 13:15.
Lund University Cognitive Science Document name
Department of Philosophy
LUND UNIVERSITYDate of issue 2015-04-18 Author(s) Philip Pärnamets Sponsoring organization Title and subtitle: Observing and manipulating preferences in real time: Gaze, morality and dynamic decision-making Abstract: Preference formation and choice are dynamic cognitive processes arising from interactions between decision-makers and their immediate choice environment. This thesis examines how preferences and decisions are played out in visual attention, captured by eyemovements, as well as in group contexts.
Papers I-II make use of the Choice Blindness paradigm. Paper I compares participants’ eye movements and pupil dilation over the course of a trial when participants detect and fail to detect the false feedback concerning their choices. Results indicate objective markers of detection with important implications for questions concerning possible demand effect or cognitive dissonance explanations of choice blindness. Paper II examines another aspect of the choice environment, namely, the social context. Choice blindness is demonstrated in small groups for the first time. It is shown that preferential change can be induced in dyads by manipulating the group’s beliefs about their choices, thus extending the preference change through choice effects beyond individuals for the first time.
Paper III examines how visual attention differentially supports both decision and memory processes depending on the amount of taskrelevant information available to participants. Participants’ performance and visual attention dynamics was found to vary depending on the amount of information available to them, and the results indicate that decision outcomes are heavily influenced by encoding prior to traditional choice phases used in decision research.
Papers IV-VII concern decision-making in the moral domain. Paper IV investigates visual attention when participants choose between difficult moral dilemmas, showing asymmetries in how participants distribute their attention depending on making utilitarian or deontological choices. Paper V introduces a novel paradigm for influencing decision based manipulating the timing of decision by measuring the direction of gaze while the participant deliberates. Using this method participants’ moral decisions were biased without their knowledge to a randomly chosen alternative. This shows that moral decision and visual attention are highly coupled and that by simply knowing where someone is looking it is possible to influence their decision process. Papers VI&VII build on the links between gaze direction and moral choice and present the first computational decision models based on eye gaze applicable to the moral domain. Paper VI for choices between
principles and Paper VII for donations to charitable organizations.
Together the papers advance novel methodological solutions to understanding preferences and decisions across a number of domains, both highlighting the important contributions of social and sensorimotor interactions to the content of our decisions as they develop over time, as well as demonstrating how decisions can be influenced by leveraging those interactions.
Key words: eye gaze, morality, decision-making, preference formation Classification system and/or index terms (if any)
Philip Pärnamets Copyright Philip Pärnamets 2015. All rights reserved.
Cover art by Josef Mellergård.
Cognitive Science Department of Philosophy Faculties of Humanities and Theology ISBN 978-91-7623-241-5 (print) 978-91-7623-242-2 (pdf) ISSN 1101-8453 Lund University Cognitive Studies 160 Printed in Sweden by Media-Tryck, Lund University Lund 2015 Acknowledgments One of the great privileges of being a PhD student, apart from being allowed to work long, dark hours in one’s heart’s delight to discover something about something probably very few others care about, is all the people you encounter during that time. Not only have I had the joy of meeting fantastically brilliant people from whom I have learnt a great deal, but I have had the chance to make new friends, but also to rediscover my existing ones – so often forgotten in lieu of yet another tantalising deadline or project – and to find my family always there where I momentarily had left them. This thesis is to all of you since without you none of this would have happened anyway.
I have had the joy of having been under the tender care of three brilliant mentors and supervisor.
Lars Hall, you have taught me far more than you imagine, from your striking way with words to always asking me just the question I least wanted to hear at the moment I probably least expected it. You have always pushed me the extra mile in my thinking and for that I am beyond grateful.
Christian Balkenius, you have been a stable haven throughout my years, always providing me a sense of calm. I value your vision and clarity, and our discussions have always made me strive harder to match your insights.
Petter Johansson, I don’t know what to say – without your friendship, support and trust I would not have come anywhere at all. Domo arigato, sensei! I hope to make you proud. (Here I should also thank Marie and Moltas for letting me borrow Petter so many times the past months!) I spent a significant part of 2012 at University College London, which proved to be one of the most stimulating times of my years as a PhD. Most importantly, I had the benefit to get to know and work with Daniel Richardson. I am humbled and thankful for have been allowed this opportunity to work with you Daniel, it has been truly fantastic (PS I still have your lab keys!).
While at UCL, I was lucky enough to work with Jorina von Zimmermann and Ramsey Raafat, who introduced me to the wonderful world of groups. In this context, I also had the joy to benefit from comments and thoughts from Nick Chater, who saved our work many times.
Much of my time at UCL was spent in and around room 201 (with special thanks to Dave Lagnado allowing me space!), where I met a great deal of fantastic people: Tobias Gerstenberg, Christos Bechlivanidis, Neil Bramley, Paula Parpart, Costi Rezlescu, Keith O’Brien, Lukasz Kopec and Adam Harris, as well as, Matt Gobel and Chris Street. Special mention to Marsha Kirichek for letting me to test some wild ideas and be part of your thesis work. Thanks to all!
I have also had the great fortune to get to know Michael Spivey. I have learned much from you and our conversations have fuelled much of my perspective on the human mind.
Much thanks to Ulrike Hahn and Rick Dale, and a wine-fuelled conversation with both of you which helped channel much of my thoughts about the meaning of my main paper.
The bulk of my PhD years have been spent here in Lund at LUCS. Here is my second home. I cannot begin to explain what a wonderful and stimulating environment it is, and what a fantastic workplace; thanks to everyone past and present: Agneta Gulz, Björn Sjödén, Paulina Lindström, Magnus Haake, Åsa Harvard, Richard Andersson, Magnus Johnsson, Mathias Osvath, Helena Osvath, Jens Nirme, Kristin Ingvarsdottir, Zahra Gharaee, Can Kabadayi, Ivo Jacobs, Tomas Persson, Elainie Madsen, Jana Holsanova, Peter Gärdenfors, Betty Tärning, Joel Parthemore and Birger Johansson. Special thanks to Annika Wallin, for always going the extra mile to engage me in exciting new projects and to teach me lessons from the real world outside the lab. Kerstin Gidlöf and Roger Johansson, for a great collaboration. Rasmus Bååth for always being willing to discuss (and help with!) the nerdier sides of statistics.
The department would grind to a halt without some people hard at work keeping it afloat: Agneta Ahlberg, Anna Cagnan Enhörning, Eva Sjöstrand, Anna Östberg, Björn Petersson and Tomas Persson. Many, many thanks for help with all things large and small!
A lot of my data recording has been done in and around the Humanities Lab and using the eye-tracking equipment residing there. There I found an excellent research environment and the ever so excellent eye-tracking seminars. Kenneth Holmqvist deserves as special thanks. I am also much indebted to Marcus Nyström for always being willing to help at critical times, and the constant encouragement you have given me over the years! Likewise, without the advice of Richard Andersson starting as a master student in lab, I would not have come far! And of course all the participants in the seminars over the years for valuable feedback and discussions.
I want to thank Emmanuel Genot and Justine Jacot for being such wonderful creatures. Organising conferences, endlessly redrafting papers and eating canard washed down with bourbon has never been more fun!
With Thomas Strandberg and Andreas Lind I have not only been sharing the same lab, supervisors and frustrations, but also many joys in and outside of work. Thomas, together was have gorged in fantastic films and eaten ourselves halfway to a mortal sin – I can’t imagine a better person to do this with!
Andreas, your constant encouragement, warnings that things are not always what they seem and the fantastic turns of our discussions have been invaluable!
My friends outside of work have given me much support, and living together with many of them in Rymdblomman has been a source of strength over and over again: Jonathan Mellergård, Kaj Svedberg (rabu rabu), Oskar Hallström, Gustav Nilsson, Sofia Nordstrand, Sara Andersson and Victor Rydberg – woop woop! In addition, a special thanks to Oliver Vikbladh for sharing your mind with me, Noomi Weinberg for many fun travels and late nights, Andreas Holmgren for always bringing my ego back to earth, Johanna Landberg for always being there, Josef Mellergård for letting me decompress so many times in your home, and David Ivarsson for all our adventures in space.
I also wish to thank Pekka Mellergård for reminding me about perspective and your friendly questions.
The dancefloor has been my therapy room, spa and funhouse and many nights and early mornings have been spent in attics, warehouses and cellars. To the people behind and in front of the decks – one love.
My family has been fully supportive over the years; despite my sometime inabilities to actual explain what I am doing and why it takes so much time. To my mom and dad especially – thank you for letting me become me.
Finally, the last years have been much brighter since I met Filippa – being with you means the world to me.
Paper I Pärnamets, P., Hall, L., Strandberg, T., Balkenius, C., & Johansson, P.
(submitted). Looking at choice blindness: Evidence from eye-tracking and pupil dilation.
Paper II Pärnamets, P., von Zimmermann, J., Hall, L., Raafat, R., Chater, N., & Johansson, P. (submitted). Choice-induced preference change in groups.
Paper III Pärnamets, P., Johansson, R., Gidlöf, K. & Wallin, A. (submitted). How information availability interacts with visual attention during judgment and decision tasks.
Pärnamets, P., Hall, L., & Johansson, P. (submitted). I see your dilemma:
Visual attention and moral choice.
Paper V Pärnamets, P., Johansson, P., Balkenius, C., Hall, L., Spivey, M.J. & Richardson, D.C. (in press). Biasing moral decisions by exploiting the dynamics of eye gaze. Proceedings of the National Academy of Sciences.
Paper VI Pärnamets, P., Balkenius, C. & Richardson, D. C. (2014). Modelling moral choice as a diffusion process dependent on visual fixations. In Bello, P., Guarini, M., McShane, M. & Scassellati, B. (eds.) Proceedings of the 36th Annual Conference of the Cognitive Science Society. Cognitive Science Society, Austin, TX.
Paper VII Pärnamets, P. (2015) A fixation dependent decision model of charitable choice.
Lund University Cognitive Studies 161.
Scope and summary Imagine walking down a street with a friend, casually chatting as you make your way towards a movie theatre. Just outside the theatre you are stopped by a homeless person asking for spare change. While stopping and considering you look back and forth between the person and your goal, the movie theatre.
Suddenly, your friend tugs your arm and hurries you to move on, less you will miss the beginning of the film. At that moment you are forced to make your decision, to give the homeless person change or not. There might be all sorts of reasons going through your head at that moment for doing one thing or the other. Some of these might have influenced your final decision, or at least you could argue so. However, here I want to ask a different question: could the precise moment when your friend tugged your arm have affected your choice?
And, if so, did the direction you were looking at when you were interrupt have any influence?
The answer to both these questions, as you will discover is ‘Yes’1. More generally, in this thesis several investigations into the interactions between social and sensorimotor processes, on the one hand, and preferences and choices, on the other, are presented. The work grew out of a seemingly innocuous question;
can something (useful) be learnt about moral choices by studying eye movements? What particularly interested me was how eye movements might reflect ongoing deliberation and choice. Over the years, while working on this thesis, the question grew and mutated, seemingly on its own - like an untended lunch box in the fridge - into a broader interest concerning how preferences and choices evolve in general. Eventually I found myself equipped with a dynamic perspective on human cognition and having done a number of thematically and methodologically linked studies.
See Paper V
The example above was chosen not only because it illustrates the main findings of what I consider to be the flagship paper of this thesis (Paper V), but also because it captures the main research themes of the papers and the introductory chapters. These themes can be summed up as dynamic cognition, decisions, and morality.