Collaborations with a psychology research lab are under consideration.
Abstract
People often make irrational choices in their everyday lives because relevant information is not made available or is not presented clearly enough, and also because the human mind is naturally subject to cognitive biases. The purpose of this research will be to design and test easy-to-interpret visualizations that help people making informed decisions and assist them in probabilistic judgment and reasoning.
Context
People have trouble making rational decisions, especially when they have to estimate the probability of real life events. A person who sees media reports of plane crashes might tend to overestimate the risk of taking an airplane (availability heuristic). Or a medical doctor might diagnose a disease too easily, without taking into account its rarity in the population (base rate fallacy). Such cognitive biases are numerous and well-documented in psychology.
Furthermore, there are many cases where information relevant to decision making is simply not available or not presented clearly enough. For example, one might stop eating some food after reading in a magazine that it has been linked to heart disease, where in fact the correlation between the two has been found to be extremely low.
Despite some preliminary work (e.g., visualizing uncertainty), little research has been carried out in information visualization and probabilistic judgment.
Objectives
The purpose of this PhD thesis will be to design, implement and evaluate visualization techniques that improve the probabilistic judgment of laymen. Possible topics of investigation include:
1) Easy-to-interpret visualizations of abstract concepts such as conditional probabilities, correlations, distributions and logical reasoning;
2) Empirical calibration of probabilistic visualizations to compensate for cognitive biases;
3) Studying the effect of visualization techniques on actual decision making (e.g., the choice of a treatment given benefits and risks);
4) Augmenting news articles with visualizations to show, e.g., the likelihood that the reported event or a similar event will affect the reader or his/her significant others.
5) Augmenting popular science articles by showing the "strength" of a discovery (e.g., magnitude of a correlation or degree of confidence in the findings), in order for readers to correctly interpret the findings and take informed decisions.
Work program
The PhD student could first get acquainted with the field of information visualization and with well-known cognitive biases (to begin with, see the list in Wikipedia), then pick a small research problem that seems novel, interesting and feasible (e.g., design a probabilistic visualization that corrects for base rate fallacies). A further step will then consist in validating and generalizing the approach.
Extra information
- David Spiegelhalter's website http://understandinguncertainty.org/ discusses how to present statistical data to the public and contains examples of animated visualizations for the layman.
- Scott Plous' "The Psychology of Judgment and Decision Making" is an excellent introduction to cognitive biases.
Prerequisite
The candidate must hold a masters degree in computer science, be creative, have a good experience in software programming (including graphics programming), and be genuinely interested in both information visualization and experimental psychology.
Détails
Expected funding
Institutional funding
Status of funding
Expected
Candidates
Utilisateur
Créé
Mardi 16 mars 2010 17:09:27 CET
dernière modif.
Vendredi 20 avril 2012 15:51:58 CEST
Fichiers joints
filename
créé
hits
filesize
Aucun fichier joint à cette fiche
Connexion
Ecole Doctorale Informatique Paris-Sud
Directrice
Nicole Bidoit Assistante
Stéphanie Druetta Conseiller aux thèses
Dominique Gouyou-Beauchamps
ED 427 - Université Paris-Sud
UFR Sciences Orsay
Bat 650 - aile nord - 417
Tel : 01 69 15 63 19
Fax : 01 69 15 63 87
courriel: ed-info à lri.fr