Back
The Psychology of Human Entropy Intuitions
{A variety of conceptualizations of psychological uncertaintyexist. From an information-theoretic perspective, probabilisticuncertainty can be formalized as mathematical entropy. Cog-nitive emotion theories posit that uncertainty appraisals andmotivation to reduce uncertainty are modulated by emotionalstate. Yet little is known about how people evaluate proba-bilistic uncertainty, and about how emotional state modulatespeople\textquoterights evaluations of probabilistic uncertainty and behaviorto reduce probabilistic uncertainty. We tested intuitive entropyevaluations and entropy reduction strategies across four emo-tion conditions in the Entropy Mastermind game. We used theunified Sharma-Mittal space of entropy measures to quantifyparticipants\textquoteright entropy evaluations. Results suggest that manypeople use a heuristic strategy, focusing on the number of pos-sible outcomes, irrespective of the probabilities in the proba-bility distribution. This result is surprising, given that previouswork suggested that people are very sensitive to the maximumprobability when choosing queries on probabilistic classifica-tion tasks. Emotion induction generally increased participants\textquoterightheuristic assessment. The uncertainty associated with emo-tional states also affected game play: participants needed fewerqueries and spent less time on games in high-uncertainty thanin low-uncertainty emotional states. Yet entropy perceptionswere not related to subjectively reported uncertainty, numer-acy or entropy knowledge, suggesting that entropy perceptionsmay form an independent psychological construct.}
@inproceedings{item_3238549, title = {{The Psychology of Human Entropy Intuitions}}, booktitle = {{42nd Annual Meeting of the Cognitive Science Society (CogSci 2020): 5Developing a Mind: Learning in Humans, Animals, and Machines}}, abstract = {{A variety of conceptualizations of psychological uncertaintyexist. From an information-theoretic perspective, probabilisticuncertainty can be formalized as mathematical entropy. Cog-nitive emotion theories posit that uncertainty appraisals andmotivation to reduce uncertainty are modulated by emotionalstate. Yet little is known about how people evaluate proba-bilistic uncertainty, and about how emotional state modulatespeople\textquoterights evaluations of probabilistic uncertainty and behaviorto reduce probabilistic uncertainty. We tested intuitive entropyevaluations and entropy reduction strategies across four emo-tion conditions in the Entropy Mastermind game. We used theunified Sharma-Mittal space of entropy measures to quantifyparticipants\textquoteright entropy evaluations. Results suggest that manypeople use a heuristic strategy, focusing on the number of pos-sible outcomes, irrespective of the probabilities in the proba-bility distribution. This result is surprising, given that previouswork suggested that people are very sensitive to the maximumprobability when choosing queries on probabilistic classifica-tion tasks. Emotion induction generally increased participants\textquoterightheuristic assessment. The uncertainty associated with emo-tional states also affected game play: participants needed fewerqueries and spent less time on games in high-uncertainty thanin low-uncertainty emotional states. Yet entropy perceptionswere not related to subjectively reported uncertainty, numer-acy or entropy knowledge, suggesting that entropy perceptionsmay form an independent psychological construct.}}, pages = {1457--1463}, publisher = {Curran}, address = {Toronto, Canada}, year = {2020}, slug = {item_3238549}, author = {Bertram, L and Schulz, E and Hofer, M and Nelson, JD} }