Definition
Plain language
Choosing between options when each is described by several cues of differing reliability — like picking a product rated by reviewers you trust to different degrees.
As stated in the literature
A canonical cognitive-science task of choosing between alternatives each characterized by multiple weighted cues; the classic rival strategies are Take-the-Best, Tallying, and Weighted-Additive (WADD), used as the test domain for the AutoCog automated cognitive scientist.
Why it matters: It captures a everyday kind of choice under conflicting evidence, making it a clean testbed for figuring out how people — and models of people — actually decide.
For example, choosing a laptop based on several reviews you trust to different degrees, weighing a glowing review from a source you rely on more heavily than a lukewarm one from a source you barely trust.
Heard on the show
“Please, because "multi-attribute decision-making" is the kind of phrase that makes people close the tab.”Episode 176 — An AI Designed Its Own Psychology Studies, Then Confirmed What It Found