Linking cognitive biases: The successes of a test case that predicted variations in endowment effect magnitudes
回顾了20年的研究，通过进化生物学视角解释认知偏差的起源，成功预测了禀赋效应在大量新物品上的大小变化，为连接和预测多种认知偏差提供了理论基础。
Abstract Cognitive biases are patterned decisions that deviate from the predictions of rational-choice models. They often have profound and adverse impacts on behavior, making them important to understand. The common approach to them catalogs empirical regularities that exist but lacks a sound theoretical foundation to explain and predict why the biases exist, why they are patterned as they are, and how we can best predict new and deeper patterns. The approach described here can provide such a foundation, link cognitive biases together (as a function of common underlying psychological processes), and increase our predictive power. This article reviews a 20-year arc of research that tested whether looking at cognitive biases through the lens of evolutionary biology—essentially, at how all brains are shaped to have behavioral leanings—can illuminate the origins of biases and thereby enable useful, testable predictions. The test case centers on the “endowment effect” and led to four interconnected discoveries about how, when, and where the endowment effect arises. It culminated in the successful prediction of a very large percentage of the variance in endowment effect magnitudes across a large new set of items—something never before accomplished. The key fact is that natural selection can bias decision-making toward choices that were rational in ancestral conditions but are mismatched to modern environments, yielding outcomes that are irrational yet predictably patterned. More broadly, the success of this test case suggests that many other cognitive biases may be similarly approached and potentially linked by this predictively useful framework.