Lecture 7 Flashcards
Law of Large Numbers
As sample increases in size, it becomes more representative of a population
Law of Small Numbers
Belief that small samples are more representative of the population than they really are. Problem: the smaller the sample size, the more it may deviate from the population
Local representativeness
People expect local representativeness - even small sections or samples to be representative of whole sequence or population. But in large population, small samples can deviate considerably.
Representativeness and Bayes’ Theorem
Representativeness heuristic ignores prior probabilities/base rates
Is chance self-correcting?
No
Overconfidence bias
People estimate their accuracy or performance to be higher than it actually is. You are well-calibrated if you predict you will be right x% of the time and that is correct. However, most people are overconfident. Overconfidence decreases as accuracy increases, and is highest when accuracy is around chance levels
Reasons for overconfidence
Attentional and motivational. Attentional: selective information search and encoding, confirmation bias. Motivational: need to appear competent and confident to others and self.
Dunning-Kruger effect
There is underconfidence among the very adept.
Planning fallacy
Tendency to underestimate resources, time, and cost needed to carry out task. Consequence of overconfidence
Mental simulation heuristic
Explains why planning fallacy persists despite evidence. Imagining steps you take to complete project often involves skipping steps and ignoring setbacks.
2 Different Descriptive Theories
For 2 different worlds: riskless and risky.
Decision making in riskless world
In riskless world, compare options along all possible dimensions and pick best option. Normative model: Multi-Attribute Utility Theory (MAUT). Descriptive model: Elimination by Aspects
