-
judgement
- evaluation of evidence
- ie how likely is it to rain
-
decision making (choice)
selecting a course of action
-
3 theories on judgement/decision making
- Normative
- Descriptive
- Prescriptive
-
normative (ideal, should)
- norm of how we ought to think in an ideal world to achieve ideal outcomes
- optimisation or consistency
-
descriptive (reality, does)
describes how limitedly rational beings actually or typically think in the real world with less than ideal outcomes
-
prescriptive
- prescribes how limitedly limited beings can or should think in the real world to achieve closer to ideal outcomes
- derived from normative theories
-
Normative Theories of Judgement
- Monty Hall Problem
- Bayes Theorem
- conjunctive & disjunctive events
- expected value and expected utility
-
Probability Notation
- P(A)= probability of A
- P(A&B) = P(A) * P(B) = probability of A and B
- P(A|B) = probability of A, given B
-
Prior Probability
- base rate, baseline estimate
- best estimate before new info
-
Common Errors (Bayes)
- confusion of the inverse (test comes back 90% positive if have cancer vs 90% chance of cancer )
- ignoring false alarm rate (no cancer but positive result)
- ignoring prior probabilities (what percentage of people like THIS have cancer)
-
Conjunctive Events
- compound/"and" events
- everything must be right to ensure a successful outcome (one error ruins everything)
less likely than any of their parts.
-
Disjunctive Events
- compound event/"or" events
- only one piece needs to be right to ensure a successful outcome
at least as likely as any of their parts.
-
how to be more normative
- use prior probabilities more appropriately
- break compound events into single events
-
Value (V)
- objective monetary value
- applies for options that are certain
-
expected value (EV)
monetary value AND likelihood of each outcome
EV = p(heads)(*(v(heads) +(p(tails)(*(v(tails)
-
utility (U)
subjective value, measured in "utils"differs across people and situations
-
expected utility (EU)
subjective value and likelihood of each outcome
-
subjective expected utility (SEU)
subjective value and subjective likelihood
-
evaluating choice options
- V: objective monetary value
- – doesn’t vary across people or situations
- EV: objective monetary value and objective probability
- – varies across situations, but not people
- U: subjective value
- – varies across people and situations
- EU:subjective value and objective probability
- – varies across people and situations
- SEU: subjective value and subjective probability
- – varies across people and situations
- – most idiosyncratic
-
expected value theory
- people should choose the option with the highest EV (EV to predict behaviour)
- keep in mind diminishing the marginal utility
- issue is pricing non-monetary outcomes
-
expected utility theory
- deals with issues of expected value theory
- an explicit set of axioms (assumptions) that underlie rational decision making
**violated axiom = not maximised SEU
-
Axioms of EU Theory
- 1. connectedness (rank alternatives)
- 2. dominance (pick dominant alternative)
- 3. transitivity (a>b and b>c, then you must prefer a>c).
- 4. independence (ignore similar info look at differences)
-
Allias Paradox
mainly to do with phrasing and our tendencies to drift towards loss aversion. Can rephrase 2 options so it's essentially the same thing (e.g. 90 percent to live and 10% to live, same thing)
|
|