-
decision making as constrained optimisation
subjective standards constrain decision making and perception of objective standards
-
Coincidence, I think not
- Confirmation Bias, ignore details that don't match
- We ask what is the coincidence of THIS vs. coincidence of A
- reward of recognising real > penalties for seeing imagined
-
gamblers fallacy
it has to come up tails now
-
reverse gamblers fallacy
universe is going to continue to make heads keep coming up
-
identifying associations
ratio of disease to nondisease symptom is the same, not associated
-
post hoc fallacy
- when event B closely follows event A, we think event B caused event A
- ie Vaccine and autism. Time to vaccine and time for screening are close so assume one caused the other.
-
If A & B are correlated....
- without causal evidence: they're related
- with causal evidence: A caused B
- with reverse causal evidence: B caused A
- with evidence for third variable: C causes A and B (they covary)
-
Conclude that A causes B
- A must precede B in time
- rule out reverse causation
- rule out the third variable
-
why assume causation
- some associations are almost perfect
- close in time and space
- easy to test
-
Maximising
- normative-ish model
- finding the best possible outcome regardless of time and energy spent
-
Satisficing
- descriptive ish model
- find the first option that meets a certain criteria
- system 1
-
System 1 (hot)
- uses memory, association, emotion
- narrative thought, vivid concrete, specific
- fast and automatic
-
System 2 (cold)
- uses reasoning and calculation
- propositional thought: logical, formal, abstract
- slow and effortful
-
cognitive heuristic
- mental shortcuts to make decisions
- advantage: reduce task complexity, cost-effective, usually close enough
- disadvantage: unconscious, lead you astray,
-
Anchoring and Adjusting Heuristic
- when estimating unknown quantity, we start at some convenient value then go up or down
- good starting point but we usually fail to adjust sufficiently
- combat via extreme opposite anchor
-
Representativeness Heuristic
evaluate the probability of Event A by the degree of which it is "representative" or resembles category X
*specific event not more likely than general
-
local representativeness
- people expect small sections to be representative of the whole sequence
- smaller sample, less likely to represent the population as a whole
-
Dangers of Representative Heuristic
- Mislead by detail
- Ignores base rate
- Chance is self-correcting
- Stereotypes can be harmful
|
|
