Thinking and Reasoning Flashcards
1
Q
Heuristics
A
Strategies that ignore part of the information, making judgements and decisions more quickly, frugally, and/or accurately than more complex methods
Heuristics can be effective but can lead to biases
2
Q
Attribution substitution
A
faced with a hard question, you ask an easier one e.g. representativeness, not probability
3
Q
Natural assessment
A
automatic and routine
4
Q
Human rationality
A
- Essential to judge probability/risk
- Learn from evidence
5
Q
Are we rational?
A
- Rational behaviour is defined by rules of probability, economics and logic
- E.g. we are asked to sold a multiplication question, you can either estimate or do long multiplication
- Long multiplication is like rules and programs and algorithms
- Estimation is a shortcut or a rule of thumb or heuristic
- Rules seem more rational than heuristics
Are thinking and reasoning more like rules or heuristics?
6
Q
Judgement
A
Process of judging the likelihood of various events using incomplete information
7
Q
Rules for judgement
A
- Evaluate judgements by accuracy (what actually happened)
- Evaluate judgements by coherence (obey the rules of probability)
8
Q
Base-rate neglect
A
- People tend to ignore the prevalence/base rate/baseline (e.g. 90 out of 100 ppl were salespeople in the quiz)
- Focus on individuating descriptions: stereotypes information
- Proper response: combine prevalence and description
- Follow Baye’s rule
- Example
- Test detects 99% of true cases of a virus, false alarm in 5% of negative cases, 1in 1000 people have the virus
- The probability of virus if you test positive is 2%
- Both lay people and medics find it hard: neglect prevalence
- Crucial to know whether someone is really ill and needs treatment
o Potentially expensive, unpleasant or risky - Medical education specifically teaches this reasoning
- Knowing about causal structure helps: how false positives can arise (e.g. a harmless cyst rather than cancer)
- Motivation helps: not wanting to test positive
- Prespresenting problems differently: natural frequencies
9
Q
Substitution
A
- Probabilities are hard
- Substitute something easier
o Related to probabilities, but more natural and automatic - Representativeness
o Decide whether person/object belongs in a category by whether they resemble a stereotype or prototype
10
Q
The availability bias
A
- ‘Availability’ related to frequency
o More frequent events are easier to remember - Availability is influenced by other factors
o Systematic errors in judgement - E.g. avoiding flying because of vividness not frequency of crashes
- Can be overridden with conscious thought
- Famous names are more available
- But people tend to say non-famous name is more frequent
o Deliberate thought
11
Q
The Conjunction Fallacy
A
- Say there are two events: A and B
- Probability of A occurring (whatever else happens) >= Probability A and B happening together
- Same goes for Probability of B occurring (whatever else happens)
- This is a very common error - cringe
12
Q
Evaluating decisions
A
- Evaluate decisions based on consequences
- Consequences alone are not enough
- Probability comes into play also
13
Q
Expected value
A
- Value of outcome x probability
- Step 1: calculate expected value (EV)
- Step 2: Choose action with highest EV
- E.g. if it lands on heads you win £200, if it lands on tails you lose £100
EV = V(Outcome) x P(Outcome)
14
Q
Expected utility theory
A
- Major theory of modern economics
- For every action, consider each possible outcome
- EU (outcome) = utility (outcome) x p (outcome)
- Choose action with highest utility
- A rational decision maker maximises their utility
- Implies a set of rules all of which people violate
15
Q
Loss aversion
A
- We are more sensitive to loss than gain
- Losses feel worse than gain feels
16
Q
Risk preferences
A
- People treat gains and losses differently
o Whether outcomes are certain or variable - Most are risk averse with gains
o Prefer certainty even if you’ll probably win less that way - Most are risk-seeking with losses
o Prefer a risky loss even if you’ll probably lose more that way
17
Q
Implications
A
- Liking and disliking are subjective
- What if I could make the same event seem like a loss or a gain
18
Q
Tricked by wording
A
- Effect is robust for risky choice framing and attribute framing (e.g. 95% fat free)
- Effects can be reduced:
o With clearer phrasing
o When asked to justify decisions
o When people are experts or highly involved
19
Q
Deductive reasoning
A
- Governed by formal logic
- Set of rules for valid reasoning
20
Q
Inductive reasoning
A
Obs 1. Met train is late
Obs 2-200. Met train is late
Therefore met trains are late
21
Q
To assess argument
A
- Identify the propositions (statements that could be true or false)
- Identify the pattern of reasoning (symbolise the argument)
- Does it match a rule?
22
Q
Modus ponens
A
if P, then Q. P. Therefore Q.
23
Q
Dual process theories
A
- Sometimes we violate rules other times we follow them
- Some stable differences between people
- Not just intelligence
o Performance unrelated to intelligence on many tasks - Thinking fast and thinking slow
24
Q
Dual process theories
A
Type 1
Type 2
25
Q
Type 1
A
intutive
automatic
immediate
easy
heuristic
26
Q
Type 2
A
analylitcal
conscious
slow
demanding
rule-governed
