Please enable JavaScript.
Coggle requires JavaScript to display documents.
Judgement / Reasoning and Problem Solving - Coggle Diagram
Judgement / Reasoning
and
Problem Solving
Heuristics
Availability
: use
availability
for
frequency
Frequent events more available - but other factors affect memory
-> overestimate rare events: more likely to notice/think -> record well in memory -> easily available
Overweights probability of events that are
recent, vivid, dramatic
e.g. more words with R in 1st or 3rd position
Representativeness
: use
resemblance
for
probability
(of whether event / person in category)
Most categories homogenous - but many aren't
Overweights probability of events that
match expectations
e.g. Steve = trapeze artist, because resembles one
->
gambler's fallacy
: believe our sequence of spins homogenous with category
Single case
-> population (representativeness heuristic)
Assume single case
representative
-> apply to population
e.g. phone is bad, because friends knows a 'man who...'
Simulation
: ease with which we construct hypothetical scenario
e.g. Mario sadder, because he 'just' missed out on ticket
Anchoring / adjustment
: start at anchor, adjust around it
e.g. higher or lower than 20% vs 80%
Judgement
: draw conclusions from evidence
Subjective, contains errors because memory sensitive to errors
To make judgements, use
Frequency estimate
: how often something occurs
Attribute substitution
: use accessible info as proxy - works if well correlated
Heuristics
: mental shortcuts - efficient strategies that usually give right answer
Affect
: use
feelings
for
risk/benefit
Danger = frightening, +ve = good - but emotion influenced by factors other than outcome
Effort
: use
effort
for
value
Takes more work for higher value - but value often independent of effort
Covariation
Illusions of covariation
: detect covariation where it doesn't exist
e.g. star sign vs personality
Because
** Ignore facts
** Consider subset of facts
Base rate neglect
: ignore how frequently something occurs
in general
e.g. 70% recover in 48 hrs - what % if no drug?
e.g. more first-borns in uni, because more first-borns
Use
diagnostic info
instead: info about particular case
** e.g. use descriptive info for lawyer vs engineer, because use resemblance
Covariation
: relationship b/n 2 variables
where presence/magnitude of one can be predicted from presence/magnitude of other
Strong/weak; +ve/-ve
Use to judge
cause & effect
Conjunction fallacy
: bank teller + feminist less likely than bank teller
Confirmation bias
: see other
Thinking types
Dual process model
: 2 distinct means of making judgements
Type 1
:
fast
,
easy
- if under
time
pressure = induction
-> Use often, even if incentivised
-> Use heuristics
Type 2
:
slow
,
effortful
- if
attention
focused = deduction
-> Use if right cues, circumstances - no time pressure, can focus attention
-> Less likely to make heuristic-related errors
Confirmation
Inductive vs deductive
Induction
: draw claims from
specific
evidence
** Bottom up
** Observe -> pattern -> hypothesis -> theory
Deduction
: start with claim, see what follows from premise
** Top down
** Theory -> hypothesis -> observe -> confirm
Confirmation bias
: sensitive to confirming evidence, ignore disconfirming -> protect beliefs from challenge
Seek confirming
Don't use if disconfirming -
belief perseverance
e.g. rate ability according to feedback, even if discredited
Believe confirming,
reinterpret disconfirming
e.g. win = win, loss = near win
Fail to consider alternative hypotheses
e.g. dog scary -> notice more scary dogs -> confirm
Syllogisms
: argument containing 2
premises
& 1
conclusion
(valid/invalid) - deductive
Easier if
concrete
(in words) vs
abstract
(all A are B)
Because of illogical strategies
Belief bias
: if believe conclusion true -> judge syllogism valid (because like conclusion)
Cultural differences
: depends on
formal education
(artificial - must be learnt)
Conditional statements
: if X, then Y
Selection task (Wason four-card task)
:
** If card has vowel, must have even number - turn which card to check
** People perform badly if abstract - A and 7 (<25% right)
** People perform better if concrete - beer & age (70% right)
->
content affects logic accuracy
Decision making
Utility maximisation
: select option with
greatest value
Use goals & values to make decisions
Minimise costs, maximise benefits
Gain/loss framing
Framing: how options/decisions/outcomes described
Gain
(+ve) ->
risk aversion
-> choose sure win
Loss
(-ve) ->
risk seeking
-> choose gamble over sure loss
e.g. options: save lives vs people die
e.g. questions: award / deny custody to parent
Opt-in/opt-out
: more participation if opt-in
Endowment effect
: higher value on things you own
Reason-based choice
: people make choice
if they have a
good reason
to make choice
e.g. award custody -> look at +ve; deny custody -> look at -ve
-> framing effects
Affective forecasting
: predict future emotions
Good
: predicting
direction
Bad
: predicting
strength & length of time
** Underestimate own capacity to adapt
** Underestimate how easily rationalise mistakes
Problem solving
Heuristics
Means-end
: choose option that make current & goal states more alike
Working backwards
: start at goal, what to do to reach it
Analogies
- use more if strategies/instructions/context helps:
Place it in memory
Find it in memory (told to use it)
Focus on
deep structure
(espec. relationships, common elements)
Need to give analogy & tell people to use it
Hill-climbing
: choose option that moves you in direction of goal
But some require backwards steps
Pictures/diagrams
: sometimes helpful to visualise
Experts
- better problem solvers because use
Deep structure: underlying structure, not surface form
Analogies
Subproblems/subgoals
Memory search
Higher-order units (e.g. chunk chess into tactical moves)
Knowledge about domain
Knowledge organised effectively
Routines
Problem solving
: seek path from start to goal
Like search through maze to find shortest path
Process search
: often too many paths -> need to use heuristics
Problem space
: all states one can reach
Creativity
Common elements of creative people
Domain
knowledge
/
skills
Personality
: risk taking, ignore criticism
Internally motivated
by work
Supportive environment
: right place at right time
Stages of creative thought
Preparation
: gather info, little progress
Incubation
: put aside, unconscious working
Illumination
: insight, new idea
Verification
: confirm insight/idea
BUT
Incubation - mixed results
: time away might help
recovery
from fatigue,
forgetting
of unhelpful ideas
Illumination - mixed results
: signals discovery of new path, not solution
3 abilities
Convergent thinking
: connect seemingly distinct ideas
Divergent thinking
: think in novel directions
Forward flow
: thoughts break from past flow
Problems
Problem solving set
: starting beliefs about problem
Benefits
: avoid pointless strategies
Costs
: neglect useful strategies
** e.g. don't see easier way to fill jug, 9 dots
Problem definition
Ill-defined
: goal state too general, how to reach goal not obvious
-> harder to solve
Use
subgoals
, add
structure
(constraints)
Functional fixedness
: rigidity about object's function
e.g. use pliers as weight
-> less likely to solve problem
-> how problem explained affects whether solved
Using Type 1 vs 2
Problem presentation
: use base-rates if presented as frequencies, not probabilities,
e.g. use if 12 out of 1000 > 1.2% or 0.012
Chance
: better judgement if role of chance prominent
(then consider sample size)
e.g. read review cautiously if know chose dish randomly
Education
: better judgement if know about sample size
e.g. smaller hospital should have more 60%+ boys days because of sample size
Thinking habits
: better judgement if take time to think