bias and decision making 1 - rationality and uncertainty Flashcards
define rationality
when decision making appears to be rational
it is a set of norms
norms with reasoning
rules of action or thought which define optimality
rationality as a set of norms - 2 features (which can be errors)
coherence –> needs to be consistent across decisions
correspondence –> needs to correspond to reality, not a mismatch
availability bias (+what error is this)
over-estimating the frequency of rare or memorable events
e.g. plane crashes
this is a correspondence error - mismatch between reality (environment) and response (belief about event frequency)
framing bias (+what error is this)
switching decision based on question framing
e.g. would you buy another ticket if you dropped it on the way to the cinema vs if you dropped £10 on the way to the cinema
coherence error - making a different decision based on the same outcomes - not consistency across decisions
the Linda problem
description of a woman and then were asked whether it is more likely she:
- works at a bank OR
- works at a bank and is an active feminist
rational response:
- option 1 must be more likely than option 2 as 2 is a subset of 1
- option 2 requires two things to be true whereas option 1 only needs one
what the Linda problem shows
most people say option 2 (she works at a bank and is a feminist)
this is in defiance of norms of laws of probability
this is making a conjunction fallacy –> an error in reasoning by assuming that the specific condition is more probable than one more general condition
conjunction fallacy
error in reasoning
assume that specific condition are more probable than a general condition
shown by the Linda problem
decision calculus (3 components)
make rational decisions using:
- logics
- probabilities
- systematic consideration of all options
4 difficulties in making decisions
- the future is uncertain
- need to assess risks and benefits
- need to choose to increase chance of positive outcome
- need to use knowledge to estimate probabilities of future events
rationality - probability based on value
expected value = value an investment will have in the future
good bet = expected value is greater than investment
bad bet = expected value is less than investment
rational choice = invest to maximise the expected value
risk aversion
tendency to accept a sure outcome over a riskier one - even if the outcome results in less gain
two components considered with decision calculus
value and utility
expected utility theory - general idea
decision making method –> calculate the option with the highest expected utility
rational = choose option which maximises our utility
premise:
- theory of decision under risk
- each option leads to one set of outcomes
- probability is known
value vs utility on line graphs against quantity
value = linear increase with quantity
utility = fast increase then curves and more plateaus with greater quantity
with utility - it gets to a point where you’ve had enough and so you don’t value it as highly as you did at the start
marginal utility
utility relative to how much of an item you have - if you have more you value additions to it less
marginal utility + wealth
diminishing marginal utility of wealth:
as amount of money you have increases, each addition to it becomes less important
this is a subjective personal point of view
e.g. giving a billionaire vs a uni student an extra £1000 means different things to them
value vs utility
utility is compressed with respect to value
e.g. you don’t enjoy 10 pizzas 10x as much as 1 pizza, but it does still cost 10x more than 1
value and cost are much more in line whilst utility is more about enjoyment/preference and how it contributes to wellbeing or satisfies desires
value vs utility with coin flip example
option a - coin flips- head win £1000, tails win 0,
option b - no coin flip = £499
expected value:
b = £499
a = £500
–>but most still opt for b
utility attached to first £499 is greater than the additional £501
can measure utility using “willingness to pay”
according to Expected Utility Theory –> people should behave to maximise Expected Utility
uncertainty affects expectation - pizza example
if you would pay £10 for 1 pizza
then how much would you pay for a 10% chance to win a pizza
or a pizza which had a 1/100 chance of making you sick
these add risk into the equation
equation for expected utility
E = p*U
E = expected utility
p = probability
U = utility
e.g. expected utlity of a bet with a 50% chacne of a £40 pay off –> E = 0.5*40 –> E = 20 –> should take the bet if E is higher than investment according to expected utility theory
equation for expected utility with multiple options
E = p1U1 + p2U2 etc…
E = expected utility
p = probability
U = utility
e.g.
a £15 bet with a 20% chance of paying £50 and a 10% chance of paying £100
investment = £15
E = 20
therefore should take the bet as E > investment
the effect of uncertainty
need to understand probability distributions
e.g. if somethings is probable good, there could still be a small chance it goes bad
can picture this as curves on a graph with high and low utility on the x axis
loss aversion and prospect theory
bias in decision making
if there is guaranteed loss –> people opt for option where loss may or may not happen rather than optimal utility
aversion to guaranteed loss
e.g. having £100 and loosing £50 feels worse than having £0 and gaining £50 even though result is the same
e.g. would you accept coin toss when heads = +£1500 and tails = -£1000
innate motive to prefer avoiding losses rather than achieving similar gains
feel the pain of loss more than the pleasure of an equivalent gain
