judgement and decision making

decision making in everyday life

• lots of decisions that we make

• some have social factors to also consider

• others are big decisions eg what career to take

• other involve intertemporal trade off - reward of certain options are distributed across timelines. Eg switch off alarm clock, gives you extra rest, whereas going to uni gives you more long term benefits

questions about how people make decisions

• are they the same across scenarios or for big/small decisions

• do we take into account similar factors when deciding

old views on decision making

• aristotle considered logic

• 17th century - mathematicians looked at probability eg bernoulli

• beleived decision making involved computation

• ompute the values of each option, evaluate the options and decide how much value we want to assign to each of the options. Whichever option has the highest value will be chosen 

• have to decide which equations to use to compute the values of the options

weaknesses of mahematical approach

• may have unknown info - how do you assign value to something you've never experienced before

• limited cog capacity -  can only hold so much in memory etc, can only perform so many computations. Also computations are unrealistic for us to do every time we make a decision 

• changing situations - eg when hungry just buy whatever. so if changing factors influence decisions, hard to come up with an equation that includes all the options

• so unrealistic in many situations

shift

• realised psych factors also at play

• Looked at by cog and social psychologists eg amos tversky and daniel kahneman (thought of as first to look at it - 1970s)

heuristics

• use cog shortcuts to make decisions

• Can be applied very easily, with limited information, they’re very adaptive to different scenarios, intuitive, effortless

• “In general, these heuristics (cognitive shortcuts) are quite useful, but sometimes they lead to severe and systematic errors.” Tversky and Kahneman, 1974

analytical vs intuitive

• Intuitive answer = black circle on left is bigger however they are equal size. 

• Cognitive shortcuts/heuristics are analogous from intuitive response you get from your visual stimulus 

• Hard to use analytical response to override what has been perceived intuitively 

• Shows how the analytical and intuitive systems work together to solve problems, in this case they lead to dif responses, but in most cases they lead to very similar  

• So in most cases intuitive is correct, but sometimes they diverge 

examples of heuristics and biases

• Anchoring bias

• Framing effect

• Representativeness heuristic 

• Availability heuristic

russo and shoemaker anchoring bias

• asked to think of last 3 digits of phone number

• add 400 to that number

• think of the resulting number as a year

• ask if battle was before or after

• Number calculated was close to when they said Attila was defeated - acted as an anchor 

  • Anchoring bias - number hold in mind becomes an answer, then start to adjust from the anchor. 

  
  
  

anchoring bias

• When estimating uncertain quantities - start from an anchoring number, gradually adjust the estimate by mentally moving from the anchor - often end prematurely, so dif people give dif answers

anchoring bias and judges

• Englich, Mussweiler, and Strack (2006)

  • 42 accomplished judges, prosecutors with > 10y experience, asked to read legal case, give estimation of sentence length

  • Asked to respond to question:

  High anchor:

  • “The sentence for the defendant will be higher or lower than 3 years?”

  • On average gave 33 months

  Low anchor:

  • “The sentence for the defendant will be higher or lower than 1 years?”

  • On average gave 25 months

  • Tells us its hard to overcome anchoring bias

framing effect and meat

• People gave dif responses when gave same q presented in dif ways

• 75% lean vs 25% fat

framing effect

• The wording of a description influences people’s decisions even in equivalent situations.

framing effect and custody

Which parent would you award custody?

• Most gave custody to B

• Who would you deny custody?

• Would deny B custody (give A)

• One explanation for this is based on attention - when ask to choose what they prefer, they look at factors that are positive (B has more), when ask who to deny, look at negative factors (B also has more)

losses vs gains for money

• Begin the game with £0

  • Gamble A: gain £20 for sure, money to gain: risk averse

  • Gamble B: 1/3 chance to gain £60; 2/3 chance to gain nothing

  Begin the game with £60

  • Gamble A: lose £40 for sure, money to lose: risk seeking

  • Gamble B: 1/3 chance to lose nothing; 2/3 chance to lose all £60

  • Game is the exact same, just worded differently 

  Result 

  • First one -  more people choose gamble A

  • Second one - more people chose gamble B

why does this effect happen

• When put in a game framing condition, people start to get more risk averse, want to make sure they can gain some money 

• When put into a lose scenario, want a chance to not lose money

coin toss and representativeness heuristic

• Person tosses fair coin six times in a row 

• People say HTHHTT is more likely than HHHHHH, even when they have the same probability 

• A series of outcomes with an approximate equal number of heads and tails is more representative

• we’re making it (the shuffle) less random to make it feel more random (steve jobs)

representativeness heuristics

• Judgement is based on how much the target looks like a typical/rep example of the class under consideration, does it look like a random sequence etc

person example and representativeness heuristic

• Linda is thirty-one years old, single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply concerned with issues of discrimination and social justice.

• Which alternative is more probable?

• A: Linda is a bank teller.

• B: Linda is a bank teller and is active in the feminist movement.

• Most people choose B

• Implicit second event (Linda is active in the feminist movement)

• For B to be true, A has to be true 

• So correct answer is actually A

what does this suggest

• people don't often rely on logic, instead use representative heuristics/other intuitive things/heuristics 

chips and walnuts rep heuristic

• Do chips or walnuts have more calories per hundred grams? Correct answer is nuts, but most people say french fries 

• French fries rep of what high calorie food looks like 

• Zou and Bhatia ask pps to list 10 words that come to mind when thinking of high or low calorie foods

• Calorie estimation is better predicted by word associations with the food rather than by the food’s actual calorie content

• If want to predict peoples judgement regarding whether french fries or walnuts have more calories, look at word associations 

letters which have k in availability example

how many english words have k as the first letter vs the 4th

• 3551 vs 5871

• But most people say first letter is more

• Easier to find cases where k is the first letter, harder for fourth 

• Example of availability heuristic

Availability heuristic

• Judgement is based on how easy it is to bring the instances to mind/availability, are they easily retrievable from memory

avail heuristic and causes of death

• Eg which is a more frequent cause of death - homicide or stomach cancer

• Although stomach cancer is more than 5 times more frequent, pps estimated homicide is 1.6 times more frequent than stomach cancer

• So many reportings of homicide killings so easier to retrieve cases from memory

• Lichenstein, Slovic etc 1978

biases from people misunderstanding heuristics

• sample size

• regression to the mean

sample size

• Intuitively - larger sample should give a more stable estimation, so fewer deviations on average. Correct answer is the small hospital will have a day with more than 50%

regression to the mean

• If have a particularly bad go, next go is statistically more likely to be good and vice versa 

• Eg if you have a really good day, by statistics tomorrow will be worse

• the more extreme the og score, the more regression we expect

why use heuristics

• Time-accuracy tradeoff - The quicker a decision strategy can be applied, the less accurate it tends to be. Heuristics (cog shortcuts) help make fast decisions

• Despite the biases they sometimes lead to, heuristics are often efficient and accurate

anchoring bias and effortfullness

• Making adjustments is effortful, making pps less cog busy leads to a greater adjustment from the anchor

• under time/cog constraint, use heuristics more 

• Grey = number given to pps, busy group = remember digits of numbers, not busy group = nothing else

• Busy = closer to anchor so use heuristics more

anchoring bias in real life

• if you dont know much about the numeric answer to a q, eg pop of cov and someone asks you is it larger/smaller than 300k, you might guess the real number might be close to the one they mentioned

• Reasonable assumption - people are overall cooperative and provide relevant info, probably asking q for a reason 

framing effect and other meaning

• A patient is considering high-risk surgery and asks a doctor about its prospects.

• Positive frame: 80% chance of surviving surgery

• Negative frame: 20% chance of dying from surgery

• Are the two frames really equivalent?

• Reasonable assumption: the doctor’s choice of frame conveys additional information.

• Positive frame: survival chance increased from 0% to 80%

• Negative frame: chance of dying increased from 0% to 20%

• Reasonable assumptions they are actively choosing the frame

representativeness heuristic in real life

• sometimes is acurate, eg butter more calorific than apple

availability heuristic k example

• in most cases 1st vs 6th 7th etc the most easily accessible is the first and there are more words with that name

• so can sometimes work

availability and estimating citites

Researchers measure time taken to recognize cities - can then identify availability

cities 2

improving our own decisions

• Make informed decisions based on the understanding of heuristics and biases

• Eg Framing effect - 75% lean vs 25% fat, levin et al, company is just choosing a frame, can try to inhibit automatic response 

• Rep heuristic - have a person and have to decide if they are tech-savvy. Consider whether they represent a person that is tech savvy - experience bias? 

availabilty heuristic in real life

• contribution of you and your partner to housework

• more likely to say you de more

• self bias

• So know to communicate to other what we have done etc to avoid arguments. Also may not know all the things other people have done 

  
  
  

anchoring bias in real life

• Anchoring is used in negotiations

  • Just don't fall for the anchors the other side use, also using anchoring techniques

  • may use it for salaries etc know to offer higher in response
    
    

influencing others decisions

• Researchers and practitioners can design behavioural interventions to improve people’s decision making

• Framing effect - scale down from deluxe pizza with 12 ingredients, default $11, each ingredient subtract 0.5

• Build up basic pizza by selecting from 12 ingredients, default $5, each ingredient add $0.5

• In the scale down condition, end up with a lot more ingredients in the end, also related to anchoring

option orders

• pps taste wines one after the other and choose their fav after tasting all the wines

• People will choose the first one, if want people to choose first one, put it in first position

• Companies are also aware of this effect - will sponsor hotels etc on website so it is the first thing people see

option order and healthy food

• Vary presentation order of food items in a breakfast buffet

• Either healthy then unhealthy or the other way round

• The first food was selected by >75% diners.

• The first three comprised 66% of all the foods diners took.

• Serving the healthy foods first - more healthy foods selected

• If you want to help people make better decisions, give them the option you want them to take before the others

option salience

• 1. Ask participants to rank all the options from best to worst.

• 2. Show pairs of options on the screen for different lengths of time. Make one of the items brighter than the other. Ask participants to choose.

• Brighter option is visually more salient

• When you increase the time they start to choose their actual preferred option

option salience in real life

• In supermarkets, place healthy foods next to a less healthy but popular option

• pre-cut vegetables next to minced meat,

• increase the salience of healthy options.

• Sales of healthy foods increased by 61.3% per customer.

default options

• Red save button - auto choice, dont have to change anything to save, other option, opt out of default. Default more popular. Also organ donation 

• Eu bans pre-ticked website boxes to aid customers

relativity in option comparison

• design decoy options to make it appear better

• eg print subscription same as print and web subscription, but more expensive than web

dictionary example

• When rated separately, price for a is higher

  • When A and B were rated separately, price they would pay is higher for A  

  • When put them next to each other - b has double the entries - make b look better, so willing to pay more