wk12 - Judgements, Decisions & Reasoning

Question 1

Which descriptions best match Herbert Simon’s (circa 1990) terms ‘bounded rationality’ & ‘ satisficing’?

Answer 1

1. Using experience to construct an expectation of how good a solution we might reasonably achieve, and halting search as soon as a solution is reached that meets the expectation
2. Humans reason and choose rationally, but only within the constraints imposed by their limited search and computational capacities.

Question 2

What were the two findings from Paul Meehl (1954) study of Clinical vs Statistical Predictions?

Answer 2

1. Clinical prediction performs very poorly relative to stats.
2. Overweights case characteristics & underweights base rates

Question 3

What is the difference between system 1 and 2, and what is question substitution?

Answer 3

System 1 is fast & automatic, 2 is slow and controlled.

Question substitution is the mechanism proposed to underly heuristics - Instead of seeking the answer to some complex question, 1 seeks the answer to an easier question it believes to be related.

Question 4

What are the three general-purpose heuristics?

Answer 4

Representativeness

• Probabilistic judgements are substituted with assessments of resemblance.
• i.e. The likelihood that X is a Y, is substituted for; the degree to which X ‘looks like a’ Y.

Availability

• Factors which come to mind easily are assigned greater weight in the formulation of judgments. We judge the likelihood/frequency of an event by the ease with which instances of it come to mind.
• e.g.
  
  • Question: “What percentage of Hollywood celebrities are divorced?”
  • Substitute: “How readily do examples of Hollywood divorces come to mind?”

Affect

• Judgments are made in accordance with the intensity of the emotion felt.
  
  • Questions “How large are the benefits of nuclear power?”
  • Substitute: “How do you feel about nuclear power?”

Question 5

How was the conjunction fallacy bias shown in the ‘Linda-Bank Teller’ study? And in which heuristic does this bias arise from?

Answer 5

• The ‘Linda is 31, bright, single & outspoken. Studied Philosophy.’
• Pts were presented w/ 8 potential job/character categories and asked to rate 1-8 most probable fit for Linda.
• Also asked to rate her similarity (representativeness) to those same categories.
• Y-axis = mean rank probability rating by pts
• X-axis = mean rank similarity rating by pts.
• Inference; they’re highly correlated, revealing conjunction fallacy - looks like, therefore probably is.

Question 6

What is the proposed mechanism of system 1 and the 3 proposed ways of improving system 2’s intervention in system 1 errors?

Answer 6

System 1: Question substitution - Instead of seeking the answer to some complex question, seek the answer to an easier question you believe to be related.

To Improve System 2 Intervention:

1. Provide rewards to motivate participants to check their intuitive impressions.
2. Ensure participants are not simultaneously having to perform other kinds of mental effort.
3. Another idea is commonly cited, but does not work: Increase metacognitive difficulty (Alter et al, 2007; Meyer et al, 2015)

Question 7

Describe the below biases that arise from the Representativeness Heuristic and the key studies/evidence/for each.

1. Insensitivity to Sample Size.
2. Misperceptions of Randomness.
3. Belief in ’The Law of Small Numbers’.
4. Hot Hand Fallacy.
5. Base Rate Neglect

Answer 7

1. Insensitivity to Sample Size: People ignore sample sizes when making probabilistic guesses. BUT Small samples will reveal either much larger or much smaller values than is representative at the population level.
   
   • Evidence: Small vs Large Hospital abnormal birth rates of boys. Pts think the same rate… but it is a small sample that would produce more abnormal rates.
2. Misperceptions of Randomness: The tendency to impose spurious causal narratives on random data.
   
   • “Randomness usually has more repetition than most think”.
   • Evidence: 6-sided dice Green/Red probability.
3. Belief in ’The Law of Small Numbers’: The law of large numbers “implies that large samples tend to produce results that represent the population.
   
   • Example: rolling a single die many times will produce an average value very close to 3.5.
   • The bias is the reverse - too much faith in study designs w/ small samples - termed by KT.
4. Hot Hand Fallacy: Ppl who’ve achieved recent success has a (temporarily) increased propensity to achieve more success.
   
   • Players who score 3 x in a row are not more likely to hit the next shot.
5. Base Rate Neglect: Base rates are not properly integrated with specific case-specific information but are instead neglected.
   
   • Study: Lawyer-Engineer paradigm.
   • No Description of individual = probability integrated base rate information.
   • Dick (generic description) = 50/50 - base rate ignored somewhat.
   • Jack (s/type engineer) = 95% ignored base rate and pick representativeness of description.
   • T/away: Conclusion: base-rate information tends to be neglected when specific information about the case is available.

Question 8

What is the availability heuristic? Describe the biases:

1. Due to the effectiveness of a search set.
2. Egocentric Bias
3. Outcome

Answer 8

factors which come to mind easily are assigned greater weight in the formulation of judgments. We judge the likelihood/frequency of an event by the ease with which instances of it come to mind.

• Question: “What percentage of Hollywood celebrities are divorced?”
• Substitute: “How readily do examples of Hollywood divorces come to mind?”

1. e.g. Study: Words more frequent starting with letter ‘K’ or in the third letter is “K, i.e. Kiss or Lake
2. Spousal contributing to grocery shopping - times shopping are most available..are perceived as more favourably than reality. Biased to do more than their fair share.
3. A decision with a positive outcome is rated as superior to a decision with a negative outcome even when the information available to the decision-maker was the same in both cases.
   
   • e.g. Hospital study with surgery-survival chance options
   • Op 1: 8% chance of death and raises life expectancy from 65 to 70.
   • Op 2: 2% & raise to 75 - but last patient dies.
   • Pts choose op 1 - even though worse outcome..

Question 9

How did the Assertive-Unassertive Listing study propose it was both easy and amount of content that underlies the availability heuristic (Schwartz et al., 1991)?

Answer 9

2 Groups:

1. List 6 egs of being assertive & 12 unassertive
2. List 12 egs of assertive & 6 of unassertive.

• After recall task, assertiveness questionnaire is filled out.
• Red numbers are higher & indicate:
  
  • Listing only 6 instances of being assertive was “easy” & lead to the conclusion the individual was indeed assertive.
  • Thinking of 12 instances of being unassertive was experienced as “hard” & lead to conclusion that the individual must not be unassertive.. & therefore, assertive.

Question 10

What is the affect heuristic and the two biases: Negative risk/benefit Correlation and Insensitivity to Numbers?

Answer 10

Judgments are made in accordance with the intensity of the emotion felt

• Questions “How large are the benefits of nuclear power?”
• Substitute: “How do you feel about nuclear power?”

Negative risk/benefit Correlation:

(Finucane et al., 2000) risk-benefit Nuclear Power study.

• In real-world risk & benefit positively correlated / In peoples minds & judgement: negatively correlated.
• Study had 4 condition:
  
  • A. Information NP is high benefit > pts rate risk to be low.
  • B. Info NP is low risk > pts rate benefits high.
  • C. Info NP benefit is low > rate risk high
  • D. Info NP risk high > rate low benefit
• Even though information of NP is only about benefit or risk, participants infer the other. People are consulting affective heuristic.

Insensitivity to Numbers.

Slovic & Peters (2006): Airport Safety Study

• Improve safety by purchasing equipment which is exy.
• Pts asked to rate support of each equipment all rated as being effective in saving lives to a different degree.
• T/away: saving a % of 150 lives (second bar 98% of 150) rated higher than saving all 150..

Question 11

What is the so-called anchoring heuristic and how did the African UN-Wheel of Fortune study offer evidence for its existence?

Answer 11

• Wheel of fortune - rigged to land on 10 & 65.
• Pts asked: is the number higher or lower than African UN member countries?
• Then asked to give an estimate percentage. Each was ANCHORED their rigged number.

Question 12

What are the 3 criticism of KTs work?

Answer 12

• Lack External Validity
  
  • Minor changes to wording/context can debits people
  • Emphasising role of random sampling.
    
    • Responses greatly improved under this system,
    • Experiment but got the subjects to blindly sample the description papers from an urn. SELF SAMPLE.
      
      • Gigerenzer et al. (1988) Lawyer-Engineer Replication.
      • Elicit frequencies instead of probabilities.
• 3. Vaguely Theorised & Lack of formal Modelling.
  
  • Availability & representativeness are so general they can explain a wide range of findings post hoc.
    
    • Formal models should be proposed that make quantitative predictions.
• Overstate the Problems caused by Computational Limitations of our Brains.

Question 13

What is the recognition heuristic?

Answer 13

If one of two objects is recognized and the other is not, then infer that the recognized object has the higher value.

• San Diego vs San Antonio population question.
  
  • Americans 62% correct
  • Germans 100% correct
• What if we recognise both? Take the Best.
  
  • Next cue.. has it hosted Olympics… is there a famous film? yes.. Anchorman - SD.

Prediction Wimbledon 2003: When heuristics work.

• Recognition accurately predicts the Wimbledon winner better than rankings… i.e. Fed.

Question 15

What were some regrets from Kahneman?

Answer 15

1. Using the term heuristic
2. Including anchoring.
3. Substitution mechanism not restricted to uncertainty.
4. He missed a crucial third heuristic the affect heuristic.