What was Tversky and Kahneman (1983) Linda problem?
- Students given a statement which they had to rate what options were more probable
- Linda is 31 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, and also participated in anti-nuclear demonstrations.
What is The conjunction fallacy in the Linda problem?
- 85% of undergraduates judged h (bank teller and feminist) to be more likely than f (bank teller)
- But: h is of the form ‘Linda is both X and Y’ f is of the form ‘Linda is X
- So: h must be less likely than f
What was argued about the Linda problem regarding to how many options there were?
- Perhaps because there are so many options, people simply don’t notice the relationship between f and h?
- Tversky & Kahneman (1983) gave 142 people a simplified version of the problem containing only the two options of interest (f & h).
- Still, 85% made the conjunction fallacy
What was argued about the Linda problem regarding to the language?
- Perhaps people interpret ‘Linda is bank teller’ to mean ‘Linda is a bank teller & not active in the feminist movement’?
- In that case, their judgements could be correct
- Tversky & Kahneman (1983) changed the statement to: ‘Linda is a bank teller whether or not she is active in the feminist movement’
- 57% of people made the conjunction fallacy
What is Heuristics?
Quick procedures that frequently find the correct solution but are not guaranteed to do so (‘cognitive short-cuts’ and ‘rules of thumb’)
What are Algorithms?
Procedures that are guaranteed to find a solution if one exists, but may take a very long time
What is the Representativeness heuristic in the Linda problem?
-“Linda is a bank teller and is active in the feminist movement” is more representative of the ‘kind of person’ Linda seems to be.
What were the findings of the Linda problem?
- Separate groups asked to rank the eight ‘Linda’ outcomes on either representativeness or on probability
- Correlation of 0.99 between the mean rankings
- Similar proportions in each group made the conjunction ‘fallacy’ (note this was not a fallacy for the representativeness group)
What task did Tversky and Kahneman (1983) do looking at the USA and the Soviet Union?
- Participants in International Congress on Forecasting. Different groups evaluated the probability of statements a and b
- a. A complete suspension of diplomatic relations between the USA and the Soviet Union, sometime in 1983
- b. A Russian invasion of Poland, and a complete suspension of diplomatic relations between the USA and the Soviet Union, sometime in 1983
- Probability estimates for b higher than for a
- Different task
What did Tversky & Kahneman (1983) find looking at -ing words?
- If you pick 2000 words at random from a novel, will there be more words ending ‘ing’ or more words ending ‘n’?
- Most said ‘ing’ which must be wrong
- Availability heuristics, more likely to bring it to mind
What happened in the task looking at words starting with ‘r’?
- If you pick a word (over three letters long) at random from an English text, is it more likely that the word starts with an ‘r’ or has ‘r’ as its third letter?
- Most participants said ‘starts with r’ was more likely. Actually, many more words have ‘r’ as their third letter, but these are harder to bring to mind.
- Easier to generate words beginning with r, than ending with r
What is the Availability heuristic?
-The easier it is to bring an event to mind, the more likely that event is judged to be.
What did Lichtenstein et al (1978) find when looking at Asthma and tornados?
- Asked participants to judge which was the most likely of two causes of death
- For example, is it more likely that someone will die in a tornado or from asthma?
- Identified “many, often severe, misconceptions”
- 58% of student sample thought tornado was more likely (actually asthma is 20 times more likely)
- May be due to availability, unlikely to hear stories of people struggling with asthma
What did Slovic et al (1982) find when looking at death?
- Over estimation and under estimation of what cause death
- E.g. smallpox vaccinations vs motor vehicle accident
What did Coombs and Slovic (1979) find about death and media coverage?
- Newspaper reports over-represented more ‘dramatic’ causes of death (disasters, accidents, homicide etc.)
- People’s risk judgements were related to frequency of media coverage
- More evidence for use of an availability heuristic
What did Young et al (2008) find when looking at diseases with high media presence?
Diseases with high media presence (e.g. avian flu) judged more severe than other diseases
What did Clark and Teasale (1985) find when looking at mood?
- Positive and negative memories recalled more in the appropriate mood (in patients with depression)
- Therefore, words and associations change their availability
What task showed end anchoring?
- Tversky & Kahneman (1974)
- Group 1: Is % of African countries in UN < > 10%? Now estimate number of countries (24%)
- Group 2: Is % of African countries in UN < > 65%? Now estimate number of countries (45%)
- Anchoring participants at either 10% or 65%
- Group 1: estimate product of 8 x 7 x 6 x 5 x 4 x 3 x 2 x 1 (2,250)
- Group 2: estimate product of 1 x 2 x 3 x 4 x 5 x 6 x 7 x 8 (512)
What Range of heuristics are there?
– Representativeness
– Availability
– End anchoring
What do mathematicians say about single events?
- Some say that probabilities can never be assigned to single events
- Others disagree
- Is it fair to test people’s judgements against this norm?
- Chase et al., 1998
How did Cosmides and Tooby (1996) change Cascells medical study?
-Rephrased Cascells et al.’s (1978) medical diagnosis problem in terms of frequencies
– Input information changed from percentages to frequencies
– Question changed from single-event probability to frequency judgement
What did Fiedler (1988) when looking at the frequency version of the Linda problem?
- Frequency version of ‘Linda problem’
- 100 women fit this description: – Single, outspoken, very bright. Studied philosophy at university, deeply concerned with social justice and an antinuclear protester
- How many are bank cashiers?
- How many are bank cashiers active in the feminist movement?
- 77% correct
Frequentist approach: Evaluation
- Performance on some tasks improves when problems are phrased in terms of frequencies rather than single event probabilities
- But might this simply make the calculations easier?
- Also, other tasks are performed badly despite frequentist phrasing
