Availability and Representativeness
(This page last updated 25 November, 2005.)
Induction: Starting with specific facts and drawing general
conclusions from them.
Deduction: Starting with a general principle and working out
specific facts that follow from it.
On this page, we will examine two different specific problems of
induction and how we solve them.
1. Determining the Frequency with which Something Occurs.
Availability
- Strategy for estimating the frequency with which something
occurs.
- Try to recall as many instances as possible. The more
instances recalled, the higher the estimate of frequency.
- Kahneman and Tversky propose that it is a general strategy
used in many different situations.
- In most real-world situations in which it is used,
Availability usually works well.
- However, it is not guaranteed to give the right answer.
- It is a heuristic, a quick procedure that produces answers
that are usually right, but not always.
Experiment to Demonstrate Availability
- Subjects see list of names. Then asked whether list contained
more men or women.
- Some names are famous, and some are not.
- All lists have equal numbers of men and women. If the men on a
list were relatively more famous, subjects report that the list
has more men.
- If women on list were relatively more famous, subjects report
that the list has more women.
Recent Memories and Availability
- Recent events are easier to remember than earlier events.
- If Availability is used to estimate the frequency of an
occurrence, then something that has occurred recently will seem to
occur more frequently.
- This explains:
- why people are more likely to fasten seatbelts after seeing
an accident.
- why you might be more likely to back up your computer files
after a friend loses theirs.
The Conjuntion Rule: P(A&B) <= P(A)
In some cases, Availability can lead to conclusions that violate
the conjunction rule.
2. Determining the Probability that an Exemplar Belongs in a
Category
Representativeness
- Another heuristic proposed by Kahneman and Tversky.
- Estimate probability that an exemplar belongs in a category by
assessing how representative that event is of the appropriate
category.
- In the example in class, Linda is more representative of a
feminist than a bank teller.
- Therefore, descriptions including feminism seem more likely
than "bank teller" by itself.
- In this case, representativeness leads to a violation of the
conjunction rule.
- P(A&B) <= P(A)
- The probability of (bank teller and feminist) will always
be less than or equal to the probability of (bank teller).
Base Rates
- Base Rate: How often an item of a particular type appears in
the population.
- People generally ignore base rates in categorizing things.
They tend to rely more on representativeness.
- Example in text, p. 393.
next class: Deduction
Psych 315H: Cognitive
Psychology
Kyle Cave
Psychology Dept.
U.
Mass.