Neglect of base rates
Ignoring valid info eg statistically there are more farmers than librarians (Steve and Bill)
Attribute substitution
Occurs when we have limited sources of info
- use easy measure as substitute for objective property => what makes an easier judgment: the similarity of Steve to our expectations for the different occupations
When i think of a librarian, does it fit with Steve? => we know how to do this process
(Matching examples based on little info)
Representativeness heuristic
How representative / fits prototype of our concept
Why did we neglect or didn’t use potential info for Bill and Steve?
Bc we were not trying to figure out the likelihood that Steve is a librarian, rather, we substituted it with “does Steve fit our definition of a typical librarian?”
Representativeness heuristic => errors and biases
Characteristics / stereotypes may work sometimes, but perceived similarity is based on limited set of features => may not be objective info and tell us a lot
Conjunction error
- mistakenly judging a combination of two events to be more likely than either one separately
- can be explained in terms of representativeness
- take smth that seems likely and pair it when smth unlikely
-A] seems probable (mixed), like it fits okay (feeling) quick judgment and we think how well Bill’s description fits it
=> representativeness does not work like probability! Basically, cognitive-wise, A&J is more representative than J alone
Avoiding conjunction error / neglect of base rates
- reword question
=> imagine a hundred ppl like Bill, how many do u think are accountants? How many play jazz? How many are A&J
=> imagining pool of ppl helps with probability
Application: hiring and interviews
Potential biases
- encourage judgment thru representativeness
- does the candidate fit the typical image of what is expected?
- the way they talk, conduct themselves, their appearance => activate ur stereotype of ur ideal candidate
Being a female programmer in a male dominated field: bias against female candidates. How u imagine a programmer may not fit, and stereotypes may not be related to performance, depend on prior representation
Peak end rule
Peak (most intense point, could be best or worse moment) and the end (of the event) of an experience can predict your overall (subjective) evaluation on this
Factors not influencing:
- duration
- proportion of enjoyableness
So why is the peak end so important?
- not averaging the sum, rather 2 parts
=> our mind calls to the most representative aspects
- the peak and end are the ones that stand out the most!
- peak: intense and salience
- end: recency effect
Availability heuristic
- how easy to generate an example
Emphasis on ease
In the names demo, the female names were from famous ppl, thus our mind remembered this and thought there were more women than men
Famous is easier to recall
Is the availability heuristic biased?
Depends, could provide a good estimate in some cases, but the ease of retrieval doesn’t depend solely on how common it is (what about repetition from the media??)
Also, retrieval of examples may be difficult
Non-representative exposure
Retrieval bias
- how easy it is to think of smth?how many examples can i come up with?
=> 2 examples i can come up, so yes i believe pollution will decrease. Can’t come up with 8, so pollution will not improve
=> can come up with 2 examples of why pollution is not improving, but I can’t with 8, so i believe it will improve
Exposure bias
Frequency of occurrence correlation with commonness
Error: exclusion and not representativeness
Salience / vividness in an anecdote
- a friend’s recall of first hand experience will stick out more than a consumer report (vividness and personal news!)
