Ch. 4 Flashcards
Anchoring and adjustment heuristic
- people make an estimate by using some starting value and then adjustng it to arrive at a final conclusion, however people’s final answers are not suficiently adjusted away from the initial anchor
- when people fail to make suficient adjustment away from an initial figure that they treat as an anchor.
- can be self-generated anchor
eg. wheel spinning and pop estimate by Tversky/Kahneman
What’s different about anchoring/adj. vs. representativeness, availability?
Frederick proposes that it dosn’t operate by replacing one type of judgement by a different judgment type, e.g. replacing probability judgment with a similarity judgment for representativeness)………
More gerneral than those, applies equally well to frequenvt judgments, values, magnitude, even causal attributions
How robust is the anchoring effect? What is some evidence of its robustness?
- the effect has been observed in a variety of domains, including pricing, judgements, etc.
- although people use it, they rarely report being influenced by it
- people use anchors even when warned not to use them
- experts and non-experts both affected by anchors
- incentives are unsuccessful in reducing anchoring
- extreme or implausible anchors also affect judgments
What are the underlying processes of anchoring effect?
seletive accessibility account - The comparative question task activates information in memory that is subsequently more accessible when making an absolute judgment.
Various supporting findings; e.g. people with greater domain knowledge are less susceptible to the effects of irrelevant anchors (Wilson et al, 1996).
But this account does not seem to have any need of the adjustment element.
selective accessibility account
and pros/cons
the comparative question task activates info in memory that is then more accessible when making a judgment
support:
people attent to shared features more than unique ones
ppl with greater domain knowledge are less susceptible to irrelevant anchors
anchroing speeds recall of rando associated words
cons:
adjustment doesn’t seem ecessary
low external validity - in real life anchors are self-generated
Epley & Gilovich “other” findings -
self-gnerated anchors
head nodding
reporting
more self-reports of adjustment when self-generated anchors rather than experimenter anchors
head nodding/shaking led to les/more adjustment with self-generated anchors…… but no effect with eperimenter generated ones
Reasons why Epley (2004) think adjustment is insufficient?
- People don’t know the right answer, so unlikely to adjust
- adjustment requires attention, people may be lacking in cognitive resources
- people who dislike effortful thinking adjust insufficiently from self-generated anchors
Cognitive processes underlying external anchors vs. self-generated anchors
external - activates related material in the mind that is used to produce a response
self-gen - people generate a salient figure that they know to be wrong, which they then adjust away from
Hindsight bias
- Uncertain outcomes often seem more likely after it is known that the outcome occurred
- creping determinism - that an outcome was inevitable or more forseeable than it rly was
- Two classic studies: The Gurka study and the Nixon study
creeping determinism
that an outcome was inevitable or more forseeable than it rly was
Gurka study
what design did it use?
hypothetical design
People received decriptions of four unfamiliar historical events… 4 outcomes were described
some participants told that one outcome had occured
this knowledge altered ppls assessment of the relevance of rano scenario info
people seemed unaware that they were being influenced by this knowledge
Nixon study
memory design
prior to trip to China, people asked to predict likelihood of various outcomes
2 weeks after the visits, the ppl were asked to give their original probabilities of what would happen
probabilities were inflated for what had occured and reduced for what hadn’t happend
hypothetical design vs. memory design?
Gurk - hypothetical - judgments of those with end knowledge compared to those without
Nixon - memory - present self judges, compared to future self’s recall of prior judgment
no difference in hindsight bias between these 2 designs once accuraetely recalled answers were removed
How robust is hindsight bias?
It still occurs when ppl are made aware of it….
however it is reduced when participants…
- gave reasons for their diagnoses
- were given a written record of their prior thoughts/ideas
- were asked to engage in counterfactual thinking
counterfactual thinking
thoughts that shine light on other possible outcomes
What’s the impact of cultural and individual differences in hindsight bias?
Choi/Nisbett - East Asians showed more hindsight bias than westerners
Musch - personality factors related to self-presentation and suggestibility play big role in hypothetical design
Alternative theories of hindsight bias
- Outcome info is immediately assimilated into memory structures
- restrospective reconstructino of judgment to lign up with extra info on outcome…. more misrecaling of the initial story in ways that endorsed the outcome
- Pezzo’s model - initial surprise at an ourcome leads to sense-making activity…. if it makes sense it reduces surpise and increases hindsight…… if unsuccessful then surprise is maintained and can lead to reverse hindsight bias (the true outcome is seen as less predictable)
HINDSIGHT ARISES FROM SENSE-MAKING cognitions
Pezzo’s sense-making model
pros
Pezzo’s model - initial surprise at an ourcome leads to sense-making activity…. if it makes sense it reduces surpise and increases hindsight…… if unsuccessful then surprise is maintained and can lead to reverse hindsight bias (the true outcome is seen as less predictable)
after the feedback, thouthgts inconsistent with the outcome were associated with lower hindsight
in scenarios that were difficult to make sense of no reverse hindsight was found……. suggests that the east of sense-making, rather than “suprisingness” is most important factor in hindsight
e.g. when doesnt make sense ppl have less hindsight
Sense-making caveats
people are more confident about answers that they can retreive more quickly (availability?)
resrouce depletion - generating too many counterfactuals leads to more hindsight bias!
How to avoid hindsight bias?
Slovic & Fischoff - people read 4 descriptions of studies…… told that they had either just occured or would occur…….. asked to give prob that results would be replicated
just occured got higher estimates….. however, when asked to consider reasons why EITHER outcome might happen, it reduced the hindsight bias (though not eliminating it)
thus, consider how things could have gone differently!
Hindsight summary
- Tendency to retrospectively view events as more foreseeable or inevitable than they actually were.
- Can be reduced by reminding people of their earlier thoughts or asking them to engage in counterfactual thinking.
- 2 broad types of theory: (1) automatic updating of memory given outcomes, (2) retrospective updating at the time of judgment.
- There is evidence that this second type of theory may involve a process of sense-making.
Anchoring summary
- Quantitative judgments are sometimes affected by the presence of an anchor.
- Externally-provided anchors heighten the accessibility of a small range of values in memory (selective accessibility account).
- Self-generated anchors are known to be wrong, so are associated with greater adjustment. But adjustment may still be insufficient for various reasons (true answer not known, competing demand for cognitive resources, lack of motivation).
