cog wk 11 Flashcards
downsides of group descision making
- conformity
- authority
- Polarisation
- Groupthink
Benefits of group descision making
- wisdom of the crowd
- improvised reasoning
why is authority a problem in group decision making
evidence (don’t rlly need to remember)
Those in lower authority may be scared to raise concern to those in higher authority as they believe they should’ve already spotted it.
larger authority gradients lead to less communication from lower rank to higher
found that the biggest authority gradient between army helicopter officers in cockpit, the more likely there was to have an accident.
40% of junior co-pilots reported not relaying concerns about safety to senior pilots
led to development of crew resource management
how is polarization an issue in group decisions
take indivs, take their attitude to risk. assume average would be group attitude. BUT when they start talking to each other, they become more extreme in their views.
Groupthink as a problem for group decisions
cohesive groups striving for unanimity and avoiding conflict or criticism.
Exacerbated in homogenous teams, where members are, or become too similar.
e.g. space shuttle exploded but engineers raised the concern that it was too cold for a component. The managers ignored this concern
Wisdom of the crowd
guess the weight of a cow. collect all p.p answers, the median value was within 1% of the true weight of the cow.
this is true with multiple prediction.
by looking at medical diagnoses, what did a study find about similarity in wisdom of the crowd
groups of doctors with similar levels of diagnostic accuracy will outperform the best doctor in the group
Surprisingly popular algorithm - wisdom of the crowds
in niche specialist knowledge, wisdom of crowd usually doesn’t hold.
BUT if ask
- what is your answer?
- what will most people answer?
look at predicted percent agreement with own answer.
if there is a very low percent, we discover that someone has niche specialist knowledge
wisdom of the crowd (within)
study
asked random history question
2 conditions:
- reliable condition (just give another estimate)
- dialectical bootstrapping condition (- assume you were wrong - think of a few reasons this could be. - what do these considerations imply. - based on new perspective, make a new estimate)
Found
- no improvement in repeated
- significant improvement in dialectical. only 4%
- 7% increase in one other person.
systematic errors in wisdom of the crowds
- eliminates random errors
- Not so good at systematic errors e.g .anchoring
improved reasoning in group reasoning
p not q task.
in indiv task, 9% chose correct, 43 chose PQ.
in group task, 75% chose correct response.
groups of more than 2 tend to better than the best individual could’ve