Rationality Flashcards
two main questions in debate of rationality
what does it mean to be rational? are humans rational?
example of irrational statements
more red apples than red things (conjunction fallacy), 2/3 survive better tahn 1/3 die (framing effect), make sure no one goes to concert without ticket means make sure everyone with ticket goes to concert
standard/classical view of rationality
rational is to be in accordance to principles of reasoning based on rules of logic/probability
what classically rational beings are good at
means to end reasoning, decision-making, math, probabilities, logic, consistent beliefs
linda the bank teller
tversky and kahneman (question rationality), Linda cares about social justice, is she a bank teller or bank teller AND feminist, 89% get wrong - conjunction fallacy
conjunction fallacy
A cannot be less than A and B, but get wrong
how subjects response demonstrates conjunction fallacy (linda)
less likely to be bank teller AND feminist than just bank teller
whether responses on linda come from improper education
85% of decision science grad students at Stanford also get wrong
Lawyer v engineer
30 engineer, 70 lawyer, choose 100 descriptions, first sounds like engineer, second is neutral, third has no description, version 1 = lawyer dominant, version 2 = engineer dominant, do base rate neglect so think first is engineer
base rate neglect
ignore all you know when you get a little bit of information in the other direction, like person athletic with fast car - nurse or pro basketball player? would say basketball but statistically is nurse
bayesian reasoning
prior probability, updating, probability of hypothesis on given body of evidence depends on prior probability of hypothesis, if all 19 die vs 19 of 20 die update
cases where subjects commit base rate neglect
1/10 guys hipsters, 7/10 hipsters have mustaches, so higher chance that guy is hipster, also cliff diving example
asian disease
us outbreak where kills 600 people maybe… case 1: program A saves 200, program B 1/3 probability save 600 and 2/3 probability save no one, case 2: program A 400 die, program B 1/3 probability no one die, 2/3 probability 600 die - actually have same probability but 28% do risky with save and 78% do risky with die, framing effect
framing effect
framing of works have effect, choice involving gains are often risk averse and choices involving losses often risk taking
asymmetry in risk taking shown by subjects in asian disease cases
subjects take less risk with save and more risk with die
wason selection task
turn over to see if card has vowel and odd number on other side, should pick E and 4 - need to turn over even number to make sure isn’t vowel and turn over vowel to make sure it has odd number
subjects answer to wason task (incorrect answer)
pick E and 5, but 5 could have a vowel or consonant so doesn’t violate rule
T and K idea of representativeness
people predict by representativeness and select outcome by how much they represent the essential features of the eidence, so when there’s things that affect likelihood but not representativeness then get rules wrong - using heuristics
how representativeness explains mistakes of linda and laywer v engineer
base rate neglect, because sounds representative of engineer, ignore probability that is being offered
overview of T and K argument and which ideas of ecological rationality will disagree with
rational only if can reason with logic, not good at reasoning and logic, so not rational, idea that rationality is bounded so it has limited time/knowledge/capacities
ecological v standard/classical reationality
ecological takes into account real world environment with limitations of organisms, whereas rational is just rational due to rules of logic
idea of satisficing
make adequate solution and move to next due to limited capacity/abilities, first best choice
idea of bounded rationality with ecological rationality
organisms have limited time/knowledge/computational capacities and each environment has its own challenges, Herbert Simon
modules and central processing
modules are domain specific and give input to conceptually based, domain general central processing, where central processing carries out higher cognitive processes, system 1 (modules)
evol. psychology views on modules - massive modularity
idea that even higher cognitive processes are carried out by modules, contrasts with other psychologists’ views
evol. psychology view on why modules sometimes fail
modules are specialized so it’s not surprisig if they function well only under specific circumstances
frequentist hypothesis
some of reasoning comes from probability but they’re made to take frequency information and input and make frequency as output, if problems concern probability in terms of frequency then do much better at it
altruism and the conditions under which it is and isn’t favored by natural selection
give resources to non related people is bad unles it is reciprocal so must be able to detect when it will be reciprocal
cheater detection module
module that tracks if people are cheating in social exchange, only works in right situations
new version of linda
ask of 100 people fit description above, how many are bank tellers or bank tellers and feminists, only 22% get wrong
new version of wason selection task
better at concrete examples, like if drink beer need to be over 20, 75% get right and pick beer and 16 year old to flip, same with sandwich for money and meat - I give you thing and you didn’t give me thing, make contractual
fast and frugal heuristics
human rationality is bounded, so best way to maximize and make efficient is fast a frugal
city naming study
san diego or san antonio larger, german students all right, US students 62% right, use best heuristics and first differentiation is where german students stop because only have heard of san antonio, Gigerenzer and Goldstein
take the best heuristic
choose between alternatives, search in order of validity and stop when cue discriminates then pick the best of the 2
general idea behind fast and frugal heuristics and how successful they are
1/N strategy, satisficing, default, imitate majority/successful, find they often do better than using more painstaking methods
1/N strategy - fast and frugal
investing equally between stuff does better than wallstreet advice
default - fast and frugal
go with default unless there’s a problem with default
imitate majority/successful - fast and frugal
do what they do and usually works out okay
