10 - Investigating the causes of PR Flashcards
what was the objective of Tversky Slovic and Kahneman (1990)
discriminate empirically between explanations (incentives, intransitivity, procedure invaraiance violation) using ordinal payoff scheme and diagnostic tool
- distinguish between intransitivity and violation of procedure invariance as explanations of standard PR
if invaraince principle holds
choice and monetary value are the same
if standard PR observed - and it is not because of violation of invariance what is the other possibility
subject must have intransitive preferences
what is the diagnostic procedure used by TSK
they check if subject has intransitive preferences by checking whether M($) > X > M(P)
what was experiment TSK
- get them to elicit M(P) and M($)
- 3 pairwise choices between P, $, X
if intransitive preferences what did TSK expect to find
M($) > X > M(P)
they choose $ over X
they choose X over P
P > $ ~ M($) > X > M(P) ~ P
what would be the explanation behind TSK finding
the opposite of intransitivity
X chosen over $
P chosen over X
invariance must be violated
X chosen over $, even though monetary value of $ is higher than X
* inference = over valuation of M($)
if P is chosen over X - even though X > M(P)
* Under valuation of M(P)
what did TSK find
results
violations of invariance dominates over intransitivity
- only observed 10% cases of intransitivity
- 90% of observations make transitive choices
- violation of invaraince play big role in standard PR that TSK observed
what was the OPS
used by Cubitt, Munro, Starmer (2004)
ordinal payoff scheme
- choose between P and $ bets as a pair
- put value on each P and $ separately
- after tasks completed, random device selects choice or valuation
- if choice: subject gets bet they chose in choice task
- if valuation: subject gets bet they put higher value on
what is GET
how is it related to OPS
generalised economic theory
- theory that implies facing a task subjects will work out which options are best and then choose their preferred option - behave to get their preferred choice
- in OPS - no GET can explain dominance of standard PR over counter PR
what are the 3 psychological views of standard PR given by CMS
- people think differently about choice and valuation
- prominence hypothesis
- scale compatability hypothesis
- task-goal hypothesis
what is prominence hypotheiss
how does it explain if win-probability is prominent
an attribute weighs more heaviliy in choice than valuation
even if you value the $>P, they think that the most important attribute is the win probability - so choose the P bet - more likely to win
- depends if the think probability or the payoff is more important - regardless of value- and that attribute weighs more heavily in choice task than valuation task
what is scale compatability hypothesis
explains standard PR
an attribute weighs more heavily the more compatabile its scale is with response mode of task
- when task is to give monetary value then you place more weight on the one that has the most money
what is task goal hypothesis
explain standard PR
weighs more heavily when goal of task seen as differentiating between items
- trying to differentiate you put different weights then when you try and equate things
what was CMS experiment
tests whether scale compatability predicts standard PR
2 treatments
* monetary valuation - set the MV so that the 2 options are equal
* probabilistic valuation - set the PV so that both options are equal
- responses are numbers
what did CMS expect to find about the hypotheses
- prominence - standard PR in both PV and MV
- task goal - no PR
- scale - standard PR in MV, counter PR in PV
why would scale compatability predict the directions
MV and PV manipulation should only make a difference here
- standard PR in MV
- you weight money more than when scale in money
* put higher valuation on $ bet than P bet - counter PR in PV
- you weight win probability more when scale in probability
what did CMS find
results
MV treatment
- PR observed
- standard PR > counter
- means that P choosers are more prone to PR than $ choosers
what did CMS find
results
PV treatment
- counter reversals more frequent than standard PR
- more PR amongst $ choosers - because in valuation they weight probability more because scale is porbability
what did CMS find about
scale compatibility hypothesis
suggests that scale hypothesis is a psychological hypothesis that explains economic behaviour
- only hypothesis to predict treatment difference is scale compatability
- so it matters what scale - what kind of PR is seen
what conclusions can we make about task sensitivity
what does this mean for us
task sensitivity of preferences exists
warrants caution about generalising from any one measure of preference
- use multiple measures
- appears that people have different preferences depending on the task they face