our experiment
TRANSCRIPT
How would you bet?An experiment by
Anne Marie, Claralinda & Marina
LABREVEALED
PREFERENCES
MotivationExperiments aim at studying actual behavior of people
REALITYTRUE
PREFERENCES
PROBLEM
≠HYPOTHETICAL
BIAS“the difference between hypothetical
and actual statements of value” List & Gallet, 2001
MotivationHolt & Laury, 2002
risk preferences with high payoffs subjected to HBpeople are less risk adverse in choices involving high hypothetical payments (compared to real payments)
Are there experimental designs that minimize HBwhen high payoffs are used?
Bushong et al., 2010
WTP affected by physical presenceWTP increases sizably when people evaluate a real object (compared to just text or a picture)
QuestionRICH HYPOTHETICAL DESIGN incentives still hypothetical but enriched by simulation ✦ incentives have physical dimension ✦ ask people to consider the incentives as real and theirs
Have RH incentives a different effect on people risk-decisions?
SPARE HYPOTHETICAL DESIGN no care in let people imagine the incentive as real
Is the effect of RH incentives affected by the type of task (boring or engaging)?
NEW usual
MAIN QUESTION
CORRELATED QUESTION
ExpectationsBet (RH) < Bet (SH)
Bet (BT) > Bet (ET)
revealed risk-preferences closer to realitypeople are mainly risk-adverse
Why?
utility of people affected by funpeople get utility from taking risk in boring task
Why?
Design
TASK 1 THE HORSE RACEBet on which is the winner of the horse race
2 SIMILAR INDEPENDENT TASKS ‣ 6 equiprobable options ‣ 100€ available to bet (multiple of 5€) ‣ 9x gains
the engaging task
TASK 2 THE RANDOM GENERATIONBet on which is the generated number by random.org
the boring task
The enhanced experimental design ‣ PLAY MONEY
‣ SIMULATION STATEMENT
RICH HYPOTHETICALGROUP 1
Design2 GROUPS physically separated
the treatment group
SPARE HYPOTHETICALGROUP 2A common experimental design
the control group
YOUR NAME YOUR NUMBER
In a few minutes random.org will generate a number from 1 to 6.
You have 100€ that you can find in the envelope we give to you. Feel free to examine these money as much as you want. Please behave as if this money were your real money!! Think that it comes directly from your own wallet.
You can choose only on ONE number. On that number you can bet an amount of money from 0€ to 100€, in multiple of 5€. For example you can bet 0€, 5€, 20€, 50€, 75€, 95€ etc. If your number is generated, you will get 9 times your initial bet. For example, if you bet 10€ on number 1, and number 1 is generated, you will gain 90€.
Please fill in the form below with your choice. Remember to think of this money as if it were your real money. You have 2 minutes to take your decision. If you have any question, please raise your hand and and we will come to you.
I bet on number the amount of €
After you made your decision, please put the money you choose to bet in the envelope we gave to you. We will collect it.
Good luck!
TASK 2
DesignVISUALITAZION & TASK ATTRIBUTION
Questionare there differences in how rich hypothetical incentives impact on risk propensity compared to spare hypothetical incentives?
only with respect to high-stake payoffs.
Taskstwo different task, one more engaging and one more boring to study if the effects of both the incentives are influenced by the type of task/play.
Groupstwo different group. One group will have the rich hypothetical version of the two tasks and the other the spare version of the two tasks.
BORING TASK
ENGAGING TASK
RICH
SPARE
people does not reveal true preferences
Can we move to them?
behave regardless the type of incentive
Holt & Laurysubjects become increasingly risk averse as payoffs increase with real incentives, whereas increasing the scale of hypothetical incentives has no effect on behavior.
Goal Lower/reduce cost in economic experiments
HYPOTHETICALREAL
REVEALED PREFERENCESSTATED PREFERENCES
INCENTIVESCHOICES
choices with no consequences for decision maker
choices with consequences for decision maker
CHOICE EXPERIMENT
any situation in which a decision maker is asked to rank/choose from 2/more alternatives
Background differences in real vs hypothetical choices
hypothetical bias in hypothetical experiment
STUDY REALITY IN THE EXPERIMENTAL LAB
respondent bear some real consequence from making the decision
DIFFERENT EXPERIMENT FRAMING
MONEY MAKE PEOPLE MORE RISK-ADVERSE
RICH MAKE PEOPLE MORE RISK-ADVERSE?
RISKY CHOICES / GAMBLE
the effect of actually playing gambles was to make subjects more risk-averse
Holt & Laurypayoff scale effect
RANDOMIZATION TASK1 EXPLANATIONDISTRIBUTION TASK 1 COLLECTION
TASK2 EXPLANATIONCOLLECTION
PAYMENT?
R groupinstruction paper 1Rmoney envelope
S groupinstruction paper 1S
R grouppaper 1R + envelope
S grouppaper 1S
DISTRIBUTION TASK 2
bag & numbers slides 1
S groupinstruction paper 2S
R groupinstruction paper 2Renvelope
slides 2
S grouppaper 2S
R grouppaper 2R + envelope
COLLECT RESIDUAL MONEY
HORCE RACE
video
last slide
RANDOM.ORG
link
last slide
COLLECT RESIDUAL MONEY R group
Marina
Anne MarieClara
Anne MarieMoney
Clarainstructions
Marina
Clara
Anne Marie
Clara
Marina
Clarainstructions
Anne Mariemoney
Anne Marie
Marina
Clara
Anne Marie
Clara
ResultsRICH GROUP vs SPARE GROUP
63,33$
49,77$
0$
10$
20$
30$
40$
50$
60$
70$
Spare$Group$ Rich$Group$
Means&Mean average bet for the group calculated in both tasks
p-value 0,268
ResultsBOX PLOTS
RH 1 SH 1 RH 2 SH 2
ResultsBOX PLOTS
RH 1 SH 1 RH 2 SH 2
0
20
40
60
80
100
April JuneJuly
greater difference in average in TASK 2
ResultsTASK 1 vs TASK 2
48,48$
65,22$
0$
10$
20$
30$
40$
50$
60$
70$
Task$1$ Task$2$
Means&
Mean average bet for the task calculated in both groups
p-value 0,117
47,71%
66,82%
0%
10%
20%
30%
40%
50%
60%
70%
80%
Female% Male%
Means&
ResultsFEMALE vs MALE
Mean average bet for the group calculated in both tasks
p-value 0,0725
SIGNIFICANT AT 10%
Discussionour data
‣ tell us that the pattern of bets is sufficiently variable ‣ don’t tell us if there is or not a true effect
If there were a TRUE EFFECT, we would need a larger size sample (N≥23) in order to detect it.
Definitive conclusions can not be taken from our sample
Further research
Is the behavior in RH settings similar to
behavior in real situation?
– Anne Marie, Claralinda & Marina
Thank you for your attention!
If you have any question we are here :)