System 1 vs System 2
System 1
fast, automatic, emotional, unconscious
Tasks
2+2, sudden sound, read words on billboards
skills are born or become automatic
always active
System 2
slow, effortful, logical, conscious
focus on a voice from 1 in a group, look for a person with red jacket in crowd, check validity of compley logical argument
neccessary condition: focus and attention
Experimental Research
treatment structure
gruppe 1 bekommt geld, gruppe 2 nichts
beetween / within subjects
incentives
typically monetary incentives
limit budget: prize winner
replication
repeat event
gießen + milwaukee gleiche studie
control
possibility to controll extraneous factors
Interplay between System 1 and 2
System 1 -> impressions become beliefs -> 2
System 2 -> takes over when needed
System 1 is prone to biases
Was sagt die Studie in Bezug auf System 1 und 2 us?
Cognitive reflection test (Frederick, 2005)
MIT Studenten am besten abgeschlossen
more reflective people are more patient as they choose the later larger reward
correlation between Score and cognitive measures
high correlation between cognitive reflection and college admission
decision making under risk
Risk vs unvertainty
under risk
probabilities of occurence available or at least assessable
under uncertainty
probabilities of occurence unknown and impossible assess
objective vs subjective
objective
probability to roll a 3
subjective
tommorow’s rain probability
Bayes’ theorem
what we have
what we look for
Rechnung
Prior probability: 𝑝(𝑦𝑖)
Conditional probability likelihoods : 𝑝(𝑠𝑗|𝑦𝑖)
Posterior probability: 𝑝(𝑦𝑖|𝑠𝑗)
𝑝𝑦1∗𝑝(𝑠1|𝑦1) // £ 𝑝(𝑦)∗𝑝(𝑠1|𝑦𝑘)
Conclusion
prominent bias
positive
good to infer likelihood of unobservable event from an observable signal
negative
difficulties whet it comes to implementation of bayesion updating
=> in reality: people tend to make mistaks when faced with the challenge of updating probabilities
=> prominent bias: base rate neglect
expected utility theory (EUT)
value function V
E(a) = £ (p*a)
St. Petersburg paradox (Bernoulli, 1738)
theoretisch unendlich Gewinn möglich
warum setzen Leute begrenzt Einsatz?
Leute achten nur auf expected utility
vergessen expected value zu bewerten
gutes Beispiel für utility function U = ln(x)
Certainty equivalent (CE)
Risk Premium (RP)
CE = e^E(ln(U(x))
RP = E(x) - CE
Different risk attitudes
risk aversion
risk seeking
risk neutral
E (U(x)) < U (E(x))
Nutzen vom erwarteten Value ist größer als erwartete Nutzen
its not avoiding risk -> acceptance of risk in excahnge for compensation
e.g.: St. Petersburg lottery, insurance policy, investment
E (U(x)) > U (E(x))
Risk aversion am Beispiel
E(x) = 25
E(U(x))
U(E(x))
CE
RP
expected utility theory (EUT) (Neumann, Morgenstern; 1947)
utility function assigns utility value
higher expected utility (E(U)) -> stronger preference
axiomatix system => acception axiom => accepting EUT
Kern der rational decision theory
4 Steps
Step 1
alle Outcomes aufzählen
Step 2
Ranking Outcomes
Step 3
Indifferenzwskt für jedes Ereignis erstellen
Step 4
Entscheidung auswerten anhand expected utility
EUT
Kritik
eignet sich nicht als deskriptive theory
=> knowing all decision implications, the underlying axioms would not be accepted in the first place
Was war das erste economics experiment?
St Petersburg lottery
Between subject design vs within subjects design
between subject
2 untergruppen mit unterschiedlichen treatments
within subject
jeder Teilnehmer durchläuft beides
Advantages of experimantal research
jederzeit wieder durchführbar
control the extraneous factors
Real vs hypothetical payoffs
real
higher stakes leads to risk aversion
hypothetical
no significant effect
Is there a Gender effect when stakes are high?
Wich Gender is more risk avers when stakes are low?
affluent individuals are
Men, younger and taller and highly educated are
Last changed5 months ago