Beliefs Flashcards
what are the three main examples of non-standard beliefs
overconfidence,
law of small numbers,
projection bias
what is overconfidence
overestimate one’s own skills (or the precision of estimate) (stock markets think know esact price when in reality impossible) P̃(good statet)>p(good statet)
what is the law of small numbers
gambler’s fallacy and overinference in updating P̃(st|st-1)
what are the two types of overconfidence
over-optimism: overestimate one’s ability, over-precision: overestimate precision of one’s estimates
when is overconfidence greatest
difficult tasks, forecasts with low predictability, undertakings lacking fast, clear feedback
motivating facts behind barber and odean (2001)
men more overconfident than women about financial decisions, test hypothesis: men trade more than women, and by trading more they hurt their investment performance more than women, compared investment decisions across men and women
difference between overconfidence and risk loving
overconfidence - you think only yourself are going to have a good outcome, risk loving - you would apply this to any other decision as well
who tested the link between gender stereotypes and overconfidence
D’Acunto (2015), stereotypes can be (experimentally) manipulated
what is priming
evoking associations through stimuli which may change subjects’ behaviour
what is the design of the first gender stereotypes experiment
D’Acunto (2015), 1) Elicit subjects’ risk attitudes with lottery choices, 2) randomise assignment to i) control, ii) male prime and iii) female prime, control just read neutral short reading, treated read text priming strong gender identity 3) Elicit subjects’ risk attitudes again
results of gender stereotypes first experiment and drawback
D’Acunto (2015), male primer did increase risk tolerance, women became less risk tolerant (insignificantly), preferences or beliefs? (is it changing men’s risk tolerance or beliefs in positive personal outcomes
design of second experiment on gender stereotypes and what paper
D’Acunto (2015), same as Exp.1 but different outcomes (all post treatment), guess # of wins in 50/50 lottery for i)yourself ii)a random peer, reward for correct guess
results of second experiment for gender stereotypes
D’Acunto (2015), significant effect of priming only on OWN # of wins, only for men, priming changes men’s beliefs of winning game of chance, evidence for beliefs in positive personal outcomes
what is a summary of D’Acunto (2015)
experiment 1: gender priming -> ∆RiskTol, experiment 2: gender priming -> beliefs in pos. personal outcomes
what is the design of the third experiment for the gender stereotypes experiment
D’Acunto (2015), same as Exp.1 but different treatments/primes, Task: recall instance where one was 1)relaxed (control) OR 2)powerful OR 3)successful, 2)->overconfidence in one’s power (also over random-ness) 3)->overconfidence in one’s ability (doesn’t help w/ random-ness)
results of third experiment for gender stereotypes
D’Acunto (2015), overconfidence primer significantly increased risk tolerance, success primer did not cause any significant effects
what is the gambler’s fallacy
after a streak of red on the roulette wheel, one “expects” next draw to be black
law of small numbers explained
law of large numbers: with very large N, the share of H and T is going to get very close to 0.5, law of small numbers: this doesn’t need to hold for small N individuals expect random draws to be exceedingly representative of the distribution they come from
what does law of small numbers mean in terms of the distribution of an event
individuals expect random draws to be exceedingly representative of the distribution they come from
what is a conceptual way of understanding law of small numbers from Rabin (2002)
iid signals (H or T) from urn drawn with replacement, many people believe as if drawn from urn size N
what are the two cases from Rabin (2002) ways of looking at small numbers
Case 1: probabilities known (≈ Gambler’s fallacy) Case 2: probabilities not known -> overinference, after signals of one type, expect next signal of same type, people exaggerate likelihood that short sequence of iid signals resemble the long-run rate at which those signals are generated
what is an example of overinference
firing decisions of football managers
what is the setting of case 1 for the experiment on law of small numbers
Rabin (2002), probabilities θ known, returns to mutual fund drawn from urn with 10 balls, 5 Up and 5 Down (with replacement), θ=P(U)=0.5, observe ‘Up, Up’ compute probability of another Up, Bayesian 0.5, law of small numbers 3/8<0.5 (in your head you take out the two that have been taken), expectations driven by how representative outcome would be given underlying distribution
what is the setting for case 2 for experiment of law of small numbers
mutual fund with manager of uncertain ability, replacement with 10 balls, Probability 0.5: fund well managed (7 Up 2 Down) Probability 0.5: fund poorly managed (3 Up 7 Down), observe sequence ‘Up, Up, Up’ what is P(Well|UUU) law of small numbers guy overinfers from the distribution they see
