Thinking Fast and Slow - Study Notes - Version 2: 10 - Part 2: Heuristics and Biases

Part 2: Heuristics and Biases

S1 cannot handle statistics — e.g. lowest kidney cancer rates in U.S. counties are in mostly rural, sparsely pop’d, & located in tradit’ly Republican states in Midwest, S., W.; to explain S2 begins work, uses causal facts & suggest’ns fr. S1, (clean air? rural lifestyle?) — but highest kidney cancer rates are in same counties — correct explan’n is: sample is small, i.e. statistical, not causal
even sophisticated researchers have poor intuitions, poor underst. of sampling effects

in practice psychologists do not use calcul’ns to decide on a sample size, use their judg’t, which is often flawed —often chose samples too small, 50% risk of failing to confirm true hypotheses
we all hv bias to believe that small samples rep. the whole pop’n

S1 cannnot disting. betw. degrees of belief — e.g. “In a telephone poll of 300 seniors, 60% support the president,” info abt survey ignored — people not adeq’ly sensitive to sample size
S1 does not doubt, suppresses ambiguity, creates coherent stories — unless the message is immed’ly negated, its assoc’ns spread as if the message were true
law of small numbers (small samples) is example of bias that favours certainty over doubt
we are prone to exagg. consistency & coherence of what we see — believe in small samples — (cf. halo effect, think we know more than we do) — S1 creates a rep’n of reality based on little data, makes too much sense

S1’s assoc. method looks for causes of an event — but statistics simply relates the event to other possible events, not intst’ed in causes
looking for causes à errors in underst. random events — e.g. sequence of births of m/f babies, BBBGGG, GGGGGG, BGBBGB, last seems more likely but all are equally likely, random — we expect the regular seq’s to result fr. cause, not random — widespread misunderst. of randomness — e.g. bombing of London, pattern w. large gaps, spies? no, randomness can look like clusters or regularity — basketball, 3-4 baskets in a row à belief in “hot hand,” but just random, cogn. illusion — investor’s seq. of good years, CEO’s acquisitions, probably all luck, random
best schools are small, but so are worst
focus on content of messages, little attention to info abt reliability — —result in view of world simpler & more coherent than data justify — e.g. exagg. faith in small samples
statistics often produce observ’ns seemingly calling for causal explan’n — but search for causes inappropr. — many facts ß chance, incl. accidents of sampling
causal explan’ns of chance events = wrong.

“Yes, the studio has had three successful films since the new CEO took over. But it is too early to declare he has a hot hand.”
“I won’t believe that the new trader is a genius before consulting a statistician who could estimate the likelihood of his streak being a chance event.”
“The sample of observations is too small to make any inferences. Let’s not follow the law of small numbers.”
“I plan to keep the results of the experiment secret until we have a sufficiently large sample. Otherwise we will face pressure to reach a conclusion prematurely.”