Chapter ?? Extremes outcome are more likely in small samples. Confidence is driven by cognitive ease and coherence.  Chapter 15: less is more We have problems assessing probability (confused with plausibility). Two joint statements are mistakenly seen as most probable because it adds coherence to a story. Given a context where we learn that Linda has social concerns and is involved in demonstrations:

Linda is a a bank teller Linda is a bank teller and is active in the feminist movement

When asked people say that the second is more probable. We think in term of representativeness. People wouldn't do this mistake if the statement didn't add coherence, like: "Linda is a bank teller and ate a banana yesterday". Out of two sets of objects, one with more objects but out of them some are broken, people tend to give a higher value to the set where no broken objects are present ("less is more"), because we tend to make an average value of the set. This is not true if people were given a side by side listing  first. About probability again. We don't always see that a set is including another one. The more specific one is less probable.

I will lose the first set I will lose the first set but win the game

This isnthe same when asked about percentage of categories (one including thebother one). Yet when asked "how many", instead of "what percentage", people are less prone to errors, maybe because it is easier to represent the elements. If a die has four Grenn faces and two Reds. The following sequences are sorted by probability:

RGRRR GRGRRR GRRRRR

People tend to think the second one is more likely since it is more representative (but it contains one more than the first one). Chapter 16 Bayesian statistics. Example of the cab drivers

In a city, 85% of the cabs are green, the others are blue. A witness saw a blue cab, but he judges colors with an accuracy of 80%. What is the probability that the cab was blue?

Example 2

In a city, cabs are distributed equally between blue and green. Green cabs have 85% chance t

The two examples are mathematically the same but they feel different because one is giving "statistical base rates" ("independent" information about the population of cabs), the other gives a "causal base rate" (something we can make sense of in the context). About learning. A (true) experiment revealed that 4 out of 15 students would not run for help when hearing calls on the speaker once they assume someone closer would help. If we hear surprising statistics we will still have problems applying them later when we see individuals and are asked if they are helpers.  If we are faced with individuals, and judge them and are surprised we are wrong, then we will more likely generalize and apply this statistic. Nisbett and Bordiga summarize this as:

Subjetcs' unwillingness to deduce the particular from the general was matched only by the willingness to infer the general from the particular.

You are more likely to learn something by finding surprises in your own behavior than by hearing surprising facts about people in general. Chapter 17 - regression towards the mean Regression towards the mean. Exceptionnal/mediocre performances cannot be expected to be reproduced since they are less probable. The next perfomance is likely to be closer to the mean. Theory développés by Sir Francis Galton. Chapter 18 - taming intuitive predictions  Our intuistion tend to predict rare events from weak evidence. System 1 does overconfident judgment. It's less biased to consider the correlation between the evidence and the conclusion and account for it by doing a regression to the mean.  Chapter 19 - the illusion of understanding Hindsight: we think that we knew what was going to happen, but we are biased because now we do know. Knowing things affects how we think we thought about it in the past. Chapter 20 - the illusion of validity The illusion that we understand the pasts fosters overconfidence in our ability to predict the future.   There is a 50% chance that hitler would have been a woman. There is no correlation in the performance of a trader from one year to the next. Knowing more allows slightly better predictions, but experts tend to be overconfident and less reliable.  Chapter 21 - intuitions vs. formulas Using a formula, Meryl managed to predict wine prices with a correlation of .90. In all fields investigated, formulas have equaled or outperformed experts judgement. Humans are inconsistent and this is destructive for predictions. Experts tend to contradict themselves 20% of the time when judging the same thing under different occasion. This can partly be attributed to priming,our system 1 is sensitive to context/environment.  The authors suggest the following for any evaluations: define 5-6 criteria, find some factual questions for each criteria, ask them all for one criterion at a time, come up with a grade after each group of questions, sum up all the grades of each criterion, possibly add a final global grade given with eyes closed.  Funny formula:

Marital stability needs frequency of love making higher than frequency of arguments

Chapter 22 - expert intuition, when can we trust it. Intuition cannot be trusted in the absence of regularities in the environment. Subjective confidence cannot be trusted neither. Chapter 23 - the outside view  Before doing an estimate it's best to look at the general statistics from similar cases. People tend to be over optimistic on the gains and underestimate the losses/failures. Chapter 24 - the engine of capitalism Optimism is good for your health. Maybe not so much for your business. People tend to be overconfident, and this is determined by the coherence of the story, not so much about the facts.

90% of drivers believe they are better than average 

We live in a society where uncertainty is not well seen (financial analysis, CFO, doctors, etc). But listening to overconfident expert can have dramatic consequences. People ten to give confidence intervals four times too low. Klein suggests a "premortem" procedure to come up with threats that might have been neglected: ask experts to imagine they are one year from now and the project has fails, they need to write the history of what happened.  Chapter 25 - Bernoulli's error People are not rational in their decision, they don't look at expected values. Fechner suggested that suggestive sensations follow a log scale compared to a true physical quantity: Increasing from 10 to 100 feels the same as from 100 to 1000.  Bernoulli derived a theory were indeed the increase of utility between 1 and 2 was higher than between 9 and 10. This explains why people tend to be risk adverse. This theory works well but doesn't account for the value of the starting point (i.e. Your wallet before you gamble).  Chapter 26 - prospect theory People become risk seeking when all their options are losses.  Chapter 27 - the endowment effect. People are reluctant to give up an object they already own.  Chapter 28 - bad events  Our brain gives priority to bad news.  We are highly loss averse. Not achieving a goal is way worse than the good feeling of exceeding it. Animals (and people) fight harder to prevent losses than to achieve gains. This is why animals defending their territory more often win.  Chapter 29 - the fourfold pattern We do know weigh our decision based on probabilities but based on a decision weight which is driven by two effects. The possibilty effect: we overestimate outcomes that have a low probability. This leads to gamble (for gains, where we are risk seeking), and insurance policies (for losses, where we are risk adverse). The certainty effect: we underestimate outcomes that have very high probability (with an intensity even higher than the possibility effect). This leads to the acceptation of unfavorable settlement to ensure a gain in fear of disappointment (risk adverse), and to the hope that a large loss will be avoided (risk seeking). People take desperate gambles leading to catastrophic consequences instead of accepting the large sure loss. This loss appears too painful and the hope of complete relief is tooenticing to make the sensible decision that it is time to cut your loss and not go further.  No mater which corner of the fourfold pattern, the fact that the decision weights differ from the probability will be costly in the long run.  Chapter 30 - rare events Rare events are either ignored or most often overweighted. Also, the probabilities of rare events and overestimated.  The reason the probabilities are overestimated is due to a memory bias, an activation of system 1 via emotions and availability. To estimate a large probability we often look for its opposite and go through memory to find rare events. It's best to evaluate all kind of scenarios and make sure that it adds up to 100%. The sources of overweighting are: obsessive concerns, vivid images (tsunamis, earthquakes), concrete representation (1 out of 10 instead of saying 10%), explicit reminders.  The denominator effect is the fact that we tend to look at absolute numbers and not do percentages (divisions). When given a concrete reprensentation (e.g. 1 out of ten) we can react/overweight twice as much as if a percentage was given to us. This can be use to manipulate people's reaction/option/decisions one way or another.  The chapter also raise the question of our sensitivity to probability depending if the question is dealing with emotions or money.  Chapter 35 - two selves  There is the experiencing self and the remembering self. Peak-end-rule: We remember things based on the highest emotion and the emotion at the end (colposcopy and hand in cold water experiment). Duration effect: the duration has no effect of the ranking of pain/pleasure  Tastes and decisions are shaped by memories but memories can be wrong.  What we learn from the past is to maximize the qualities of our future memories not necessarily of our future experience. Chapter 36 - life is a story The two effects, peak end rule and duraction neglect, also apply when evaluating quality of lives of other people.  Also we are concerned about the quality of the stories more than their feelings. For our own lives, the remembering self is the one taking decisions (choosing our next vacations). We want to store memories, removing those reduces the quality of the experience. Chapter 37 - experienced well being The chapter differentiate "experienced well being" and "life satisfaction".  Most of the time we experience well-being in the moment, and this is improved by awareness (e.g. thinking that you are eating), or action, switching to passive activities (tv), to more active forms (socializing and exercise).  Physical health and social contacts are good measure to added if someone had a nice day.  Life satisfaction is increased by education and money. But more educated people tend to report increased stress, and more money does not increase well being above a level. Money may reduce the ability to enjoy the small things (when priming students with ideas of wealth, they don't express as much pleasure thinking of chocolate). Religion does not reduce depression and worry but it has a favorable impact on positive affect and agree reduction.  Just like a colonoscopy: the evaluation of our lives and the actual experience are related but also different. The easiest way to increase happiness is to control your use of time: can you find more time to do the things you enjoy doing? Chapter 38 - thinking about life Regarding happiness, there are two selves as well, the experiencing one and the thinking one. We usually make "affective forecasting" errors, we predict wrongly how we will feel about things (typically decision to get married): nothing in life is as important as you think it is when you think about it. We also make focusing illusions, we exaggerate the effect of changed circumstances on our future well being, e.g. climate is not that determinant for well being. Most of the time we do not think about those things, or they are only felt temporarily. It's true to bad and good situations (e.g. Paraplegics). We neglect our capacities to adaptation as well! In these illusions , we often neglect time, and we have a bias. You will pay little attention to your car with time, so it might be better to engage in a social group where you meet many times.  The author argues that experienced well being is in average unaffected by marriage, because it changes some aspects for the better or the worse.  Genetics affects experienced happiness and life satisfaction.  (Teenage, graduate) Goals make a large difference on experienced happiness. Reaching them brings satisfaction.