Processo decisionale e razionalità, Sbobinature di Psicologia della Comunicazione

Scarica Processo decisionale e razionalità e più Sbobinature in PDF di Psicologia della Comunicazione solo su Docsity! INTRODUCTION: JUDGEMENT, DECISION AND THE VIOLATION OF RATIONAL AXIOMS  Doing nothing is also a decision, we don’t always have to decide to do something. Often we have almost two alternatives, but more alternatives you have harder is to decide something, you don’t know what to do. Less alternatives make the decision easy.  We think about what alternative in the future we will regret less, for this reason less alternatives make easy your decision. DEFINITION OF DECISION  As All decisions have a common feature, they are made: 1. Intentionally: we do something that we supposedly want to do 2. Consciously: when we do something we have control of our behaviour This view of decision making corresponds to the way everyone of us believes humans behave  Hastie e Dawes (2001) proposed the following definition of decision: a decision is our reaction to a situation characterized by three different components:  A chance to evaluate more than one path of action  alternatives. It also considers the case where we must decide to do something or not to do something. Not doing something is also a decision  The presence of expectations (subjective probabilities) regarding the likelihood of achieving the outcome of each path of action.  We have an idea of the likelihood that we have of achieving the outcome that is associated with every alternative. It considers how much that outcome is important to us and our perception of the likelihood to get it. o The percentage is subjective, not for everyone a number has the same meaning.  The presence of consequences associated with the outcomes  consequences that can be assessed based on personal values and objectives of the decision maker. We make o When we decide we gamble, it is uncertain the outcome o In real life is difficult understand the probability and having an immediate feedback of our decisions, or knowing the consequences of our past decisions (many things are not in our control, something will happen in a more far future or something is the result of more than only one decision that was made in one specific time).  Decision tree: We must decide which train to take to go to a meeting. The first train is more expensive, but it arrives earlier so there’s a better chance to make it on time for the meeting. The second train is cheaper but there’s a chance you will not make in time to go to the meeting. Depending on the person, someone will give more importance to spending less money or to get on time for the meeting. These are the reasons that make someone choose something and not something else.  The probabilities are calculated quickly in an intuitive way DECISION, JUDGEMENT, CHOICE  The previous example allows us to analyse the differences among three terms that people often use as synonyms. Decision is referred to the whole process of selection of an action: 1. Representation of the problem 2. Estimate of expectations  cognitive representation of the data 3. Evaluation of outcomes  What will happen 4. Selection of an alternative  how we will feel if happens something and what we prefer  Judgment: It is referred the part of the process that is before the actual choice, to the components of the decision process relative to:  Evaluation of outcomes.  Evaluation of the consequences associated to each outcome. It is referred to our understanding of the value of every alternative (good or bad) 1  Choice: Indicates the sub-process of the decision in which an alternative is chosen among several other options. The choice is the last part of the process of decision and it consist in the selection of the alternative that we think is the best one.  There is also a post choice phase where we evaluate the consequences of our decisions and we learn how to improve them in the future. We are very good in doing that if we have immediate feedback (our brain thinks in terms of causal connections), but as we know in real life the real outcome is often delayed in time and there are other factors which interfered with the final outcome. UNCERTAINTY, RISK AND AMBIGUITY  One of the points in which the normative and psychological approaches differ the most is people’s attitude towards:  Choices under uncertainty.  Risk perception.  Choice under ambiguity.  Decisions are made under uncertainty when: we do not know exactly what is going to happen, but we know the likelihood of each possible outcome.  When we decide under uncertainty, we are taking a risk.  Example (it considers also a risk): Imagine you are asked to bet on the drawing of YELLOW ball from an urn filled with 100 balls, 50 yellow and 50 red. If a yellow ball is drawn from the urn you will win €50. How much are you willing to pay to play this gamble?  In this circumstances people evaluate risk (the likelihood that the expected outcome will not be achieved) in a subjective way:  As we will see, there is no such thing as risk aversion.  Rather, psychologists speak of loss aversion  Decisions are made under ambiguity when: we do not know exactly what is going to happen and we do not even know the likelihood of each outcome to happen (main difference with uncertainty).  You know the outcome, but you don’t know the probability of each one to happen, because they are completely out of our control.  It doesn’t permit to calculate the possible value of the outcome  Example: Imagine you are asked to bet on the drawing of a YELLOW ball from an urn filled with 100 balls, both yellow and red. If a yellow ball is drawn from the urn you will win €50. How much are you willing to pay to play this gamble  Usually when the two examples are presented together, we prefer the first one because we know the probability, but if they are presented separately, we choose both indistinctly. This because the presence of a reference point, so an alternative with more information, is more attractive. But separately they are the same for us because we cannot make the comparison (we valuate more the probability).  In the second example the people don’t usually think about the likelihood (that is the same of the first one), but in the first example it is relevant (they are salient) because it is explicit, so they are more predictable.  This form a rational point of view is wrong because the rational model says that we have stable preferences. THE THEORY OF RATIONAL DECISION  The rational agent has all the information: he considers, understand and can compute all the probabilities, but in reality it is a subjective process that create variability.  It was developed by philosophers and mathematicians that were trying to understand how we make decisions, and what is a rational decision. In particularly they tried to understand what’s the rational way to make a decision when the outcomes are uncertain?  The start of this kind of studies has been traced back to 1738 and to the work of a Swiss mathematician, Bernoulli who asks himself: 1. How much are people willing to pay to play a game based on a coin toss? 2  Most scholars who are currently working on rationality do it by relying on the notion of Instrumental Rationality.  In this view, our psychological states and mental processes are rational if they help us achieve our goals.  In this way, rationality does not consider the correctness of our cognitions and inferences, but just the efficacy of our actions. BOUNDED RATIONALITY (THE SATISFICING PRINCIPLE)  However, Simon showed we can define rationality in a different way (procedural rationality) and proposed the concept of Bounded Rationality.  Even a rational agent would not make a further effort if it is not bigger than increasing quality he can get from that effort. We accept a no optimal decision if the effort it requires is not followed by a personal gain/increased quality of the outcome.  Sometimes to make an additional effort is not worth because we don’t gain enough in doing it.  He was the first to analyse the processes  Moreover, in contrast with the assumptions of the neoclassic theory of economics, people: 1. Are not fully informed -> we decide depending on what we see and we know, we don’t look for more additional information 2. Do not have stable preferences 3. Have limited cognitive and computational resources -> our brain hasn’t got the ability to make difficult compute  Simon suggests that cognitive limitations, together with the context in which the decision is made, determine the procedures that decision makers can actually use. Under this view, cognitive limitations of real decision makers and “environmental uncertainty” due to incomplete information are linked together in shaping behaviour.  In other words, people use different procedures and make different choices depending on the context in which they are making a decision: 1. Number of alternatives available. 2. Amount of information about the outcomes of each alternative. 3. Environmental factors (time constraints; social pressure).  In Simon’s view, bounded rationality is connected with the notions of Heuristic and Satisficing  Heuristic = we use simplified strategies to decide. We preserve energies but we are ble to make good enough decisions  Satisficing = we make a decision that left us satisfied even if it is not the best. There is efficiency in doing the optimal decision possible and not the best. o The level of satisfaction depends on the importance of the decision.  Seldom people, in their everyday life, apply the procedures required to select an alternative which maximizes expected utility. o These procedures are very demanding from a cognitive perspective o They consume a lot of cognitive resources (attention, memory, self-control).  People use very simple strategies (heuristics) that do not always allow to make the best choice, but lead, most of the times and with limited amount of time and effort, to satisfying results.  From a practical point of view, heuristics allow people to make an effort-accuracy trade-off: o We accept, implicitly, to be less accurate. o However, we can also make decisions without too much effort.  The term heuristic comes from informatics: they are the processes that combine different information from different sources to give as a simple outcome when we press a button or when we give a command to the computer. In humans it is the process in which our brain brings, combine and evaluate different information to give us an answer or decision. They are useful when we must make decision quicky (ie: in a dangerous situation). They permit us to make good enough decisions, that without experience are often big mistakes. 5  Thanks to his studies on heuristics, Simon showed that people cannot avoid using strategies that allow to: 1. Minimize the exploration of all possible solutions. 2. Consider just a few data so that memory and attention resources do not get overloaded.  Most of the evidence presented by Simon was gathered thanks to field studies: with managers and with chess players. For example chess is a difficult game that requires a lot of compute analyses, but it is a quite predicable game (players predict only some moves because it is too exhausting to remember all the possibilities).  As a result, under conditions of high complexity and uncertainty:  The choice of a satisfying strategy is an approximative form of optimization. It is impossible for a human being to do the best decision, but we try to do the best we can.  It is rational to search for a better solution up to a point and then stop searching when the effort surpasses the increasing quality (it is counterproductive to make an effort without a gain). o At that point, better solutions do increase our knowledge about the problem, but force us to face very high costs (e.g., in terms of time) compared to the advantages that can be gained.  From this point of view, the choice of a satisfying strategy could be considered as rational because: a rational economic agent should stop when the additional benefit does not overcome the additional costs.  Rational in this case is not to do the best but to be the more efficient  Simon’s conclusion is that although the rational agent is fully informed, he will end up using heuristics that simplify the problem and do not require to explore all information. Simon connects the rational agent and the real person: the first uses rationality, the second heuristics, but both are the most efficient possible RATIONAL AXIOMS  One of the most important implications of Simon’s work is that it started a discussion on:  The extent to which people are able to make truly rational evaluations.  The procedure that are used when decisions are made accordingly to optimization vs. maximization criteria.  The most central rational axioms, according to the theory of expected utility, are the following (von Neumann and Morgenstern, 1947; Savage, 1954; Luce, 1965): 1. Independence axiom (or Cancellation) = our decision should be based on what is different between alternatives. 2. Transitivity axiom = we have always the same preferences and based on these we create rankings. 3. Dominance axiom = we choose the alternative that is better on one attribute and not worse in all the others. 4. Regularity axiom = if we add an additional alternative to others, none of the original alternatives should increase its popularity 5. Invariance axiom = the way we measure a option should not make any difference  It is a good normative model, but at the descriptive level (the level in which psychology works) we don’t behave in this way, in fact we violate them.  If the transitivity and invariance axioms are violated the theory of the expected utility collapse. INDEPENDENCE  The independence axiom states that:  Preferences should depend exclusively on the outcomes that are different between alternatives. o The preference depends of what I gain choosing one option and not another one.  In other words, the utility maximization is achieved by eliminating the states of the world that are independent from the chosen alternative.  Allais' Paradox (Allais, 1953). Choose which of the following gambles you prefer: 1. GAMBLE A: 100% chance to win €1 million. 2. GAMBLE B: 89% chance to win €1 million. 1% chance to win nothing. 10% chance to win €5 million. Most people prefer Gamble A, because is a secured or sure gain of money.  Now, choose which of the following gambles you prefer: 6 1. GAMBLE C: 11% chance to win €1 million. 89% chance to win nothing. 2. GAMBLE D: 10% chance to win €5 million. 90% chance to win nothing. Most people chose Gamble D.  In the first case having less money but for sure for us it is a satisfied decision. We don’t want to risk losing everything, so we prefer a little gain. In the second case we are economical rational, in fact we risk a 1% to win more money, but at the end both gambles are the same, and we change preference.  However, the two pairs of gambles (A-B and C-D) are actually the same:  In the second pair (C-D), the outcomes that were independent from choosing A or B were deleted.  In other words, C and D are the same as A and B, minus: A 89% chance to win €1 million.  Despite this, people preferences changed in the second choice compared to the first and this contrasts with the independence axiom.  Certainty effect (Kahneman & Tversky, 1979). Choose which of the following gambles you prefer: 1. GAMBLE A: A 100% chance to win €3.000. 2. GAMBLE B: A 80% chance to win €4.000. A 20% chance to win nothing.  Now, choose which of the following gambles you prefer: 1. GAMBLE C: A 25% chance to win €3.000. A 75% chance to win nothing. 2. GAMBLE D: A 20% chance to win €4.000. A 80% chance to win nothing.  Again the two pairs of gambles are identical (A-B e C-D):  In the second pair, all probabilities have been reduced to one quarter of the original values: 100% became 25% and 80% became 20%.  Again the certain alternative is the most popular, but when we change the alternatives people is more keen to risk to win something more. In the second case there is more uncertainty, because it is a risk leads people to choose the rational decision. Generally we prefer the certainty win or the option that doesn’t lead to any loses.  Certainty effect (Kahneman & Tversky, 1979) works also with alternatives that are not about money. Choose which of the following alternatives you prefer: 1. ALTERNATIVE A: A 50% chance to win a 3-week trip to England, French, and Spain. 2. ALTERNATIVE B: A 100% chance to win a 1-week trip to England.  Now, choose which of the following alternatives you prefer: 1. ALTERNATIVE C: A 5% chance to win a 3-week trip to England, French, and Spain. 2. ALTERNATIVE D: A 10% chance to win a 1-week trip to England.  Once again, the relationship between the alternatives has not changed from one case to the other since the second pair has been created by reducing the chances of both A and B by 1/10.  Reducing the probability of outcomes by the same amount for both alternatives should not influence people’s preferences. What they preferred before (between A and B) should remain their favourite alternative (when choosing between C and D): but in both examples, what we observe is actually a change in preference. We call this phenomenon “preference reversals.”  The alternative that was initially sure is impacted by the fact that it became uncertain. This demonstrates that, in our minds, sure outcomes have a special status.  An equal reduction in the likelihood to win impacts the option that was originally sure more than the one that was already uncertain.  This is not a mistake in real life, because in ancient times it was fundamental for the survival accepting something sure even if less. TRANSITIVITY  The transitivity axiom states that:  If a person prefers alternative A to alternative B, then it should always prefer A to B.  If a person prefers alternative A to alternative B, then the addition of an alternative C, inferior to B, should not modify the decision to choose A. 7  This decision problem can be represented graphically as follows:  In such a situation (both alternatives have something good and something bad), the manager may experience a conflict regarding which project is going to be the best for the company.  Project A has a higher chance to be completed successfully. However, it is more expensive than B (uses up a larger part of the budget) and still it is not a sure bet.  Project B has a lower chance to be completed successfully. However, it is less expensive than A (allows to save half of the budget to fund other projects).  As a result, depending on the decision maker, some may prefer Project A while others may prefer Project B  Under this new condition, the regularity axiom states that: The popularity of the two original alternatives (Project A and Project B) cannot increase. What happens if we add Project C (Cost €50.000. 55% chance to complete it successfully)?  Popularity could decrease if some managers choose Project C.  Or, it could stay the same if no manager chooses Project C.  However, many experiments showed that, once Project C is added into the set of alternatives, the popularity of Project B increases.  Project C is dominated by Project B and, correctly, no one chooses it.  The popularity of Project A decreases.  But this is a violation of the regularity axiom, in fact if we compare the options two by two separately, we understand that Project C is irrelevant, but we use it as a new information that changes our preferences and choices. It make more attractive an option rather that another one.  Compromise effect (Simonson, 1989). Once again, the regularity axion is violated by people’s decisions. This time, the addition of an extreme alternative (Project A) leads to an increase in the relative popularity of the alternative that was originally more expensive.  Choices for Project D increase.  Choices for Project B decrease.  Very few people choose Project A  An explanatory theory of this effect is Reason based choice: the alternative with the best reasons is the one we choose. o We make a compromise between the choices o Since when we have to make a choice, we are afraid of doing a mistake, when we find a justification for that alternative to be the best one, it solve the cognitive dissonance. INVARIANCE  If violated, we can’t predict people’s behaviour. If we want to preserve the theory, we have to create new criteria that account for the variability and the mistakes that people make and that will make the theory more complex, more difficult to apply to more domain and we will lose the big strength of this theory: to be simple and valid for all situations.  The invariance axiom states that: regardless of everything we don’t change our preferences and decisions. But our decisions are not consistent, and we 10 don’t have static ranking of what is best. We are not rational in economic terms, we choose online, by the time we see the alternatives  Procedure invariance: how is measured our preference or how we are asked to express our preference. The way we measure preference influences people’s decision, even if for this axiom shouldn’t be like this. If an alternative is preferred to another one, such preference should not change as a matter of the method used to measure it: o Choice = choosing a or b o Judgment on a rating scale = in a scale 1 to 7 how much do you like instead of b o Willingness to pay (WTP) and willingness to accept (WTA) = how much are you willing to pay for something. If you prefer it you should choose to pay it more or choose to sell it to a bigger price.  How we describe the problems: preferences should not change or “reverse” depending on how they are presented. o Problem described in positive vs. negative terms = presentation in terms of what we can gain and what we can lose. o Joint evaluation vs. separate evaluation = people looks at ranks when they are comparing how they’re doing today respect how they were doing yesterday (yesterday I was richer then today?). This also influences the level of happiness. Moreover happiness and the decision’s ranking depends on the reference point, because it make more easy to evaluate attributes (especially numbers). For example we could be keener to buy a music dictionary with less words but a new cover, instead ones with more word but with an old cover. But with the reference point they were able to choose the dictionary with more words but older.  Preference reversals (Lichtenstein & Slovic, 1971; 1973): when we choose something in one condition, but we choose something else in other conditions. Imagine you are asked to choose which of the following gambles you prefer to play: 1. P-Bet (probability bet = more probability of winning less money): A 28/36 chance of winning €10. 2. $-bet (dollar bet = less probability of winning more money): A 3/36 chance of winning €100. How much would you ask to sell the P-Bet? and how much would you ask for the $-Bet?  The P-Bet offers a high chance to win a low amount of money, whereas the $-bet offers a low chance to win a high amount of money. When asked to choose, most people prefer the P-Bet since it offers a higher chance to win. As a result, the P-bet appears more attractive than the $-bet despite offering a lower win. People prefers having more probability of winning even if the money prize is lower.  When asked to choose, most people prefer the P-Bet since it offers a higher chance to win. As a result, the P-bet appears more attractive than the $-bet despite offering a lower win.  Lichtenstein and Slovic asked the same people to set a WTA price (the minimum amount they were willing to accept to sell the gamble to another player): people give an higher prize to the $ bet, but this is the one they didn’t choose (they give it a higher value). 1. P-Bet: A 28/36 chance of winning €10. 2. $-bet: A 3/36 chance of winning €100. How much would you ask to sell the P-Bet? and how much would you ask for the $-Bet?  Compatibility principle: when someone ask more money, the answer is compatible with the two dimensions of the gamble: how much win. So we give more money to the gamble that offers a higher money prize. Disregard the probability, that become less relevant respect to the money, because the answer is given based on the other dimension that has the same importance.  We can choose something with a less utility value but with less risk, as the probability bet, but we should also ask more money for that  When asked to set a price for the two gambles, the majority of people sets a higher price for the $-Bet than the P-bet, because it offers a chance to win a larger sum of money. As a result, people attach a higher value to the $-bet despite the fact that the chance to win is much lower than for the P-bet. 11  From a rational standpoint, our preferences should always consider both dimensions (likelihood and amount of a win). Even if we value one of the dimensions more than the other, we should be consistent in every circumstances, no matter how our preference is elicited. However:  When people make a choice, they focus on the likelihood of winning.  When people set a minimum price, they focus almost exclusively on the amount of money (outcome). o The money dimension is compatible with the request to set a price to sell the gamble. o Therefore, it is overweighed in this case but not when making a choice.  The type of answer that we give makes us more sensitive to the attribute of the gamble that is more consistent, compatible with the money. It is an attention bias, when there’s a money prize, people don’t look to the probability giving a price. The real outcome of the gamble is the probability for the prize, so if we are economically rational we would compute the expected value and choose the higher. In this case the expect value is the same, but the probability changes: the idea is that we should be consistent, but we value more the bigger prize, but when we have to win something we choose the higher probability.  Framing effect (Tversky & Kahneman, 1981): In a first experiment, people were asked to read the following scenario: 1. Imagine that the U.S. are preparing for the outbreak of an unusual Asian disease, which is expected to kill 600 people. 2. Two alternative programs to combat the disease have been proposed.  A first group of participants had to choose between the following programs: 1. Program A: Adopting this program, 200 people will be saved. 2. Program B: Adopting this program, there is 1/3 probability that 600 people will be saved and 2/3 probability that no people will be saved. When there is a gain people choose the sure option.  A first group of participants had to choose between the following programs: 1. Program C: Adopting this program, 400 people will die. 2. Program D: Adopting this program, there is 1/3 probability that nobody will die and 2/3 probability that 600 people will die. People choose the risk because don’t want to kill people  As shown in the example, the framing effect depends on how relevant information about the problem is described: On many occasions, the same information can be presented either in positive or negative terms. These could be complementary versions of the same information and should not change people’s preferences.  The frame, so the context where the problem is describe, make people change decision  In a second experiment, a group of people was presented with the following scenario: Imagine that you have decided to see a concert where admission is 50 Euros per ticket. As you enter the stadium you discover that you have lost a 50 Euros bill. Will you pay for the ticket? People says yes  A second group read the following version of the theatre problem: Imagine that you have decided to see a concert where admission is 50 Euros per ticket. As you enter the stadium you discover that you have lost the ticket. The seat was not marked and the ticket cannot be recovered. Will you pay for another ticket? People says no, because it seems like a second purchase  From an economic standpoint, the two scenarios are identical since in both cases the decision maker loses 100 Euros: 1. Scenario 1: 50€ lost + 50€ ticket purchase = - 100€ 2. Scenario 2: 50€ lost ticket + 50€ ticket re-purchase = - 100€ As a result, people’s choices should be the same in both versions. However, the second case is different because the money spent for the ticket that went lost has already been assigned to the “leisure time” mental account. People don’t think to their loses and their economic situation, but they think on what did they 12  Thousands of people watched the videos and half of them did not see the gorilla the first time.  Many people did not believe the gorilla was actually present and thought it was added in the replay.  When we focus on something important we become “blind to evidence” and “blind to our own blindness” INTERACTION BETWEEN THE TWO SYSTEMS  Usually, we rely on the intuitive system, while the analytic system remains inactive. This way we can save cognitive energies and spend as less energy as possible. The analytic system should intervene every time something it’s too difficult to be done by system 1.  The intuitive system produces stimuli and thoughts all the time that the system 2 should evaluate. If the analytic system accepts those intuitions then they become beliefs and voluntary actions.  If we don’t know what to do or we don’t know anything about that we trust our system 1 and his intuitions.  PROBLEM!! When everything looks fine, as it happens most of the times, the analytic system accepts the conclusions coming from the intuitive system without modifying them.  We believe in our impressions and behave according to our desires - that’s usually fine  To make a good decision you have to be motivated, to care about it, and you must know what to do, or else you will analyse things with ha automatic system  Only when the intuitive system fails to reach a plausible conclusion the analytic one steps in and assesses the situation (For instance, the analytic system activates when we are asked to compute 17 x 24).  The analytic system activates when there are events that go against our worldviews, because these are the reference points used by the intuitive system. So when events don’t go the way they should or the way we expected and we activate system 2 to figure out what is going on. Moreover this depend on our emotions: if we are in a bad mood we are more analytical than when we are happy (we are more intuitive)  Further, the analytic system is in charge of our behavior and regulates self-control when, for instance, we are about to offend someone during a discussion  The analytic system activates only if we have energy and we have to care to intervene  The interaction between the two systems is very efficient because it allows to minimize the effort and optimize decision quality. The intuitive system is generally able to:  Have a good representation of familiar situations.  Can make short time predictions.  However, it is also prone to errors and biases CONFLICT BETWEEN THE TWO SYSTEMS  Stroop effect: usually we do not care about the color used to write a word, because we automatically focus on its meaning. As a result, when we are asked to report the color instead of the actual word we experience a conflict.  The meaning of the word is elaborated automatically and interferes with the task we are asked to complete (reporting the color).  The analytic system has to inhibit the intuitive response and this makes the task more difficult (and our responses become slower).  This is what happens when we see a situation that seems familiar but we are not considering all facts (monkey illusion)  The conflict between intuitions and analytic elaboration of information is quite frequent in our everyday life.  When we drive on snow, if the car starts to go sideways we need to be cool, steer on the opposite direction, and inhibit the temptation of hitting the breaks.  When we talk to someone we rely on the analytic system to inhibit the intuitive reaction to offend them or reveal something that we were asked not to say. COGNITIVE ILLUSIONS 15  Even when we know a preceptive illusion we have troubles not being affected by it and have to remember what the reality is.  We may know there is no white triangle on top, but still it looks like it is there!  These examples deal with perceptive illusions. However, there are also cognitive illusions, which apply to our worldviews, opinions, and beliefs rather than perceptions of the physical world.  We are so convinced on what we believe that we cannot understand that others can have other ideas. Moreover when we search information about something we do that in a confirmative way that make us believe even more the idea. That makes harder to discuss about things  Kahneman’s example of the psychotic patient that praises the therapist to have a special treatment and then breaks the relationship. After the terapist feels bad because it was attached to the patient.  These illusions are very difficult to prevent since the analytic system does not recognize the mistake made by the intuitive system. Even when there are “hints” that we could be wrong, an error can only be avoided through a careful monitoring of the situation.  However, it is very hard to think carefully about every single thought or decision we face.  The analytic system is too slow to substitute intuition in our routine behaviors.  We must learn to recognize the conditions in which we risk making a mistake in order to focus and pay attention when it matters DUAL-PROCESS, SELF-CONTROL, EFFORT  The analytic system allows us to direct our attention towards what we are interested in while completing a task or making a decision.  This makes cognitive tasks an effort that consumes our mental (and physical) resources...  The same is true for self-control, as it is easier to exert control when we are relaxed rather than tired.  Example: Imagine you should memorize the following string of numbers: 1893572. While you reiterate the string in mind, I ask you to choose one of the following desserts: Chocolate cake or fruit salad. Since system 2 is busy doing the task, the system 1 decide the chocolate cake, choosing the option they feel more than the healthier. System 2 is busy and cannot exert self-control (so choose the fruit salad)  When people have made a cognitive effort, they are more likely to: 1. Choose in an hedonic way. 2. Be egoistic. 3. Make sexists or stereotypical remarks. 4. Make superficial judgments during social interactions.  Worrying about our performance in a specific task decreases the quality of the outcome because anxiety and thoughts interfere with working memory.  Baumeister and colleagues showed that people can easily spend their finite reserve of self-control. When it happens, the quality of their decisions worsens significantly.  Every time people make a cognitive effort, they become less able to exert the same effort in a subsequent activity (ego depletion): People who are asked to suppress their emotions while watching an affect loaded videoclip obtain a worse performance in a subsequent task involving physical effort.  When we are tired is difficult to do something we don’t want to do  All activities that involve the analytic system also require self-control and maintaining self-control is difficult and not too pleasant, therefore this opens the door to the intuitive system to step in and lead the action. SELF-CONTROL  Activities that require self-control: 1. Inhibit one’s own emotional reactions. 2. Make decisions when there is a conflict. 3. Trying to impress other people. 4. Being nice when someone is treating us unfairly. 16  Behaviors that depend on a lack of self-control: 1. Over-eating while on a diet. 2. Impulsive shopping. 3. Being aggressive when someone provokes us. 4. Obtaining unsatisfactory results in cognitive and logic tests.  A baseball and a ball cost $1.10 in total. The bat costs $1.00 more than the ball. How much does the ball cost? Most people answers that the ball cost 10 cents, but is wrong, it costs 5 cents. If you have self control, you try to find the right answer and you use the analytic system, but if you trust your intuitive system, maybe because you’re tired or you don’t care, you fail.  Mischel & Ebbesen, 1970: under certain conditions, a child’s success in delaying the gratification of eating marshmallows or a similar treat was related to later cognitive and social development, health, and even brain structure.  Feelings guides our decisions a lot and we can be impulsive when the emotions are too strong to resist them (es. marshmallow)  Il doesn’t replicate so well, especially the second part of the experiment (when they were adults). That because they track all the original kids (small sample), moreover kids not from the middle-up class didn’t resist because from their experience they never know if will be another occasion like that. INTUITION AND ASSOCIATIONS  When we think about something, automatically we associate this to other thoughts, that connect and activate other different ideas with less force.  The reaction we intuitively have toward something will determine our behaviour. PRIMING  We can see this when a stimulus is presented to people but this doesn’t arrive to the consciousness  This mental process can lead people to activate specific associations or to experience a certain emotional reaction. When this process happens without people being aware of it, we call it priming. The priming effect can be induced in many different ways: word, image, video or more complex tasks. But more complex is the situation where we present the priming, more difficult is for it to work. PRIMING IN REAL LIFE  It is easier to find priming in the lab, but it’s more difficult to find it in real life because of the big amount of stimulus there is.  We can assist to priming effects even outside the lab:  Vote on public schools in the U.S.: people who voted in the school voted to give more founds to the school than that is other places. The place they were in was consisted with what they were voting for  Images of God or images of a leader (e.g., Mao in China): make prime giving an authoritarian image or respected that make us behave in different ways  Of course, there are several limits that make priming effects outside the lab much less effective than in the lab. Nevertheless, each of us can be influenced by events without being aware of it. OTHER FEATURES OF THE INTUITIVE SYSTEM  The intuitive system can create “fictitious” causal relations: for example thanks to its associative functioning, it is very effective at filling informative gaps and create coherent stories.  On many occasions, our need to find reasonable explanations leads the analytic system to trust excessively the conclusions produced by the associations upon which the intuitive system relies.  The intuitive system can create “fictitious” causal relations, because is really difficult to make prediction of the future: Nassim Taleb wrote about a financial example related to the price of U.S. bonds on the day Saddam Hussein was captured.  In the morning, prices were going up and Bloomberg used this headline: “U.S. bonds go up; Hussein’s capture did not reduce fear of terrorism”. So people were buying safer than stocks. 17  2 decks are “advantageous” (the gains they offer are not so great, but losses, when they occur, are quite small). In the IGT, people are asked to select 100 cards from any of the four decks.  Usually, after sampling about 50 cards (mostly from disadvantageous decks), participants start to select cards mostly from the “advantageous” decks. However, they are unable to tell explicitly which strategy is the best (which decks are offering the best results).  After drawing about 80 cards, participants are able to explicitly explain which decks are advantageous and why.  Bechara and colleagues recorded skin conductance and heart beats finding that, as the task progresses, people are increasingly showing stress reactions when they are about to select a card from one of the “disadvantageous” decks.  At some point they implicitly learn from the reaction from their body that something is not right in decks A and B, and choose C and D decks. So emotions are important because they can communicate us the consequences of our behaviour and from these we can make decisions.  People with lesion at frontal cortex (areas about decision making – as also using emotion to decide), because they are noy able to use their emotion to judge, they will choose more the disadvantageous decks and loose money. But because of their lesions, they don’t understand this.  People that are better in recognising their emotion are better also in labelling the situation using them and in regulating them  In real life we have these amount of choices in more long time, we don’t have so much chances to learn, the context changes and the feedback is delayed in time. So, in real life is difficult to have implicit leaning, usually we use other strategies  The affect heuristic shows us why people perceive risk and benefit as negatively correlated (although in real world the correlation is often positive): when we are exposed to a stimulus, in an automatic way we associate to this stimulus something that is positive or negative. It is a feeling that automatically and with low intensity influence how we process the stimulus, making us like it and approach it or avoid it.  When we see something, we associated it with positive or negative things and consequently creates our emotions and our opinion about the stimulus.  Kahneman show that in risk perception there is a tendency for people to have negative relationship with benefit and risk of technology (same as stocks).  How does it work? As we see in investing stocks has high benefits, so they expect to gain a lot of money, they will also expect to be a lower risk because the benefit influence positive emotions. If I see that stocks are risky, this will lead a negative feeling about not gaining money. So, as in reality people underestimate the chance in gaining money in stocks and are not investing in them.  Putting before benefits then the risks about something: the order effect will make people focus on the first information and create a good impression of the thing they are evaluating because they will read the other information keeping in mind this information and the associated net. PROSPECT THEORY  It was a response to the Rational theory, the way we know that works in psychology  It tells us how people make decisions in real life  Economic and utility theory became the norm of the rational decision making. It is not how people behave  Prospect theory is a descriptive theory and underlies the effects that influence our decision. PROSPECT THEORY: A DESCRIPTIVE MODEL OF DECISION-MAKING  Starting from work on the violation of rational axioms, Kahneman and Tversky (1979; 1992) developed a model to describe more accurately how real people make decisions: the Prospect Theory.  Prospect Theory is not in opposition to the Theory of Expected Utility, but aims at integrating and extending it. 20  Expected Utility Theory provided a model about how people should make the best possible decision (normative model). Prospect Theory provides a model about the processes that induce people to make suboptimal decisions (descriptive model).  Prospect is a gamble: from a formal point of view is a probability and an outcome attached to it, but it something we create in our mind. It is our perception mediated by our brain  In Kahneman and Tversky’s (1979) model, Expected Utility becomes the benchmark model to assess the quality of choices made by people in several domains (economics, health, etc.). In choices under uncertainty, people have a tendency to simplify the decision as much as possible in order to save cognitive energies.  In other words, cognitive limitations (in terms of memory or attention) make it difficult to execute the complex computations required to obtain a measure of expected utility.  It doesn’t capture the variability of the people, so describe the behaviour of most of the people.  It is descriptive also because tests what people do in specific situations: first they observe the behaviour and then they explain it and create axioms. Rational theory does the opposite. WHAT IS A PROSPECT?  The label «prospect» substitutes the economic term lottery and aims at conveying a stronger meaning of subjectivity regarding how alternatives are represented in the decision maker’s mind.  A prospect is the combination of all possible outcomes of an alternative and their respective probabilities.  Prospect X: (x1, p1; ... ; xn, pn) dove p1 + ... + pn = 1 o X is outcome and p is the probability  To simplify, we can omit null outcomes and use the form (x, p) to indicate the following prospect (x, p; 0, 1 - p) that offers an outcome x with likelihood p and an outcome 0 with probability 1 - p.  We can compute the expected value, but we prefer save mental energy. Prospect theory tries to understand why we don’t compute.  A riskless prospect, which offers to obtain the outcome x for sure, is represented with the form (x). THE TWO STAGES OF THE DECISION PROCESS  Prospect Theory states that decisions are made in two phases: 1. «Editing» phase: This is a preliminary assessment of alternatives and leads to a simplified representation of the available prospects. 2. «Evaluation» phase: The simplified prospects, as they emerge from the editing phase, are evaluated and the prospect with the highest value is chosen. THE «EDITING» PHASE  The preliminary analysis of the prospects is made, for the most part, at an unconscious level and uses a wide array of mental operations. The most relevant editing operations are the following: 1. Coding. 2. Combination. 3. Segregation. 4. Cancellation. 5. Simplification. 6. Detection of dominance  The prospects in our mind become something else, that may be not recognisable after our transformations, so our perception is how we constructed the prospect in our mind. The prospect after editing operation is different, and that’s way we have a subjective perception of the prospect.  We apply these operation in different order, and the order we used them opens the door to make another type of transformation. So the information is subjective and leads to different decisions. CANCELLATION  When something is common to all the options we don’t consider that.  To simplify the choice among several prospect the decision maker could (mentally) cancel, or delete, the components that are common to all prospects:  Example, if the choice is among the following prospects: 21  Prospect A: (€1000, .25; - €100, .75)  Prospect B: (€1000, .10; €500, .40; - €200, .50)  Someone could say that 10% of probability of winning €1000 is very low, so they don’t consider it in they prospect computation and decision They could be simplified in the following way:  Prospect A’: (€1000, .15; - €100, .75)  Prospect B’: (€500, .40; - €200, .50) SIMPLIFICATION  Often the decision maker can simplify the elements of a prospect in order to make it easier to evaluate.  For instance, he/she could round the values of outcomes and probabilities:  Prospect X: (€199, .50; - €201, .49; - €100, .01) Could be simplified in this way:  Prospect X’: (€200, .50; - €200, .50)  However, although such operation makes the prospect simpler, it also changes its overall value.  A prospect that was originally slightly negative, is now neutral. EDITING PHASE  These operations are applied to prospects without following a planned order, but depend on the information each decision maker is more likely to focus on. This is an issue when we want to predict people’s evaluation (and their subsequent choices), because:  The use of a specific simplifying strategy can undermine the use of another one.  As a consequence, the order with which these strategies are used becomes fundamental to understand the final decision.  Many intuitive strategies that we will cover in future lectures (i.e., heuristics) arise from the use of simplifying strategies during the editing phase. From this point of view, the editing phase is a central stage of the decision  process. THE EVALUATION PHASE  During the evaluation phase the simplified versions of the prospects are compared and the final choice is made. The evaluation phase relies on two functions that people use to judge, subjectively, the value of outcomes and their likelihood. 1. Weighing function (applies to likelihoods). 2. Value function (applies to outcomes).  Both functions tell us which is the subjective value we give to probabilities and to outcomes WEIGHING FUNCTION  The weighing function casts light on two important features of how people judge probabilities:  Low probabilities are usually overestimated (we perceived a big difference between 0% and 2%: something is better than nothing).  Medium and high probabilities are usually underestimated (certainty effect: we perceived a big drop from 100% to 98%)  So we only perceived right 0% and 100%of probability  The certainty effect explains why we don’t pay insurance: we don’t think we will make an incident  As a consequence, very unlikely outcomes are overestimated compared to having no chance to obtain them.  In contrast, very likely outcomes are underestimated compared to when they can be obtained for sure.  This explains the certainty effect and the example presented while discussing the violation of the independence axiom.  Certainty effect: o Going from a sure win to a very likely win (e.g., 98%) makes that prospect significantly less attractive (high probabilities are underestimated). 22 b) 50% chance of gaining €200 or 50% chance of gaining nothing  Results: Even if expected value is the same (100€), most people will choose a because they don’t want to risk. Important is the sentences because it gives a reference point. Now, imagine you are €500 richer than you actually are, then choose one of the following alternatives: c) Losing €100 for sure. d) 50% chance of losing €200 or 50% of losing nothing.  People don’t consider the information about how much richer they are (€500), they only look at the alternatives and in this case they prefer taking a risk (choosing the gamble).  But even if the reference point is different, options a and c give the same utility value: 400€. So a rational person, even without seeing both problems, will choose option a and c, because the end result is the same (having 400€ is better than having 300€).  But we don’t look to the alternatives, we choose based on loss aversion. As we can see, people’s choices are inconsistent, therefore irrational from an economic standpoint, because the outcome are the same in both versions of the problem:  A (€300 + €100 = €400) is the same as C (€500 - €100 = €400).  B (€500 or €300) is the same as D (€300 or €500).  Tversky & Kahneman (1986): During a financial crisis, a company is in economic trouble. The company is based in an area that has been hit hard by the crisis and there is very high unemployment, with an inflation of 12%. For this year, the company decided that they will rise the employees salary by 5%. How acceptable do you judge this decision?  Money illusion: people makes exchange in nominal terms rather than considering the exchange rate. When we think in a increase in our salary, we don’t see that this increasing is lower that the one of the inflation.  It is not very fair but people thinks it is During a financial crisis, a company is in economic trouble. The company is based in an area that has been hit hard by the crisis and there is very high unemployment, but zero inflation. For this year, the company decided that they will cut the employees salary by 7%. How acceptable do you judge this decision?  In both cases there is a 7% gap (we lose purchasing power), but in the second case we don’t see this, we see only the cut from the salary. In the first case we see an increasing salary (not enough to help to fight inflation though) and it is interpreted more positively from people because they see only the nominative value.  A consumer behavior study tested the framing effect using products labels such as “% lean” versus “% fat”. Participants were asked to assess and judge a piece of meat. There were two different experimental conditions: 1. Condition 1 «only label»: Half of participants in this condition were told the meat was “75% lean”, while the other half was told that the meat was “25% fat”. 2. Condition 2 «label + tasting»: Participants were given either one or the other information and where also allowed to taste the meat All participants judged the meat on four different dimensions using 0-7 scales:  Bad taste vs. good taste.  Oily vs. not oily.  Low quality vs. high quality.  Fat vs. lean Results: even when they taste the mean the difference between the two labels are small, but the labels influenced the judgements. The label fat made believe the meat fatter, lower in quality, more olly and with a worse taste. This is why on the packaging of food they put only the percentage they want to attire more people.  Levin, Schneider and Gaeth (1998) described three different types of framing: 25 1. Risky choice frames: This is the case in which different descriptions of the same problem change people’s risk preferences (like in the gambles problem or in the Asian disease problem). a. This type of framing is measured through people’s choices. There are any differences when outcomes are described as gains vs. losses? 2. Attribute framing: This is a case in which a dimension of a stimulus is expressed along complementary positive versus negative terms/labels (the meat experiment). a. This type of framing is measured through evaluations (judgments) of an alternative described positively (% lean) or negatively (% fat) with a rating scale. 3. Goal framing: This framing aims at changing people’s behavior by using messages that maximize the likelihood of achieving the goal (for instance, describing the risk associated with smoking versus the benefit associated with not smoking) a. When we give recommendation about a particular behaviour (example: sun cream) b. Since the aim is to actually change people’s behavior, researchers will compare the amount of people who change their behavior depending on the message they received (how many individuals stop smoking after learning about the risk of smoking vs. the benefit of not smoking) ENDOWMENT EFFECT  The endowment effect has been introduced and studied in depth by Knetsch, Kahneman and Thaler (1989; 1990; 1991; 2001). This bias concerns the asymmetry that exist between willingness to pay (WTP) for a product and willingness to accept (WTA) to sell the same product. Usually:  Sellers set a price that is higher than the price buyers are willing to pay.  It is also a great strategy to negotiate (you ask more to sell the objects at the price you want), in fact this bias is well-know to negotiators  This asymmetry arises because those who own the object attach a sense of ownership that leads them to overestimate the value of the object compared to those who do not own it.  It is a reaction driven by loss aversion since people do not want to give away (lose) the object.  It is only driven by the ownership of the object  These researchers have shown that this bias can be found even when the object to sell has been given to the seller in that very moment (no ownership).  A direct experience or the memory of how (when) the object was used is not required for the endowment effect to appear.  The coffee mugs studies (Kahneman, Knetsch, & Thaler, 1990). The goal of a first study was to “neutralize the reasons that may influence the experience of ownership.” This way, researchers could minimize the relation between the value of ownership and the affective reactions. The authors created three different conditions (between-subjects) in which people were asked to express their preference for either a pack of candies or a coffee mug: 1. Group 1: These participants received the coffee mug and were asked whether they wanted to exchange it for the pack of candies. 2. Group 2: These participants received the pack of candies and were asked whether they wanted to exchange it for the coffee mug. 3. Group 3: These participants did not receive neither the coffee mug nor the pack of candies and were asked to choose one or the other Results: 1. Group 1: 89% of participants in this condition decided to keep the coffee mug rather than exchange it for the candies. 2. Group 2: 90% of participants in this condition decided to keep the candies rather than exchange them for the coffee mug. 3. Group 3: These participants did not show any clear preference for one product or the other (56% chose the mug; 44% chose the candies) 26 o The explanation is that when something is told is yours, we have difficulties to give it away. This does not depend on the type of object o It is also connected to the status quo bias: you stay in the situation you are.  In a second study, Kahneman and colleagues (1990) created two different conditions, assigning participants to one of two roles: 1. Sellers received the mug once arrived in the lab and were asked for the lowest price they would be willing to accept to sell the mug (WTA). 2. Buyers were asked for the highest price they would be willing to spend to purchase the mug (WTP). Results: 1. Sellers were not willing to go below an average price of $5.25. 2. Buyers were not willing to purchase at a price higher than $2.75 o So it was very unlikely to find someone ready to buy: the exchange would happen only if the buying price would be higher than the selling price. Usually the price exchange was find in the middle. o Sellers overvalue the mug, and the buyers undervalue it. STATUS QUO BIAS  When there is uncertainty about the possible outcomes, people can choose “not to choose” and simply keep things the way they already are. This happens for different reasons: 1. Inertia and procrastination: changing things need an action that can require effort, time, or even money, with no certainty regarding the results. 2. Uncertainty: changing obliges the decision maker to face the risk of obtaining an outcome that is not as good as she wanted or hoped for. o In investment people don’t sell because there’s the chance to gain more another day. 3. Inexperience: those who do not have enough experience often resist to changing a familiar, albeit unsatisfactory situation, because they are scared to make a mistake. o But sometimes we are only overestimate the chance to commit a mistake.  Car insurances in New Jersey and Pennsylvania (Johnson & Hershey, 1993). In the early ’90 the laws on car insurance changed in both states. To reduce cost a new policy was introduced that excluded coverage for theft and fire. However:  In New Jersey, new drivers had to buy the new policy and could then choose to upgrade it to the full coverage one (opt-in solution).  In Pennsylvania, new drivers had to buy the full coverage policy and could downgrade to the new, cheaper policy (opt-out solution). An analysis of drivers’ decisions in the two states showed that the status quo option had a significant effect on the insurance they ended up with:  Only 20% of drivers in New Jersey chose to upgrade to the full coverage policy (for these drivers the status quo was the cheaper policy).  In contrast, almost 75% of drivers in Pennsylvania chose to keep the full coverage policy (it was the status quo for them)  In both the states people stayed where the law makers put them (the coverage where they were). It is very likely that participants have inertia, didn’t want to make the effort, and that how had the standard package didn’t want to pay an additional cost (they thought they would lose something). The actual cost in Pennsylvania and the future cost in New Jersey would be the same, but the participants are feeling different.  In Pennsylvania people are undervaluing the cost (there is benefit in the money I am spending) and in New Jersey people is overvaluing the cost (too much money for a potential benefit). So in one case people think about what they will lose and in the second about what they will pay. o The value function: the chance to gain is not as bad as an actual loss, so losing was worse for the Pennsylvania drivers than the gain for the New Jersey drivers. 27  They are an intuition  The term comes from technology, from the time in which they tried to doing analogy with computers. REPRESENTATIVENESS HEURISTIC  The representativeness heuristic is used to answer questions like: What is the likelihood that object A belongs to class B? What is the likelihood that the event A originates from process B?  Likelihoods are assessed on the basis of how much A is representative of B. If A is highly representative of B then the likelihood that A originated from B is judged as high  People make mistakes when they judge whether a particular event belongs to a specific process or category.  Stereotypes are an example of representativeness heuristic  In other words, likelihood judgments are made based on how much an event is representative of the process being evaluated. In doing these evaluations people end up NEGLECTING sample size or the laws of probability: so we apply the law of probability also to the small samples. But as we know from statistics we need a big sample to have reliable results and representative of the population.  For a large part, research on this heuristic was made by Kahneman and Tversky who were at the forefront in studying the Theories of Judgments Distortion  Example: Imagine we flip a (fair) coin for six times. Which off the following sequences of outcomes is more likely? T T T H H H or T H H T H T  Results: people think the second one is more likely because is more representative of random events, but in reality they are independent probabilities so they are both equally likely.  When the series of events (and outcomes) is so short the likelihood of any sequence of coin tosses is going to be equally likely, but we don’t consider that the sequence al too short to make an evaluation (= gambler’s fallacy)  The representativeness heuristic is the direct cause of many errors made by gamblers. Such errors led to the definition of “Gambler’s fallacy”.  Gamblers who are playing at the roulette think that: “If red was the outcome for several times in a row (for instance, 4) then it is sure that on the next round the outcome will be black”. The same applies to the lotto game, some people believe that if a number is not drawn for a long time, it becomes more likely than a number who was extracted recently.  Actually, in both cases (coin toss and lotto), we are facing a probabilistic game with replacement: at every coin toss (or lotto drawing) any result has the same chance to happen as in the previous round SAMPLE SIZE  Law of the small numbers = we apply the law of large numbers also to small numbers. When we have small sample we think we can use the same rule of large numbers but this doesn’t apply because the sample is too small (one outlier can make everything different)  The previous examples show that people believe that a SAMPLE DRAWN FROM THE GENERAL POPULATION. Should have the same characteristics as the population itself.  However, a big issue in these cases was sample size (how large is the sample?). When a coin is tossed 5 times, it is entirely possible to get always the same outcome (for instance, always heads)... Things would be different if we toss the coin for 100 times. In this case, the sequence of outcomes will start to even out and the number of Heads and Tails will be much closer.  IN OTHER WORDS, WE HAVE A TENDENCY TO APPLY THE LAW OF LARGE NUMBERS EVEN TO SMALL SIZE DISTRIBUTIONS!!  The representativeness heuristic leads people to neglect base rates when they are judging the likelihood of an event. So we value small number as larger numbers  The BASE RATE is the frequency of an event within a specific population. 30  Example from the work of Kahneman e Tversky (1974): In a town there are two hospitals. In the bigger hospital 45 babies born every day, in the smaller hospital 15 babies born every day. As you know, on average half of the newborns are males and the other half are females; however, the percentage of males and females who are born in each specific day is not always the same. For an entire year, each hospital kept a record of the days in which the number of male newborns was higher than 60% of all babies. In which hospital, did the number of males was above 60% more often?  Of course, the correct answer is “the smaller hospital.”  People says that is equally likely to find more than the 60% of male babies, but in reality is more likely in the small hospital, because the smaller is the number of children that is born and the higher is the probability to go over the 60%. In small samples just a few outliers (days in which are much more males) can influence the result. But people don’t consider the base rate (how big is the sample)  Small samples are less representative of the population compared to large samples and it is more likely that observations based on small samples differ from what actually happens in the general population.  Regression to the mean: Extreme events are likely to be closer to the population average when measured again. Extreme events tend to converge to the mean gradually when we measure something.  Air force fighter jet pilots training: teachers saw that after a prize the future pilot made worse and after they were criticize, they made better. In reality the fact is that who is not an expert will have very fluctuating performances. So after a extreme performance, the after one will be nearer to the average.  School tests.  People don’t understand how poor is the reliability of a small sample. CONJUNCTION FALLACY  Linda is 31 old, single, outspoken, and very bright. She majored in philosophy. As a student she was deeply concerned with issues of discrimination and social justice, and also participated in anti-nuclear demonstrations.  Linda is a teacher in an elementary school.  Linda works in a library and takes yoga lessons.  Linda works as an assistant at the psychiatry division.  Linda is a member of the League of Women Voters.  Linda is a bank teller. (B)  Linda is an insurance seller.  Linda is a bank teller and she is active in the feminist movement. (B&F)  So people tell the last option is the more likely based on the description, but it is impossible that a conjunction has more probability than the singular and bigger thing, that includes the other.  The likelihood of the conjunction is smaller than the bigger event. The description doesn’t matter, you need only see the base rate.  Tom W example and hiring people example. AVAILABILITY HEURISTIC  It answers to what is the likelihood that something will happen, recalling in the mind similar experiences. These examples permit to construct the probability: more easily the examples (less are enough) came to mind larger will be the likelihood. Of course if you experience something, you will remember it more easily and you will have an idea of how much frequent it is.  Availability heuristic is not always based on how often the event happens (likeness) and from the experience, but how vivid and emotional your memory is and the association you make.  More an event is recent and easily to recall, more vividly we remember and make the associations that lead to the likelihood.  We make perceptual errors, so same events are easier to remember because of the images that we have in memory (as dramatic images) 31  This is the way terrorism works  People rely on this heuristic when they need to make judgments regarding the likelihood of future events. In making this type of judgments they use the information that comes more easily to their minds.  However, our memory:  Is not perfect.  Does not work by employing statistical principles.  Is influenced by psychological factors like our cognitive limitations.  The psychological processes behind the availability heuristic are particularly powerful and pervasive. Some of them are:  Likeness-driven associations.  Perceptual errors (some events are easier to remember and remain impressed in our memory).  Memory is selective.  When judging the likelihood of an event people try to remember or mentally generate examples that can be used as useful hints. Since this a psychological phenomenon, we already know it won’t follow precisely the rules of randomness (no normal distribution).  For instance, vivid events look more frequent than they actually are: think about terrorist events or people who hurt you.  The examples that come to our mind are not like objective information, but a “psychological sampling” that we use to make inferences regarding the likelihood of an event. Usually, the examples that come more easily to our mind are those:  Vivid.  Linked to our previous experiences.  Most recent  EXAMPLE: It is well know that airplane accidents are judged as more frequent that what they actually are. On the contrary, the frequency of car accidents is usually underestimated: thinking of how many flights there are everyday and how many incidents there are. A big role is played also by the news  EXAMPLE 1 from Tversky and Kahneman: People were asked whether in English there are more:  Words that start with an “R”.  Words with an “R” as the third letter The majority of participants answered that, in English, there are more words that start with an “R” - WRONG ANSWER!  Estimates were influenced by how likely it was to come up with words that start with an “R” and words that have an “R” as the third letter.  It’s easier (more fluent) to recollect words that have the “R” as the first letter that the ones that has the R in other positions, but in reality, there are more combination that have inside the R.  EXAMPLE 2 from an experiment by Tversky and Kahneman: This study shows the tendency to judge more frequent stimuli that are more relevant or familiar to us. Tversky and Kahneman presented participants with two lists of names of famous people who were active in the show business. 1. LIST 1: 19 names of famous women and 20 name of well-know men. 2. LIST 2: 20 names of well-known women and 19 names of famous men.  A WELL-KNOW person could be someone who is familiar but not so much to be known by everyone.  A FAMOUS person, instead, is someone that everyone knows and recognizes  Results:  In the first condition, participants judged the number of (famous) women in the list to be higher than the number of (well-known) men. Being more familiar, the names of the women in LIST 1 were easier to remember.  In the second condition, participants judged the number of (famous) men in the list to be higher that the number of (well-known) women. 32  REAL LIFE EXAMPLE: Anchoring effect in the real estate market. Both Business School students at the University of Arizona and actual real estate agente from Tucson (AZ) were asked to estimate the value of a house based on a set of information provided by the researchers:  Real estate trend in that neighborhood of Tucson (referred to properties similar to the one they were evaluating).  Information regarding the property they were asked to evaluate (area, rooms, price it was listed for, etc.). Both students and real estate agents were allowed to: Visit the house and other properties in the neighborhood. They were also divided in two groups and each group received only one of these two prices:  Group 1: House listed in the market for $65.900.  Group 2: House listed in the market for $83.900. These were the estimates provided by participants in the two conditions:  The first group gave an average estimate of $63.571.  The second group gave an average estimate of $72.196. No statistical difference was found between the estimates made by the business school students and those made by the real estate agents. SCARCITY MONETARY SCARCITY  Economic: Scarcity is everywhere, everyone have a limited amount of money.  objective perception: You objectively have fewer resources than needed in these contexts.  Psychological: Scarcity is not just a material limitation.  subjective perception: You subjectively perceive having fewer resources than needed (or wanted) in these situations.  Individuals throughout the income spectrum can feel as if they “do not have enough”.  A scarcity mindset is thus not triggered only by an objective lack of resources. If that was the case, a scarcity mindset would be observed only in poor individuals. FIRST STUDY ON SCARCITY  After the Second World War, a group of researchers from the Minnesota University (Keys, Brožek, Henschel, Mickelsen, & Taylor, 1950) investigated how to reintroduce food in a population that experienced a condition of famine during the war.  32 subjects for 1 year.  Physiological and psychological aspects constantly monitored.  The experiment was focus on the denutritional aspects than the psychological ones. Participants were always thinking about food, going beyond the practical benefits of these thoughts.  “It was impossible not to think about food in every possible declination and this was not intentional”  It was the famine (or the perception of not having enough food) that caught their attention, in particular what was missing in their food routine.  It addresses to how scarcity capture people’s mind WHEN SCARCITY CATCH YOUR MIND  Can you stop your mind from thinking about the deadline when you are getting closer and closer? You are worried about the deadline, you cannot stop thinking about that and you are less focused on other problems.  It has positive impact because it helps you to focus on what you have to do and to not distract. FOCUS DIVIDEND  But at the same time you forget about everything else so if you are doing something else in the same time it is detrimental. So you are focused on what is important, but if you have to think also to something else, it is detrimental. 35  Example: If you need to organize it and select some ideas, scarcity could help you. But if you need to improve your project, scarcity is a problem. TUNNEL EFFECT  Focusing on what seems to matter the most could be considered as positive, but however, letting scarcity to cause the tunnel effect and to bring us to neglect other aspects cannot be considered that advantageous. (Mani et al., 2013; Mullainathan & Shafir, 2013, 2014).  The power of staying focus, it is also the power of exclude something.  It depends on the fact that the activation of something in our mind, reduces the possibility of recalling other concepts that so are inhibited.  Tunnel effect: Scarcity leads people to concentrate on something that is considered important and to neglect all the rest.  It is not to convenient when we have to consider also other aspects of our life.  we reduce the energies and resources destinated to other things ATTENTIONAL BANDWIDTH  The human cognitive system has limited capacity (Baddeley, 2012): For instance, people cannot pay attention to a large volume of information.  Our attentional capacity is determined by both structural factors, such as the limits of our working memory (Cowan, 2010; Logie, 2011), and the processing speed of neurons (Bays,2015).  Given this limited capacity, engaging in one process consumes cognitive resources needed for another, thus causing interference (Tomm & Zhao, 2016).  There is a sort of interference between what we are doing and what can potentially do.  When people are paying attention to and processing information (Meuris & Leana, 2015; Chugh & Bazerman, 2007), allocation of cognitive resources to one issue, inevitably diminish the amount of cognitive capacity left for others, which can result in superficial evaluation and/or neglect it.  Poor people haven’t a less effective bandwidth, it is a situation that everyone lives.  The problem of scarcity is also connected to a reduction of the mental sources that people can use in their daily basis.  The perception of not having enough resources leads to the tunnel effect, and this consumes the mental bandwidth and attention available for making decision that are usually used for our daily life. The result are impaired functions and capacity. IMPAIRED FUNCTIONS  Parts of the reasoning that impairs  Fluid intelligence: fundamental to solve problems, retain information and engage in logical reasoning  Executive control: essential to plan, attend, initiate, or control impulses  Consequences of scarsity:  Prospective memory (it is related to remembering to perform a planned action or recall a planned intention)  Working memory  less efficient  Lack of self control  Sleep reduction  Negative affect and stress  Increase in cortisol levels in the blood  Higher risk aversion 36  Exacerbates time discounting  Imagine an hypothetical situation in which people were stopped in a store and have just been told that repairs on their car would cost $1500 (or $150). Participants were given 2 tests of fluid intelligence: Raven’s matrices and a second test in which was measured their ability to think and act fast. Results: the less well-off participants scored much worse if they were facing the higher bill.  Participants in the lower income condition scored more when they were told they had to pay the higher bill. IMPAIRED FUNCTIONS IN REAL LIFE  Mani (2013) ran an experiment in India, where farmers receive their salary once per year after the harvest.  They supposed that right after the farmers had salary their life was easy to manage, but what happen when they are far from receive the salary and they have to manage the expenses?  they measure money scarcity with a cognitive (performance) and behavioural measure (how many times they asked for a loan or a pawn shop). Imagine that farmers receive the money in June, in the following months it is quite easy to manage everything. But how easy is going to be in April or May? Mani measured farmers’ cognitive performance pre and post sugarcane’s harvest. Results:  Pre harvest: 78% went to pawn shop 99% asked for loan  Post harvest: 4% went to pawn shop 13% asked for loan  Considering performances in fluid intelligence and in executive control, the same farmer had worst performance in the pre harvest condition than in the post harvest condition (+25%)  Participants had a different performance between the pre and the post harvest and that the cognitive test on the fluid intelligence was higher post harvest. WHAT KINDS OF THOUGHTS OCCUPY THE MINDS OF THE POOR?  Concerns about money come to mind spontaneously among the poor. This is related to the fact that they have a objective scarcity (they try to manage money in the best way possible) and they relate to the fact that there is a persistent stress because of the economy situation  Scarcity has a robust effect on the associations that the poor perceive. And those associations, when the relevant concepts are encountered, raise the specter of scarcity in the minds of the poor.  Monetary “cocktail party effect” = they are more sensible to the economical aspect related to the daily life. They have low threshold on perceiving the economic dimension in the daily routine. So they are more likely to entertain monetary consideration even when the situation doesn’t involve money.  Poor individuals have a low threshold for seeing the economic dimension of everyday experience. That is not to say that poor individuals are always thinking only about money. But if an experience has some economic dimension, the poor are more likely to tune to it than are the rich.  Studies on that  More likely to remember what things cost (Mullainathan & Shafir, 2013)  More likely to notice hidden taxes (Goldin & Homonoff, 2013)  Spend their resources more efficiently (Shah et al., 2012)  More likely to consider tradeoffs and opportunity costs (Shah, Shafir, & Mullanathan, 2015; Spiller, 2011)  More likely to conform to certain normative economic principles and are less susceptible to certain context and framing effects than are the wealthy (Shah et al., 2015) THINK ABOUT THE COMPROMISE  To have one thing you must give up something else. Mullainathan and Shafir (2013), asked people in a Boston metro station to list the features that motivate them to buy a television. 1. Size 2. Screen resolution 3. Price 4. Quality 37 MARKETING PERSPECTIVE  People always want what they cannot have and it is hard to obtain.  Scarcity marketing is a marketing strategy that plays on a prospect’s fear of missing out on a purchase (FOMO).  Limited product availability  Exclusivity  Urgency A product that had limited availability is perceived as exclusive and transmit urgent. COVID-19  The initial fear was that there would not be enough resources, and time uncertainty increased the perception of resource-related scarcity.  That leaded to the panic buy. This behaviour is related to reduced anxiety, reduced anticipated regret and social influence (having something that others don’t have)  Scarcity doesn't always increase interest in a specific product, knowing the context is essential. Vaccine scarcity decreased willingness to receive a vaccine, while this effect was moderated by trust in doctors and science. HOW TO ESCAPE FROM THE TRAP?  People experiencing scarcity need an effective plan to break the loop, otherwise it is difficult to take action due to lack of resources.  The more you worry about money, the more likely it is that you will be unable to take the right decisions to escape from poverty. To escape from poverty, it is important to have strong and enduring self control.  ….but unfortunatly is not limitless!  Important is self control  Is it more expensive the limit itself or an additional burden on the bandwidth?  Hang deadline into the tunnel could draw the attention to the limits that we want people to considered, but at the same time we could be taxing attentional bandwidth even more.  Useful deadline (urgent!)  More frequent deadlines  bandwidth as stress  we could move a deadline in smaller deadline that are perceived more urgent to be effective and this help people to focus on the little problem that was solved.  Incentives and programs from the government aimed at people with low bandwidth are often designed to be too costly, precisely in terms of bandwidth. But the fill people had to fill are very hard to do  College & Financial Aid Programs. 3 experimental condition:  Watched the application for grants, so they studied them and they fill by themselves  Specific information with fiscal experts who can support them to fill the forms  Help by the fiscal expert filling out forms o + 29% college enrollment. o Form filled out automatically! Results: the best condition was the last one were people was helped in filling a form that was expensive in a cognitive point of view 54.30  Create periods of moderation, rather than peaks of abundance followed by periods of scarcity. Buffer stocks, created during periods of abundance: this is not easy, it is not spontaneous, because when we are in the moment of abundance we do not expect that these can be affected. Instead of having an up and down 56.40  For people experiencing scarcity, saving is important but not a chore.  Bringing savings into the tunnel  Reminder at the end of each month (text message/letter) + 6% increase in savings  There was no educational or willpower training process, just reminding of something important that can be underestimated. 40  Banks remind clients to save money to reach their monthly goal  they saved 6% more the usual SAVE WITHOUT ACTIVELY DOING IT  Automatic deduction: when there is a tendency not to think about it, it is more effective to change the outcome to which that tendency leads, rather than fighting it.  Keep the change: another bank program where the change from the payment goes automatically in the saving account CHOICES: CONTROL OR ONE-TIME?  To help people to set their choice in the best way to help people 1. Control: Repeating the same decision over and over again 2. One-time: only once (or very infrequently). Under scarcity conditions, it is easier to make the right choice when you make it once. o Advisable to convert behaviors that allow it into one-time choices. o Is related to something that is very infrequent. BANDWIDTH VARIATION  How cognitive resources change? Bandwidth does not remain constant over time, so it would be wise to leverage this time trend when implementing policies and scheduling intervention programs. By understanding the bandwidth timeline, it will be easier to schedule the right times for interventions.  As we see our performance was better after the salary. In one study in Kenya tried to gave the option to buy the fertilazer (that can increase the money from the farm and that they usually buy only when they had no money) only after the salary and when they really needed during the production.  Results: most of the farmers used the fertilazer and understood how useful it was. NUDGING  It is an approach to policy making that maxims effectiveness and reduces costs.  it permits to use psychology in different context to make people behave better.  Based on what we present to the people we can create a context to favour a behaviour. But nudging is not persuasion, because in nudging people is conscious of what we are trying to change in them. IRRATIONAL DECISIONS  People is irrational when decide to do something that is not good for them  Sometimes biases lead people to act in ways that can have very serious consequences: For instance, U.S. workers do not save enough for retirement. Often patients do not adhere to medical therapies or refuse treatments that can help them heal more effectively. COLLECTIVE CONSEQUENCES OF INDIVIDUAL BEHAVIORS  The effects of sub-optimal decisions do not only affect individual lives, but the whole society: for example if people decide not to comply with waste and garbage disposal systems, the whole society will pay the consequences of these actions. HOW TO INTERVENE?  Recent economic approaches based on judgment and decision-making research suggested that we should modify how we look at economic interest in our societies:  Traditionally the only “practical” interest of economics was to gain a profit from consumers’ choices. The gaol has always been to convince consumers to eat a lot, smoke, buy lottery tickets, make debts using their credit cards...  Even companies that apparently want to help consumers (e.g., antinicotine or low-calorie products) are actually pushing for people to smoke or develop obesity, to sell more their products.  The new approaches aim at finding solutions that promote behaviors that favor both:  Individual well-being.  Companies profit and an effective use of public resources  One particular solution is to “induce” people to behave in a more virtuous way.  Such a solution may seem to put two opposite views of public policy in conflict: 41 1. Liberalism (freedom of choice): Each person must be free of choosing which alternative or behavior they prefer (e.g., whether to save for retirement or not). 2. Paternalism (reduction of choice freedom): When people are unable to choose the best alternative it is right to force them (e.g., help savings through a withdrawal from workers salaries).  These two positions seem incompatible. However, they both have pros and cons, and the aim is to mix both. NUDGING (LIBERTARIAN PATERNALISM)  Several authors proposed less radical approaches to paternalism characterized by an attempt to preserve freedom of choice while at the same time using decision biases to induce people to behave virtuously.  They suggest to use our knowledge of psychology and decision making to design more effective decision contexts.  SMART: This is a program to enroll people into pension funds in the U.S. It was developed by Benartzi and Thaler (2002) and it is based on the principles of nudging.  These approaches have been labeled in various ways such as: nudging; libertarian paternalism; light paternalism; asymmetric paternalism. EXAMPLE OF NUDGE  RunCadence app:  This up is used to increase runners’ cadence (number of steps per minute).  Once a specific cadence is set up, the app starts a metronome. The runner must try to hit the ground at every ring of the metronome. The sound of the metronome is particularly annoying, but if the runner hits the right cadence it stops ringing.  This way there is an incentive to achieve the right cadence to make the sound stop and the feedback is easy to understand.  There are various strategies to change behaviour, sometimes also changing the environment POSSIBLE INTERVENTIONS  By taking advantage of decision biases it is possible to design public and economic polices that leave people free to choose how to act while at the same time increasing the likelihood that they behave in the most advantageous way.  People is free to decide, but also we try to influence them  Several psychological phenomena can be used to achieve this goal: 1. Defaults rules that don’t require our intervention (automatic enrollment) 2. Simplification of the process to follow to do something (reduction in the complexity of the decision context) 3. Social norms and social comparison (highlight others’ behavior) 4. Convenience (offering low-cost solutions or making healthy alternatives more visible) o In this sense interest is boosting: make people conscious of the situations and alternative to make a responsible choice 5. Disclosure (making the cost of a behavior explicit) 6. Notifications and graphic solutions (like the images on cigarette packages) 7. Pre-commitments (asking people to commit to a specific program) 8. Reminders (text messages or emails reminding people to pay bills or other expenses) 9. Behavioral intentions (messages to increase voters turn-out) 10. Feedback about the outcome of past choices (an household’s past electric expenses)  The countries develop their unit to study and use nudging to govern. NUDGE IS AN EVIDENCE-BASED APPROACH  Evidence-based policy (EBP) offers a systematic form of validation for decision-making in leadership (Ruggeri et al., 2021). So it is larger than nudging and it is based on promoting policies that are based on scientific reliable evidence. This is used as justification to improve behaviour 42 DEFAULT AND PENSION FUNDS: THE U.S. DEFAULT RULE  After a series of field tests, in 2007, the American government decided to modify the enrollment rules for pension funds and introduced an opt-out program: When an employee was hired a percentage of his/her salary would be withdrawn and placed in a pension fund (401k plan).  Workers who do not wish to participate in the pension fund scheme must make an explicit decision to leave the program (Opt-out). o Furthermore, the percentage of salary paid to the pension fund was also set by default. o As the salary increases, the percentage withdrawn increases automatically (at least up to a point)  You contribute to the plan in default and in advance you can decide how much you want to contribute every time you get a raise. Doing it in advance make people don’t see the loss and make them more keen to participate. Moreover it is the default choice so they don’t resign, so a lot less people got out from it.  Participation in pension funds (401k program)  86% of whom had the default from the first day of work, continued having the retirement plan. They are more than the ones that can choose that  To not have a risky portfolio is important to start early, because not risky portfolios need more time to make a lot of money.  Contribution rate  Among new workers (NEW), 76% chose the default contribution (3% of salary).  In the US if you contribute a certain amount the company will give the same amount until the 6%., after that cap they will always contribute 6%  Most people put 6% trying to maximise the gain and the matching with the company. Someone chose to contribute 15% maximizing always the money form the company. A lot of people also stayed in the default choice of 3%, but doing it they are losing some money because they are giving and receiving less than the company cap.  However, a rational agent would choose a 6% rate since this is the maximum value that is matched by the employer and it would be more advantageous to maximize the contribution  In Scotland, workers contribute to the pension fund by default. The default amounts contributed by workers and their employers are: 5% from the workers and 3% from the employers. A contribution of 8% is not enough to reach a comfortable retirement. As a result, the BIT developed a new intervention to increase the amount contributed to the pension fund.  Young people between 22-29 years of age are those who should increase contributions the most.  The intervention used the following conditions: 1. Control condition with the same wording as the current plan. 2. Investment + rules of thumb: pension contribution was framed as an investment rather than saving and people was given rules of thumb of the power of saving early. a. The pension was seen as a way to make money, as an investment rather than retirement saving b. rule of the thumb = they were giving an evaluation on the amount of money they could get If they had invested early. 3. Labelled amounts: 12% to stay above the poverty line (“minimum”); 15% for comfortable retirement (“comfortable”). 4. Future focus: people were asked to think about the future, retirement, and life after work. Then they were told that an extra £80 a month now would amount to £150 a month more in retirement  Results: the label amounts convince more people, and the others were not so different in results to the original question. CASE STUDY FROM ITALY (RUBALTELLI & LOTTO)  People do not save enough for retirement and several interventions have been tested in the past. 45  ENPAP asked us to help them increase the amount of income freelance psychologist contribute to their professional retirement saving fund. Participation in the fund is mandatory, but people can choose the percentage of net income (between 10% and 20%) they want to contribute.  ENPAP did not want to increased the minimum contribution rate, but pointed out that many psychologists are not contributing enough. In particular because a present gain is better than the future one, most of the people chooses the 10% option, the default one (to see the others they should click on the number and see them in the window)  Instead of the drop down menu with the window, they introduced:  Possibility to see all the contribution percentage (with buttons)  The percentage where decrescent from 20% to 10%  They put 20% as default, but they could change it  Every time they change it, there were a pop up saying that the contribution and the pension will be lower  They were told the they can detract from taxes, and they offered also the calculation of the savings they could get with this  Results:  People before the intervention chose the default option of 10%. They passe from the 3% to the 27% of people that chooses the 20% contribution option.  Most people that went to 20% where young people that understand that it was more convenient thanks to the long time and the taxes.  We compared contributions in the year before the intervention (t1) with those made the year the intervention was introduced (t2). This difference corresponds to additional contributions to the fund in excess of 13 million Euros between t1 and t2. EASY: OTHER APPROACHES  To increase compliance with taxes or rules that ask citizens to provide a series of information it is possible to prepare pre-filled forms.  It is the case of the 730 tax form in Italy. In large part, tax evasion depends on people’s difficulty to comply with complex requests coming at many different times during the year  These reduces the errors and permit people to remember to do it  In the UK, an intervention to help empty the lofts allowed to increase the number of households who chose to insulate their roofs.  The offer was not only for insulate roofs but also to empty them with junk and move it somewhere else.  The intervention worked even when the cost was higher than just insulating the roof! EASY: HOW TO DECREASE A SPECIFIC BEHAVIOR  In several countries around the world, there has been an unexpected, but very effective law that reduced motorcycles thefts. Make mandatory to use the helmet reduced motorcycles thefts.  It is possible to reduce suicides in a similar vain, by making more difficult to commit such an act.  If pills are sold in blisters rather than bottles it is significantly less likely that someone would use them to kill themselves. This because they have to pull the out one by one and so they have time to thinl. ATTRACT: HOW TO MAKE PEOPLE PAY FINES  In Australia, people who default on their speeding tickets receive a letter. In a field test, a little addition, a couple of red words were added on the top of the page.  PAY NOW: This intervention increased compliance for unpaid fines by 17% (for a total of 10 million Australian dollars). In addition, by paying the tickets about 8000 drivers avoided the suspension of their driving license (comes when you don’t pay the fine after 3 letters). Finally, the state was able to save about 80.000 dollars because they did not have to send additional letters. 46  Also it seems that people put the letter somewhere they could see it and pay it, without forget about. ATTRACT AS AN ALTERNATIVE INTERVENTION INSTEAD OF FINES  Example: speeding tickets. Until the mid 1980s, the intervention used by UK police was to have hidden patrol cars using speed guns and taking drivers by surprise.  This strategy is the best one if the goal is to give a lot of tickets. However, it did not produce any behavioral change, people did not start to drive slower or in a less reckless way.  A first alternative is to provide drivers with information about the number of other people who comply with speed limits.  Further, patrol controls with speed guns were 10 times less effective than a street sign. This is one of the reasons why nowadays speed traps are visible and paired with specific street signs.  Recently, technology made possible to use signs that provide real time feedback to a driver about the speed he/she is moving. These signs led, on average, to a reduction in speed equal to 6 - 12 Km/h. These signs are particularly useful when paired with signs that remind the driver why he/she should slow down (e.g., school crossing, dangerous intersection, and so on)  The signs that tell that they are controlling make people go slower  Using new signs with half numbers: for example 69 instead 70. This because people see only the left part, so people where going slower. ATTRACT: RECYCLE BINS  A new intervention to increase the use of trash bins has been tested in Copenhagen.  In collaboration with the city council, the trash bins were painted with very bright colors (to attract people’s attention and become more visible).  Previously, like in many other cities, trash bins were painted with very neutral colors (e.g., gray).  Furthermore, researchers painted on the ground a bunch of footprints like steps moving in the direction of the trash bin  Analyzing students’ behavior around the university campus, researchers found that the intervention increased the use of the trash bins by about 45% ATTRACT: HOW TO USE EMOTIONS  Images on cigarettes packages are also a way to exploit the attract strategy.  In Australia also changed the font od the brand making it anonymous. This because after a while the images where not effective anymore.  It is important knowing iif the nudge is effective in the long run to prepare other strategies to substitute that. SOCIAL: SOCIAL NORMS AND SOCIAL COMPARISON  Sociality is influenced by two rules  Social norms = what others in the group do, become norm and everyone follows that  Social comparison = look to what the others do to understand what is the correct thing to do and follow them.  Asch’s (1951) experiment on conformism: participants were tested together with a group of confederates. Participants’ task was to tell which line on the right was identical in length with the line on the left. Confederates were instructed to make mistakes on purpose.  Most participants, when they answered after the confederates, conformed to the wrong answer. However, they knew all too well which answer was the correct one (it was not a perceptive error!).  So we conform to what others do even if we know they are wrong  Conformism is a strong behavioral effect which depends on different processes:  Informational social influence: o Often we use others’ behavior as an information to understand how we should behave. o It is a very useful information in new or ambiguous situations. o We use what the others do to take a decision 47  People need a little gain or motivation to leave a bad habit  In this situations, a potentially effective intervention is to use incentives, like in programs that help people fight against addictions:  Everyday, people who are addicted to drugs receive a discount to buy consumer good if their urine test is negative.  Often these little incentives have a large effect, despite the fact that the people receiving them have to cope with huge losses due to their addiction.  The discount acts as an immediate incentive, while without it people would face an immediate cost (avoid drugs) in order to get a benefit in the long term (improve their lives).  Similar programs have been developed to reduce:  Nicotine addiction (Volpp et al., 2006).  Obesity (help people lose weight; Jeffrey et al., 1983).  Improve high school grades (Angrist et al., 2006).  It is important to consider that on many occasions people do wish to change behaviors they consider unhealthy (e.g., smoking): However, they find it difficult to give up immediate gratifications every day in order to be healthier later on.  If people don’t want to change, the incentives are not enough because they don’t see their utility TIMELY: LOSS AVERSION  A study showed that loss aversion can be used to help obese people lose weight. Patients were asked to deposit money or other valuable objects and sign a contract with their doctor: Patients would get back the money (or valuable objects) if they follow the weight reduction program and achieve the intermediate goals.  Of all patients taking part in a test program: 47% lost more than 9Kg and 70% lost more than 7Kg.  At the end of the program a lot of people regain weight, so it was effective for the time of the intervention, but it is not enough to make them going on also after that  By asking patients to deposit their valuable allows to reinforce the motivation to lose weight: People anticipate the negative affect they would experience if they fail to get their money back.  This type of intervention takes advantage of people’s optimism when they decide to start a weight loss program: Usually, people are quite convinced they will achieve their weight goal. As a result, patients are also quite willing to deposit some money because they underestimate the difficulty of following the diet every day  Sharing is important to incentive the person: in fact saying that they commit publicly they will achieve to don’t disappoint people FRAMING: FREE VS. DISCOUNTED  Framing was adopted to increase people’s use of a bike sharing system in the Portland, Oregon (Kirkman, 2019): they tested two different framing on people who had just moved where there were this program yet and with people who already lived there but they had just placed this service near their homes. They used two type of frame: the discount and the free to the program of bike sharing  Hypothesis:  the free frame should be more effective than the discount, because people like more having something without paying.  Those who had just moved into the neighborhood will take up the offer more often than those who had a newly built infrastructure close to their home, because is timely giving this service to who is adapting to the new house and neighbourhood. This is timely because everything is new to them, so they are really willing to see if it is a good habit.  The others instead should brake their habits and this is more difficult. Overall, number of redemptions will be low because transportation habits are hard to change and marketing is not targeted.  No long-lasting effect expected 50  Results: the free option is more effective on everyone. People who lived there were more likely to take the offer if it was free, and for who moved both offers were equally effective. When someone moves to a new neighbourhood it is a good time to change people habits in a more long lasting time then when the introduction of the program comes later on. RE-FRAMING FINANCIAL RISK  A study investigated how the framing of risk-return representations can influence investors’ perceived complexity, usefulness, and understanding of content (Gentile et al., 2015). nvestors were presented with information about the risk, return, and costs of four products (outstanding structure bond, newly issued structure bond, two stocks varying in risk level). 4 conditions: 1. Synthesized form: Aggregated market, liquidity, and credit risks information.  the aggregate information gives a general sense of the product 2. Unbundled form: No information aggregation.  explanation of all the products separately 3. What-if scenarios: Presentation of risks as a function of relevant indicators. 4. Probabilistic modeling of expected returns  the likelihood of gain in every condition Finance is very complex to understand and more you risk, more you will gain quickly. Moreover the stocks are more risky and bond are safer. Results: Complexity of information format affected investors’ accuracy of judging the products’ risk: Their assessments were more likely to be correct for the less complex formats.  The first was the easiest to understand, in fact they made the best financial decisions. Instead the what if condition was the most difficult to understand. But maybe this way of proposing things is not right because the bank has to show all the information to the possible buyer/trader, and saying the things synthesized is the opposite NON-MONETARY INCENTIVES  The incentives that can be used to reinforce healthy habits do not necessarily have to be monetary ones.  In poor countries, monetary incentives could actually open the door for counterproductive behaviors that were previously unattainable. In contrast, non monetary incentives can be directed to enhance the behavior that is targeted by the intervention. So an option could be to give them food or something they really need, giving them the knowledge for example.  An example is a field study whose goal was to increase immunization in rural and poor areas of India.  For each child, a full immunization requires 5 trips to the hospital. The field study divided people into three different groups:  Control group: Full immunization paid by the Indian government (WHO/Unicef package). Immunization administered at the local hospital.  Intervention A: A mobile clinic (“Immunization camp”) with a nurse was sent to several villages on fixed days every month.  Intervention B: Same as the previous intervention, with the addition of 1Kg of lentils for each phase completed and a set of metal plates (for each child) once the full immunization was achieved Results: the intervention B with the incentive, immunized more people than the other two conditions. DEFAULTS AND IMPLEMENTATION INTENTIONS: INCREASING VACCINE COVERAGE  The WHO estimates that vaccinations prevent approximately 2.5 million deaths each year. Still, vaccine diseases remain a major cause of morbidity and premature mortality in the world (WHO, 2013). Why people don’t make use of vaccines? 1. Complacency: They do not care about immunization or they think that the diseases is not dangerous 2. Inconvenience in terms of physical availability, affordability, or geographic accessibility. The difficult to get it (difficultly to go to the hospital for example) 3. Lack of confidence in the vaccine itself or the healthcare system. 4. Estimates that the pros do not outweigh the cons (Betsch et al., 2015). 51  Chapman and colleagues (2010) tested the effect of defaults on flu shots: 1. Opt-out: People received an email that they were scheduled for a flu shot appointment (with date, time, and location). They could change or cancel the appointment. 2. Opt-in: People received an email explaining that free seasonal flu shots were available with the option to schedule an appointment Results: who get the appointment in the opt-out frame actually went to it.  Implementation intentions is another way to increase vaccination rates (Milkman et al., 2011). 3,272 employees of a large firm received one of three different emails about workplace vaccination clinics providing dates and times of clinics at their work location: Two formats of the email included prompts for the employees to write down the date (version 1) or date and time (version 2) they planned to get the vaccine. So in one case the employees can choose the date and in the other one could choose also the time.  Results: Version 2 was the most effective and increased vaccination rate by 4.2%; this because they could choose more. PSYCHOLOGY OF DISHONEST BEHAVIOUR  Usually people are more honest than what we think that they are honest. If can cheat without being caught, they do it a little bit, in a more subtle way, because in this way they can produce a justification: having a justification means that they can say that we are honest.  So people is more honest than we think  People cheat a little bit only when they can justify themselves. INTRODUCTION  How can we explain that some (many?) people do not do what seems obvious and correct to do in these situations? Sometimes we get an advantage and we want to use that.  Sometimes we invoke moral principles and personalize them even if the situation do not require them, only to justify an opinion on a situation that we feel that is wrong or disgusting. So the reaction come before the rationalization  So when we have the possibility to cheat we are in conflict, this is resolved only when we find a justification in way that can change a little bit the categories or what is acceptable and what is not. NORMATIVE APPROACH  Gary Becker, a neoclassic economist, developed a normative model to fight (and explain) dishonest behavior:  Cheating is a rational decision based on a compromise between: o The chance of being caught (and the possible punishment). o And the benefit accrued by not saying the truth.  In such a model, to reduce cheating is sufficient to: o Increase the frequency of controls and increase the probability of catching them. o Increase the severity of the punishment. This should take from them the incentive that make them cheat  Let’s apply this model to tax evasion. The solution is quite simple: We should increase audits and associate them with very hard punishment (e.g., prison time). BEHAVIORAL APPROACH VOLUNTARY CHEATING  Psychologists try to implement simple interventions to induce people to behave more honestly (Bazerman & Gino, 2012). How?  They investigate how people justify themselves  How they judge others VOLUNTARY CHEATING  Imagine you have 5 minutes to solve as many of these numerical matrixes as possible in the poor time they gave them (for example find the couple of numbers that give 10). For each correct response you get 50 cents. 52  It is easier to feel compassion with one person than a multitude of people, this because dividing compassion between more people make it less stronger in the feeling then concentrate all the emotions in one person.  Example: the media coverage of war in Syria, was at his peak when they talked of chemical weapons, but the become lower and it is not proportional with the increased numbers of deaths.  There are different models to understand how we measure the value of a life:  A normative model values every life equally, but it is not like this: people values more the lives of people in exceptional and sudden crisis because they empathize more. In fact they donate more in these cases (earthquake, hurricane), and less when the problem is chronic and slowly developing issue (malaria, climate change)  People also donate more to things that are popular and known: for example breast cancer and not heart disease that is the first cause of death in the US  We are not aware of statistics: regardless form the numbers we give when we feel something and we are aware of something because we apply heuristics.  There is also a correlation between how much people help and how many people die: so people give more when there are a lot of deaths rather than when there is a lot of people affected. From a rational point of view is more useful give money to those that are affected and not the ones that died because they can’t use the money. Why? The emotion is more relevant, so you feel worse for the deaths, and they are also a value that is certain and simple to try to compensate with money. For the affected are more difficult to understand how much they loss: so it is an uncertaing concept, it is more difficult to value it. IRRATIONALITY IN PROSOCIAL BEHAVIOR: NATURAL DISASTERS VS. “CHRONIC” CRISES  However, people (and many charitable organizations) provided more help to those who were hit by natural disasters than those hit by chronic crises: because of such behavioral bias there is basically no correlation between how much people give and the amount of help needed  Example: In 2006, in the U.S., people donated $1.893 per person affected by hurricane Katrina. However, they gave only $10 for each American diagnosed with HIV.  One of the reasons people tend to neglect chronic crises is the difficulty to constantly face with the negative emotions they induce. Moreover they are such big problems that it look as everything one person can do it is insufficient to make a difference: the person can contribute a little bit, but this will not solve the problem.  In addition, these crises are so large that it looks like it is impossible to make any difference (“drop-in-the-bucket effect”). The need for help is such that any action seem ineffective to stop it.  Finally, on many cases, media dedicate much more time to local issues and natural disasters than chronic crises, therefore strengthening donors biases, because they are not such available in our mind.  Small (2010) provided laboratory evidence for this effect.  The explanation proposed by Small is based on prospect theory.  Sudden crisis, like natural disasters, make affected people look like they lost something (e.g., their houses, their health, etc.).  Chronic crises, such as malaria or malnourishment, are not perceived as losses but rather it looks like people affected missed out on a potential gain (e.g., getting richer/less poor or improving their health)  Value function of prospect theory: natural disasters or sudden crisis are perceived as losses. The origin of the axis is the normal life and 55 they suddenly lived a loss, so from the normal life people lost something, so for us is easier to imagine their suffer and we put ourselves in their situation and we feel moral obliged to help them to get out form that and that is perceived as very unfair. What happens with chronic diseases is that they develop slowly and they are less visible than natural disaster, so we feel that they are living a normal life, and our donation will increase their wealth and quality of life. So in people perception it is the way they always lived so why we should help them.  Volcano eruptions and earthquakes need less death to attract news coverage than food shortage. MORAL BEHAVIOR:  The trolley and footbridge problem measures the value of life. Should we kill one person to save 4 people (least bad option when someone will die anyway), or push a man off the bridge to save 4 people (they feel more responsible). So life value changes from one situation to another.  When we think about how we should behave and moral behaviour, we are guided by this moral principles and we follow them. But more recent theories suggest that we behave and then we rationalize our behaviour in the sense of what is moral and what is not. So even if we did something immoral or disgusting and then we try to find a justification and rationalize it.  Moral reasoning is usually aimed at reaching a specific goal:  However, on many occasions, it is based on post hoc justifications that people interpret as objective/analytic thinking.  Moral behavior correlates more strongly with moral emotions (such as empathy) than moral reasoning  Starting from these conclusions, Haidt (2001) proposed the «social intuitionist model»:  This model modifies the role of moral reasoning: o From being the cause of moral behavior. o To being the consequence (a rationalization of something that has already been done).  We act because of a set of moral intuitions that are driven by our emotions and explain our behavior through conscious reasoning. Sometimes it helps to behave better, but sometimes it makes our behaviour worse because 30.10  confirmation bias  First we act and then we rationalize it.  When people experience empathy towards someone who is in troubles, they use the emotional activation as a cue to grasp the moral relevance of the situation.  This reaction leads people to think about moral principles and values that could apply to that specific situation.  These processes are also supported by the “emotion-as-information” SINGLE VICTIM EFFECT  Jenni and Loewenstein (1997) showed that people are more willing to give to identified victims. In an experiment, half of participants were asked to help a single child while the other half where asked to help a group of unidentified children. Results showed that participants gave more money to the identified child rather than to the group.  The scenario described children who could be affected by lead poisoning and may need to be tested at the local hospital.  Jenni and Loewenstein found that the difference in the amount of the donation was not explained by vividness of the single child compared to the group (M = .44 vs M = .64; p = n.s.).  What made the difference was that for the single victim participants knew how many children were at risk and how many they could save (higher perceived effectiveness).  In the group condition, it was less clear how many people could need help and how many could be saved  It is not because the single child is a more vivid image, but that helping a single victim made people think that they are helping at 100% the person at risk. They knew that they will be able to really help the life of that child. 56 o It is more clear the effectiveness of the donation and when you help more people you are not aware of what is the impact of your donation. So sometimes the donation depends on emotions and empathy and sometimes on the thought of making a difference (effectiveness)  Small, Loewenstein, and Slovic (2007) studied the effect of the identified victim when compared to the presentation of statistics regarding the overall amount of help needed. Would people give more when they are informed that millions of children are at risk or when presented with a single child at risk?  Results: This experiment measured willingness to donate (WTD) and it used real donations based on the amount given to individuals as a participation fee.  They said to the participants that they should pay attention to the single victim effect, but that didn’t make any difference in the amount donated when they read only the statistical version (instead of donating more), and when they read of the single victim knowing about the effect they donated less.  Moreover is interesting to notice that the single victim effect is effective only if there is only the description of the single person and not when it is combined with the overall statistics. This because when we react not to the single person, but to the fact that we can pretend that the problem doesn’t exist. With the statistics we are more aware that the problem exists and that many other people needed help. SINGLE IDENTIFIED VICTIM EFFECT  Kogut and Ritov (2005) showed that the identified victim effect is actually very strong because it also involves a single identified victim. These authors presented two groups of people with:  A single sick child who needs money for a very expensive medical treatment (participants saw the child name, age, and a picture).  A group of 8 children who needed the same treatment as the single child in the other condition (for each child, name, age, and a picture were available). Participants were told that the overall amount of money needed to save the child or the group was the same (in this way they reduce the justification of the not helping behaviour, because they could think that their donation will not be relevant). So the donation had the same value in both the condition Moreover the children could be identified or not (the photo and the description of them).  In both conditions, Kogut and Ritov measured:  Willingness to give (with real donations).  Affective reactions  People were more willing to give to the single identified child rather than to the identified children belonging to the group. Despite the fact that the overall amount of money required was the same in both groups.  Kogut and Ritov tested two additional conditions in which both the single child and the group were not identified. This time the difference between the two conditions was not significant. Only when the children were identified there were the single victim effect  People were donating more to the single identified child than the group.  Participants had more compassion (positive affection reaction) for the single child Conclusion: So people were more willing to give to the single identified child than the identified children belonging to the group despite that the sum required of money were the same for both the groups. And the same difference is not seen in the unidentified case  Collapse of compassion function: after one person there the maximum compassion that then starts to decrease.  In this field people perceive that we should value every life/person, but the literature shows that the fewer people is at risk, the more is the help. When the number increases people don’t identify and feel a connection with the victims and they help less. Not because they don’t understand the greater need of 57 (between subjects). In these increasing situations they could help 2 children with 5$ till the last situation where they could help 90 children with 225$.  Results: 80% chosen the first scenario and only 14% the last one: in the last option was very expensive so this is not surprising. But the perceived benefit of helping 90 children is higher than helping 2 children.  Benefit increases so they understand that they are helping more children but the cost is perceived more precisely: less people chooses to donate more money.  The benefit is perceived in a very emotional way than the money: emotions are a on off system that activate with the benefit but we switch off with the money decision (situation that we live everyday)  50$ is the maximum people is willing to donate, after that the amount of donations decreases drastically even if the benefit is perceived higher than the cost, but we have to perceive a much higher benefit than the cost to donate (the different has to be enough). So to make people donate we have to show them that the difference between costs and benefits is huge  Experiment 2: they put together the previous scenarios together. Lump sum (they put all the conditions together) versus bundle of smaller programs:  Lump sum conditions: o A $150 [or $165] donation to help 60 [or 66] children living in a refugee camp in Kenya.  Bundle conditions: o A $75 donation to help 30 children living in refugee camp A in Kenya. o A $50 donation to help 20 children living in refugee camp B in Kenya. o A $25 donation to help 10 children living in refugee camp C in Kenya. o A $15 donation to help 6 children living in refugee camp D in Kenya. If they choose to donate they have to donate to all these programs. As we can see the total is 165$ and 66 children.  Results: in the bundle condition was chosen more than the lump condition, and considered the lump sum condition too much to donate. This effect is present not only in the condition with less money asked (150), but also when you ask more money (the same amount).  Experiment 3: they used the attraction effect, where they added a third dominant alternative to two that were already there. Participants were asked to help a neighbour who had suffered a loss because his house had burned down. The target was to make people donate more money: in one condition they had a lower cost (blanket and less money asked)and in the other they had a better benefit for the neighbour (blanket+pillow but more money asked). In the dominant condition they could chose a better quality blanket in exchange of more money of the first condition but less than the last one.  Results: people choose more the blanket and pillow option when was present also the dominant alternative (adapting effect). Even though the benefit of the target alternative increased a little bit and the cost went down, confronting that with the blanket+pillow condition, the last one was perceived as a better way to spend their money.  Similar results were found also when:  Quality of the blanket (benefit) was presented in a less evaluable fashion: 4.5 stars vs. 4.5/5 starts (presenting the maximum points). So they told how good was the blanket to see if this made the difference.  When giving participants: o A chance to make no donation at all. o An exit option (choosing freely the amount of money they wanted to contribute). This is a way to nudge them to give money giving them to possibility to choose.  When the target included a bundle of products unrelated to each other: Blanket and towel vs. blanket and pillow.  In line with the need to test interventions in the field, we found that the attraction effect can also decrease donations if the decoy is dominated by a low-amount alternative. 60  Experiment 4 on joint and separate evaluations: some dimensions are easier to valuate when joint and we have a comparison. Joint evaluation (JE) versus separate evaluation (SE): People were asked for their willingness to give: 1. $75 to fund a charity supporting research to find a cure for cancer. 2. $15 to fund a charity supporting an amateur baseball league.  Three conditions: 1. JE: Participants saw both causes and could support either one of them, both of them or make no donation at all. 2. SE1: Only cancer research. 3. SE2: Only amateur baseball league  Results: when cancer is presented with the baseball league becomes more important than in the separate evaluation. In all condition the main difference is that cancer is always considered the most important cause. Comparing cancer, asking more money, with baseball (less money) is considered more important, in fact people donate more.  Motivation to help is way higher for cancer when it is in the joint evaluation than in the lone one  To nudge people to donate more we can compare a more important cause with a less important one.  In a second study, we manipulated the relative cost of the two causes:  The donation to cancer research was always $50  The donation to the amateur baseball league was: $20 or $50 or $80 To prove that the cause is not the amount of money asked. They didn’t found any difference from the previous experiment. So the big difference is the cause, not the amount of money. RISK PERCEPTION  When we face a risk we understand the risk and the we try to analyse and quantify that. At the end we mage the risk doing something about it 1. An event is defined as dangerous, like the explosion of a nuclear power plant, because the negative consequences are given (although their level could vary) 2. An event is defined as risky, like talking on the phone while driving, because the negative consequences are possible but not given.  So in both cases there is a negative event that will be happen but in one case we know the negative consequences, in the other one they are not sure OBJECTIVE VS. SUBJECTIVE RISKS  OBJECTIVE RISKS: An event is defined as dangerous, like the explosion of a nuclear power plant, based on the probability that the event occurs and the severity of the negative consequences.  It can be computed using formulas and models more or less sophisticated; these computations can vary depending on a set of parameters added to the model. So it depends on objective factors and for this reason we can calculate it.  Example: The risk of a stroke in a specific age group depends on genetic predisposition, smoking, drinking habits, lack of exercise, and so on.  SUBJECTIVE RISK: An event is defined as risky, like talking on the phone while driving, because the negative consequences are possible but not given.  We can talk of "subjective risk" because the impact of the consequences is usually evaluated differently by different people and depending on the procedures they use to judge it.  Example: The risk of a stroke depends on the subjective experiences and thoughts related to this risk.  It depends on the person that is evaluating the risk and it is not based on objective data EXPERTS VS. LAYPEOPLE  EXPERTS: Expert analysts perceive and judge risks in a quantitative likelihood perspective with consequences. 61  LAYPEOPLE: Laypeople (i.e. not expert at all or just a little) generally refer to qualitative dimensions of risks.  Example: climate change (97% of scientists vs. around 50% of Americans) so scientists agree that climate change is happening CHARACTERISTICS OF RISK FOR LAYPEOPLE  Dread vs. Not dread: is this a risk people have learnt to cope with and to think about in an emotionally cold way or is it a risk that terrifies people and makes them prone to impulsive, emotional reactions?  Example: Terrorism vs. Caffeine.  Controllable vs. Uncontrollable: If you were to be exposed to this risk, to what extent, with your own knowledge, would you be able to avoid negative consequences?  It depends on how we feel in control of the situation  Example: Driving a car vs. Plane crash.  Voluntary vs. Involuntary: Does people voluntarily choose to face this risk or is something that just happens?  In this sense involuntarily is connected with the uncontrollability  Example: Extreme sports vs. Terrorist attack.  Chronic vs. Catastrophic: Does this risk kill one person at the time (chronic risk) or does it kill large numbers of people at once (catastrophic risk)?  How many people can die for this event?  Example: HIV or cars vs. Tsunami or earthquakes.  Fatal vs. Not fatal consequences: When a risk entails injuries and illness, is it likely that the consequences could be fatal?  Example: Nuclear reactor accident vs. Aspirin side-effects.  High vs. Low risk to future generations: To what extent an activity or technology puts future generations at risk?  Example: Air pollution and climate change vs. Smoking.  Immediate vs. Delayed consequences: When a person gets in contact with a substance, technology, or engages in a specific activity, is death immediate or does it happen after a period of time?  Example: Plane crash vs. Smoking  The consequences are immediate or delayed in the long term? In the latter the perception of risk is lower, because we don’t feel the immediate impact.  Observable vs. Not observable: When something negative is about to happen because of a specific activity, substance, or technology, to what extent the damage is observable?  We can see the damages of an event?  Example: Explosion vs. Smoking or pesticides.  Known vs. Unknown to those exposed (or to science): To what extent the risk is known to people who are exposed to it? To what extent risk are known by science?  How much we know and science know about that.  Example: Cars and surgery vs. GMO.  New vs. Old risk: Is this risk new or familiar?  When something is new we are no used to face and we feel more threaten by it. On the contrary if something is familiar we tend to undervalue the risk  Example: COVID-19 pandemic vs. X-rays COGNITIVE MAP OF RISKS  These characteristics of risk can be used to: 1. measure the subjective risk 2. create specific profiles for each activity and 3. compare different activities along different dimensions. 4. So the psychometric approach aims to do exactly this kind of measurements using the cognitive maps of risks where they put all together the different types of risks. 62  Knowledge  The model reports that only those who are informed about the origins, consequences and proper behaviours that need to be adopted are able to perceive and assess the climate change risk adequately, so they are more concerned about the matter.  Different kinds of knowledge: natural cause-knowledge, human cause-knowledge, impact- knowledge, response knowledge (= how to act), scientific consensus EXPERIENCIAL PROCESSING  Negative affect  Affect = a quick associative judgement or a "faint whisper of emotion" (Slovic & Peters, 2006)  Affective imagery and negative affect in general are important predictors of global warming risk perception (Leiserowitz, 2006; van der Linden, 2015), because they are reletated to how vivid is the image: in fact the more you have a vivid image, the more you have a negative affect, the more you are concerned about climate change  It has a relationship with cognition: negative affect of personal experience needs to be mentally associated to climate change in order to shape its risk perception (Weber, 2010) otherwise it will no impact the concern about climate change  Personal experience  People form affective associations through personal experiences and affective evaluations of future risks depending on the vividness of the imagined impacts of climate change. So the experience has an impact and if we have an experience about something that can determine climate change, we will be more concerned about it.  What we experience and detection of climate change: rising temperatures, seasonal weather, precipitation change, heat and warm daily temperatures. But also we experience extreme weather events, such as hurricanes, flooding, heat waves and droughts. SOCIO-CULTURAL INFLUENCES  Social norms: referring to the way other people (descriptive norms) and to the way people are expected to (prescriptive norms) act, think, or feel in specific situations. So the more people see climate change as a serious problem, and the more the individual risk perception is amplified (van der Linden, 2015).  Value orientations: they are basically the same in different countries, they are not mutually exclusive, they have been validated in many cross-cultural studies. They are three: 1. Egoistic values - oriented on maximising personal outcomes. On the contrary following our egoistic desires makes us care less of the enviroment 2. Socio-altruistic values - oriented on caring for others. They influence a lot the risk perception of the environment. 3. Biospheric values - oriented on caring for animals and plants. They influence a lot the risk perception of the environment. HEURISTICS AND BIASES  The mismatch between the environment in which heuristics evolved and their application in modern ("global") contexts makes people to misperceive or underestimate the risk of climate change in a number of important ways.  Consensus heuristic  Confirmation bias  Present bias  Status quo bias  Single-action bias  Hyperbolic temporal discounting  Intergenerational discounting  Local warming effect CONSENSUS HEURISTICS 65  Consensus heuristic: Public's perception of the degree of scientific consensus influence important beliefs related to climate change, such as concern about global warming (van der Linden, 2015).  Widespread public misperceptions of the degree of scientific consensus dampen concern about global warming.  Conveying the scientific consensus on human-caused climate change can increase acceptance of and concern about climate change (Lewandowsky et al., 2013).  If the public feel that the scientists agree on climate change, they are more concerned about that. CONFIRMATION BIAS  People generally seek information that confirms their prior beliefs and information, while discounting the ones that contradicts them.  It is determined by the absent or poor knowledge about the causes and/or consequences of climate change, as well as the action that should be taken to tackle it.  Example: Reading news articles that are consistent with my point of view PRESENT BIAS  People generally give stronger weight to payoffs that are closer to the present time when considering trade-offs between two future moments.  It is determined by the inaccurate understandings of the drivers of climate change  Example: Choosing to drive (instead of taking the bus) to the grocery store.  we undervalue the benefit that we can have from a present action for the future of the climate change. STATUS-QUO BIAS  People are less willing to change the status quo, because it would require efforts and uncertainty is also involved  It is determined by the misperception of, for example, carbon footprints associated with each action taken (i.e., they are underestimated)  Example: Buy usual products rather than try new and more sustainable alternatives. SINGLE-ACTION BIAS  People are likely to take one single action to reduce a risk that they encountered and are worried about and they are quite less likely to take other actions that would increase the protection or reduce even more the risk (Weber, 1997). The first action taken reduces feelings of worry or vulnerability, then people are less prone to take additional actions.  Example: Illinois farmers engaged in one single action to adapt to climate change, but hardly ever engaged in more than one of those practices (Weber, 1997) DEBIASING TOOLS: HOW TO ELIMINATE BIASES  Confirmation bias  Forward-looking: generating arguments for forward-looking options or considering the legacy for future generations, makes people less prone to confirm their believes and to understand more climate change  Present bias  Default: making climate-friendly options the default when you have to choose something, because instead you will chose the higher benefit in the present  Single-action bias  Identity or value reinforcement: reinforcing one's identity and values to promote subsequent climate actions that can have an impact ()as buying something  Status-quo bias  Visualization: highlighting human contributions to climate change as opposed to natural causes leads people to actually do something and no just wait TEMPORAL DISCOUNTING (Hyperbolic)  People tend to heavily discount uncertain future risks, such as climate change impacts.  Temporal discounting is the tendency to place less value on the future consequence of a phenomenon due to its increased temporal distance from the present (Hardisty & Weber, 2009). For example we prefer to have something that gives us what we want now but that pollutes, than something more challenging that will be better for the planet (ie. mono use plastic)  People think more about climate change when they are old, because the discount uncertain future risks. 66 INTER-GENERATIONAL DISCOUNTING  Inter-generational discounting is the tendency to weigh less the benefits that future generations will get rather than the costs of the one's own generation (in the present)  It is effortful to invest resources in the present to transfer consumption from ourselves for the benefit of people distant in the future (Schelling, 1995)  Such discounting leads to an underestimation of long-term projects and political programs in cost-benefit analysis. CONSTRUAL LEVEL THEORY  The past biases are related with this theory  Trope & Liberman (2003, 2010) proposed that temporal distance alters people's perception and responses to future events compared to those in the present, because increasing the temporal distance, the mental representation become less concrete and more abstract.  When temporal distances are larger, events are more likely to be represented in abstract terms that capture their central features (high-level construals). So the risk perception is lower  When temporal distances are smaller, events are more likely to be represented in more concrete and incidental terms (low-level construals). So the risks are more vivid and clear in our mind and they are higher LOCAL WARMING EFFECT  People rely on simple available information, such as daily temperature, to make judgments about a more complex and less accessible phenomenon, such as global warming (Zaval et al., 2014)  Note: due to the high variation in short-term weather, the local warming heuristic is an unstable inference tool for forming risk judgments about global warming. PRO-ENVIRONMENTALBEHAVIOURS  Pro-environmental behaviours have been studied in relation to a lot of different constructs since there are many that play a key role in promoting such behaviours:  Affect and emotions (Brosch, 2021)  Personal and group values and identity (Bouman et al., 2021)  The role of social norms and policies THE ROLE OF AFFECT AND EMOTIONS  The affective responses to climate change predicts mitigation behavior, adaptation behavior, policy support, and technology acceptance.  Positive anticipated emotions have a stronger effect on intentions to fight climate change for individuals who are already engaged in climate mitigation, while for those not engaged yet, the stronger effect is shown by negative emotions (Odou & Schill, 2020). So it depends on the people we are considering: if you are engaged, you think about the positive emotions of acting pro evniromantally, instead if you don’t care emotions as fear have more impact.  The case of warm glow (i.e, pro-sociality moved by the positive emotional experience of helping others): positive feeling that we experience when we help others. In particular, anticipated green warm glow predicted self-reported pro-environmental behaviors four weeks later we can see this effect also when there is a delay between the warm glow and the actual behaviour. THE ROLE OF AFFECT AND EMOTIONS  Induced compassion for climate change victims leads people support more climate mitigation policy  Induced empathy toward the suffering of polar bears leads people to donate more to climate change activism  Study assessing multiple which emotion helps with what: pride increases intentions to invest in environmental protection, guilt increases willingness to repair environmental damages, anger increases tendencies to punish others for negative environmental actions  Take into account the transient nature of emotional reactions (sadness effect, attenuation of time-delay and re-establishment of making a non- binding commitment) 67  People is not very good in mathematics and with probability. Let’s see an example:  Imagine you invested €2.500 two years ago. If the price of the stock gained 30% during the first year and lost 30% during the second year, how much money do you have now?  €2.500  €2.725  €3.250  €2.275  The last one is the right response: because you gained less than you loss (we applied the percentage to a higher value than the original one where we applied the gain). This effect is called composite interest. People usually fail in responding to this question because they do not understand math. INTUITION AND INVESTMENTS  Confusion between similar tickers (= string of letters that indicate the indexes) that are presented near the name of the companies when we have to invest. This acronyms should be evocative of the meaning, but people don’t know what they mean and do not go to see to what they are referred.  The MCIC - MCI case:  MCIC was the ticker used in the New York stock market to identify MCI Communications, a company in the hi-tech sector that was very popular during the “internet bubble” (between 1995 - 2001).  MCI is the ticker used in the stock market to identify the Massmutual Corporate stock fund, a Massachusetts pension fund that, during the bubble, had no technological companies in its portfolio because considered too riskier. People stated to invest on its tickers and not only on the right company MCIC to not lose money because they thought to reduce the risk (A variegated portfolio in fact permits to compensate the losses form some companies with the gains from the others). But they didn’t see that the tickets were different and indicated different typo of companies.  When MCI Communications was very active, made many acquisitions and the media reported about this company, the flow of capital towards the Massmutual Corporate pension fund increased. And when the bubble burst, MCI Communications was hit by the crisis, and eventually went bust and failed. At this point, the flow of capital towards the Massmutual Corporate fund decreased significantly.  Changing name increase more investments than changing investment strategy for a company.  So labels are very important in the market  How is that possible? The company and the pension fund have no connection… Could it be a spurious correlation? Actually, investors used the two tickers as a cue to assess whether the company and the pension fund had something in common.  Many investors may have preferred the fund to have a higher portfolio diversification. PSYCHOLOGY OF DECISION-MAKING AND FINANCE  How can judgment and decision-making help us understand investors’ behavior? 1. It allows to explain asset allocation and diversification mistakes. So we can influence how many people invest and which tools they use in the investment strategies based in who we present them 2. It helps understand investor’s risk perception. 3. It explains how the intuitive system impacts the estimates of expected returns.  According to finance models, risk and return should be positively correlated: Investors who accept to take a risk should be rewarded with higher expected returns.  By taking a big risk it is not guaranteed that higher returns will follow, but if you take small risks in the long term you will have big gains, because even if there are big fluctuations in value in the short term, at the end the trend is positive on the long run. 70  So if you invest of various different companies, not in the same field and not concentrating the investment only in one company, you can also take some risks investing in stocks and ending up gaining.  However, risk should correspond to an higher variability of expected returns (not just in terms of potentially larger losses but also larger gains).  Historically, stocks (high risk) achieved higher returns than bonds (low risk), although such difference got smaller with time. But because stocks go up and down quicky, they are considered dangerous.  Statman, Fisher and Anginer (2008) studied the perception of the risk-return correlation by interviewing a large group of affluent investors. They asked investors to assess a list of stocks from companies included in the Fortune (economic journal) Most Admired Companies list. Participants judged every company based on: 1. Affective reactions. 2. Company risk. 3. Stock expected returns for the following year  They predicted that the companies that are higher in the list will be liked more by the people because they give them positive affections.  Results: there is positive correlation between the companies at the top of the ranking and the positive feelings and the expected return and negative correlation with the risk. So the riskier companies are considered not capable of giving a great return in the following year. ASSET ALLOCATION  Stocks are considered riskier by people so they don’t invest on them and they evoke negative affects, instead bonds are seen as lower risk and the encourage positive feelings. So there is a dangerous belief that the bond performance is better also when the bonds are riskier because they are performing too well: so the government will go bankrupt.  So the relation between risk and investment is driven by affects.  Asset allocation are the different type of investment that we can have. It depends on the overall things we have: how much money we have, houses, jobs… So it represents how we manage our portfolio  When we invest too much in bonds and not in stocks because they are considered too risky the asset allocation will not be very good.  Financial models made a few assumptions regarding the way investors allocate capital to different portfolio assets:  The rational theory about allocation also says that we can create a portfolio make of all stocks in a way that maximize gains and minimize risks: it is assumed that investors are perfectly rational and able to maximize the risk-return tradeoff. To do so, they diversify their investment portfolio, selecting a broad enough number of stocks from different sectors (and possibly different countries). o In this way gains and loses balance  Maybe it don’t permit to gain always but in the long term you will gain.  They should keep track of the correlations among different stocks and choose those options that are “less than perfectly correlated”.  They should invest only in stocks.  Life cycle theory: young people should invest in stocks because they have more time to make grow the asset and to recover from the small losses. And the old people should switch to safer options because there the possibility to retire in a economical crisis and loose all the money. People usually do not do this because when we are young they don’t have a lot of income, maybe they have mortgage, they have babies… So most people start to invest seriously when it is 40-45 when hey have a more stable financial situation, but it is very late, and if they make big risks they risk to loose all they retirement pension. 71  In reality, investors do not diversify enough. In other words, they choose to invest in just a few stocks. American investors own, on average, only 4 - 5 different stocks (when pension funds are not included).  Even Markowitz, who developed the modern portfolio theory, did not invest 100% of his pension fund in stocks. He preferred to split it between stocks and bonds to “reduce his future regret.”  So people see a negative correlation between risk and returns. This depends also on the familiarity that we have with the asset (we are more positive than we should be) and if the market goes well (it make easier to take some more risks). BIASES IN ASSET ALLOCATION  Often investors follow a naive theory of diversification. Such theory simplifies a very complex financial concept and leads to serious biases.  The main behavioral implications of the naive theory of diversification are the following:  Equity premium puzzle = people don’t invest in stocks enough. They are risky but you can reduce the risk.  Home bias = people invest in companies in the wrong country  Equity premium puzzle: This is investors’ tendency to ask for an excessive risk premium when investing in stocks (because of this bias only a tiny percentage of the portfolio is invested in stocks; Mehra & Prescott, 1985). This means that investing in stocks is perceived as taking a risk and so it is seen so dangerous that they want to be offered a very big gain to make a risky decision on their money.  Home bias: This is investors’ tendency to choose disproportionately from local investments (at country or regional level). Again the outcome is a lack of diversification. ALLOCATION: EQUITY PREMIUM  People don’t invest a lot in stocks contrary to what says Markovitz, anyway they prefer the total participation instead the direct participation. Moreover in contrast with the lifetime theory people usually invest in stocks in the adulthood (30-60 yo) when they have more money and start to think to retirement and for this reason they invest in stocks and not in bonds because they need more money faster.  Direct participation mean that people invest in single companies: they are picking single companies based on what they think is better.  Total participation include funds: in this case you choose a fund that invest in a lot of companies (ie only in development countries)  The main reason that leads investors to “avoid” stocks is a phenomenon known as “myopic loss aversion” (Barberis & Huang, 2006). The stocks go up and down and if we look to the short term we only see the fluctuations, but on the long term we don’t see the fluctuation but the fact that the trend is positive. People do not see this because:  Investors have a tendency to check and assess investments too often because they are worried of losing. Moreover the news system gives everyday a lot of information that influence the investor’s behaviour.  Disposition effect: people tend to lose stocks forever and they sell winning stocks immediately (too early because they could have gained more). But in the US people tend to sell the loses to lose some wealth and pay less taxes. So people checks too often (almost once a year to reduce the taxes) and then they cannot see the trend on the long term.  People also anticipate the regret of doing bad decision, in fact if they sell a bad stock it is always the possibility that after that it could go better (so they lost a gain).  On short time horizons, stocks are likely to have a worse performance than safer alternatives (e.g., bonds) because of the high variance of their returns.  People have also difficulties in understanding the meaning of portfolio. So another reason behind the equity premium bias is investors’ tendency to consider different assets (or financial tools) as separate rather than components of a single portfolio (Barberis, Huang, & Thaler, 2006).  Especially when they have stocks and bonds, they consider them separately and not put together the all the amount they are gaining or losing. They do not consider the portfolio as a hole. 72 achieved a very good standard of performance. The company decides to give you a bonus in addition to the regular monthly salary (Shefrin & Thaler, 1981):  This bonus could be paid by adding 500€ per month to the salary. Usually in this case the money goes in routine and everyday expenses.  This bonus could be paid out as a single amount at the end of the year. Usually people use almost of them for extraordinary expenses.  This bonus could be paid in company stocks that can be sold only after three years. Usually people can’t sell them for the first years and then they forget of them or they don’t know what to d with them and they leave them where they are. In this way they become savings for the long term  BONUS PAID WITH SALARY (500€ per month): In this case, those who receive the bonus are likely to integrate it with the mental account for running expenses (pay the loan, go out for dinner, buy gas for the car, groceries or clothes for the kids,…).  BONUS PAID AT THE END OF THE YEAR (6000€ in a lump sum): In this case, those who receive the bonus are likely to integrate it with the mental account for exceptional expenses (take a vacation a faraway place or remodeling the kitchen, any unspent money will be saved).  BONUS PAID IN STOCKS (6000€ invested and available after three years): In this case, those who receive the invested money are likely to integrate the bonus with the mental account for savings BEHAVIOURAL PORTFOLIO  Maslow idea is that when we satisfy the basic needs we don’t have satisfied the psychological needs. In the portfolio the idea is the same: so we try to preserve our wealth and not to loose money and then we try to use the money we gained (not the one of the downside protection but the surplus) money to take some more risk and to increase our wealth  Most investors have two main goals: 1. Downside protection: avoiding losing money or value. It is at the base of the pyramid and they are the investment that we make to avoid losing money and to fight inflation. 2. Upside potential: gaining money, increasing wealth. It is the top of the pyramid when we try to increase our money and investment. We try to speculate and get as much potential possible. When we do this we use additional resources that we have as surplus and we do not touch the base that continues to make the usual money.  Once investors protect their capital from potential losses, they can start freeing resources to invest on riskier assets.  Even when the goal is to increase one’s capital, a large part of the portfolio is invested in safer assets. These two goals are pursued independently and the investor can assess them using two different mental accounts, therefore reducing the complexity of portfolio diversification.  Appling that to the value function graph Downside protection:  Reference point: Current wealth  Goal: Avoiding losses  A loss corresponds to any outcome which is going to decrease current wealth. So we try to maintain our wealth. Upside potential  Reference point: Future wealth goal.  Goal: Achieving a gain.  A loss corresponds to any outcome which is lower than the desired future wealth goal (even a gain can be perceived as a loss!). 75  Using this approach to diversification it is possible to associate each layer of the pyramid to a different mental account and to different investment labels. For instance:  Bonds vs. stocks: bonds are considered safer that stocks that are on the upper part of the pyramid  Fixed returns vs. value growth: the firsts indicate the investment gives always the same performance and the latter are the ones that give more gain.  Domestic assets vs. foreign assets: the assets that are more familiar are considered better than the foreign assets that are considered riskier  Looking to this type of portfolio helps people to see it as dividend in chunks: this permits to look at them differently because they have different goals.  Of course, these labels must be interpreted as general hints and not as rules that can be generalized to all contexts of the market. For instance, Junk Bonds or BTP with high spread are very risky. However, large companies with a stable business and revenue stream are usually quite safe investments (e.g., Microsoft stocks)  Google stocks fluctuate a lot rather then mycrosoft but at the end looking to the 10 year context they gained 600% and mycrosoft only the 80%. So google is riskier in the short term, but more successful DISPOSITION EFFECT  The disposition effect (= an investment in our portfolio is going well we sell it immediately, and when the investment is not going well we keep it) has been discovered by Shefrin and Statman (1985) and it refers to how people manage the investments in their portfolio:  Investors usually sell to early the stocks or bonds that had recently achieved a positive performance.  At the same time, they keep for way too long stocks or bonds that had recently lost value. o Because of this pattern of choices, a large part of potential gains is lost when performance is good, while adding losses to the portfolio when performance is negative.  This is the opposite that we should do: if the investment is not going well for a while is better to sell it, saving all money possible to invest it on something else. But all the money that is invested is not real, we don’t really have them: so if we are gaining we think about the case in which everything goes bad and how bad we will feel in losing this money so we sell it to have the real money. If a investment is going bad we think that at the end of the day it is not real money so we feel less bad and we wait from a better performance that could happen in the future. For this reason there is a stock loss, that it is a set value from which when there is a loss they system automatically sell it.  The initial explanation proposed by Shefrin and Statman was based on:  Prospect theory: when there is a gain it is considered as a sure thing selling it and not selling it as a gamble, and when there is a loss if I sell it I have to cope with that but if I don’t sell it maybe there is a chance that it will recover.  Mental accounting: we think about all as separate.  PROSPECT THEORY - Based on this theory, people are risk averse when gains are involved, but risk seeking when losses are involved:  Investors sell “good stocks” because because they want to secure the gain.  They know the good performance can continue in the future, but prefer to sell the stock since future returns are not always predictable (the stock could start losing…).  On the other hand, not selling a “bad stock” for a long time exposes to the risk of losing more, but also keeps open the door for a recovery if the stocks starts gaining value  MENTAL ACCOUNTING - As we have seen before, investors do not look at the portfolio as a single entity, but they split it in several mental accounts:  When looking at the portfolio as a single entity, it might be easier to realize that selling a loss is not a problem because the portfolio is still in the black (it is gaining money).  This would make the loss more acceptable and easier to sell that stock. 76  Since different assets are assessed independently, the decision is made only focusing on the performance achieved by that asset (gain or loss).  Recently, Barberis and Xiang (2009, 2012) proposed a slightly different explanation of the disposition effect. This explanation involves two different psychological processes as well:  Realization utility: the positive or negative emotion associated with selling a specific investment. It involves how good I feel in selling something: if I sell a gain I feel smart and good because I gained, and selling a loss makes us feel dumb and bad because it could be the wrong decision because things could have changed in the future.  Present bias: the impulsivity leading to the overestimation of the effects, especially affective ones, of present performance compared with future ones. We prefer a smaller gain now than a larger gain in a long time and we prefer a bigger loss in the long time rather than a small loss now. This because the long time is not attached to a emotion in the present.  REALIZATION UTILITY - Investors derive utility not just from the final outcome but also from intermediate stages in which they buy and sell their investments.  Selling a “good stock” induces positive affect because investors feel good about the gain and are proud of their investment decision.  Selling a “bad stock” induces negative affect because of the loss and investors feel regret for their investment decision.  PRESENT BIAS - Because of impulsivity:  A “good stock” is sold quickly to experience positive emotions.  A “bad stock” is not sold to avoid experiencing negative emotions. OVERCONFIDENCE AND PORTFOLIO TURNOVER  Overconfidence depends mainly on people’s misjudgement regarding their skills in a specific task.  It’s not just investors who suffer from overconfidence. Among groups that overestimate their skills, there are: drivers and managers. In particular they found that 80% of them thinks to be better than the average.  In the financial domain, overconfidence is associated with an excessive portfolio turnover aimed at achieving a performance well above the average. Turnover is the rate at which investors modify their portfolios selling some assets and investing in others.  It is important to think to be able to invest, but it is risky. Investors are overconfidence when start investing in a time in which the market is going up, and it is lower when they start investing in a period where the market is going down.  Barber and Odean (2001) more than others demonstrated a link between overconfidence, excessive turnover, and negative portfolio performance. Particularly those who change less their portfolio they gain more, in fact every time they change the portfolio there are transaction costs, so they have to be sure that the portfolio has enough money to cover them. For this reason, more you change the portfolio lower is the performance.  Similar data have been found for investors who switch to online trading platforms: deciding alone rather than with the consultant make the people gain less money than before and they are not able to prevent loses when there is an economical crisis. ATTENTION GRABBING STOCKS  When investors must choose in which stock (or fund) to invest, they often focus on alternatives that grab their attention (for example the stocks that are going well or a name that recalls something familiar).  This is an intuitive strategy that helps reducing the difficulty of checking and analyzing a huge number of potentially viable investments.  Whether an information attracts our attention or not depends on how much emotion it makes us feel (see the slides on the availability heuristic).  Investors disproportionately choose stocks and funds that are advertised in specialty media, which are usually those who achieved the best performance in the recent past (Sirri & Tufano, 1998; Barber & Odean, 2008). 77