Decision Making and Reasoning - Cognitive Psychology - Lecture Slides, Slides of Cognitive Psychology

Download Decision Making and Reasoning - Cognitive Psychology - Lecture Slides and more Slides Cognitive Psychology in PDF only on Docsity! Cognitive Psychology Notes 13 REASONING AND DECISION MAKING Where We Are  We’re continuing our tour of higher cognition. We’ve covered:  Categorization  Language—Structure  Language—Meaning  And we continue with:  Reasoning/Decision making  Human factors Docsity.com Logic  Here’s a rule:  If there’s fog, then the plane will be diverted.  Here are four tests of this rule:  Present a situation where there’s fog.  Present a situation where there’s no fog.  Present a situation where the plane has been diverted.  Present a situation where the plane has not been diverted. Docsity.com Logic  The rule is:  If there’s fog, then the plane will be diverted.  I’ve added outcomes to the four tests:  Present a situation where there’s fog, find that the plane has been diverted.  Present a situation where there’s no fog, find that the plane has been diverted.  Present a situation where the plane has been diverted, find that there’s no fog.  Present a situation where the plane has not been diverted, find that there’s no fog.  Evaluate the tests and put a T or F on your paper for each. Docsity.com Logic  The rule is:  If there’s fog, then the plane will be diverted.  The four tests:  Fog, diverted.  No fog, diverted.  Diverted, no fog.  Not diverted, no fog.  The tricky part is that only two of these tests are valid ways to check out the hypothesis. Two of them have no bearing on the hypothesis. So, your answers should have been:  T  I can’t know  I can’t know  T Docsity.com Logic  How do you know what the correct answer should be? Consider this problem:  If the red light appears, then the engine is overheating. Two valid tests: ○ Modus ponens:  The red light appeared.  Therefore the engine is overheating. ○ Modus tollens:  The engine is not overheating.  Therefore, the red light must not have appeared. Docsity.com Logic  How do you know what the correct answer should be? Consider this problem:  If the red light appears, then the engine is overheating. Two invalid tests: ○ Denying the antecedent:  The red light did not appear.  Therefore, the engine is not overheating. ○ Affirming the consequent:  The engine is overheating.  Therefore, the red light appeared. Docsity.com Logic  In general terms:  If p then q: ○ Modus ponens:  p.  Therefore q. ○ Modus tollens:  not q.  Therefore, not p. ○ Denying the antecedent:  not p.  Therefore, not q. ○ Affirming the consequent:  q.  Therefore, p. Docsity.com Logic  Let’s go back through the other examples and figure out why the correct answers are what they are… Docsity.com Logic  The interesting cognitive question is: Why are people so bad at this?  Illicit conversion. People tend to reverse the order of the terms or make it into a biconditional problem. They read “if p then q” as “if p then q and if q then p.” However, the order matters. Except for writing things down and being careful, there’s not a tip to avoid this kind of confusion. Docsity.com Logic  The interesting cognitive question is: Why are people so bad at this?  Illicit conversion. Consider this: ○ If you smoke, then you will get cancer.  Valid: ○ I smoke and didn’t get cancer, so it’s wrong.  Invalid: ○ I got cancer and I never smoked, so it’s wrong.  If you turned it around to: If cancer, then smoked, the invalid test becomes valid. Docsity.com Logic  Another interesting cognitive question is: Why are people so good at some conditional reasoning problems? If you’re under 21 then you should not be drinking alcohol. 18 coke 43 beer Docsity.com Logic  Most people guess 18 and beer with no problem. Why? One hypothesis is contextual support. Another is that you have evolved an ability to detect cheaters and are good at permission situations. 18 coke 43 beer Docsity.com Logic  The Wason selection CogLab examined this, let’s turn to that now… Docsity.com Heuristics  Some implications of representativeness:  Lottery play. Does 1 2 3 4 5 6 seem like a good set of numbers to play? Most people think it’s a bad choice because it’s “so unlikely.” But, every set of numbers is equally likely. ○ If you’re thinking about playing, ask yourself if you would play 1 2 3 4 5 6. If the answer is no, you understand the odds and shouldn’t play. ○ But, 1 2 3 4 5 6 is actually very representative of numbers other people won’t play, which means a lot of people do play it, and that makes it a bad choice (you’ll split the pot with more people, decreasing the expected value of the lottery payoff). Numbers > 31 are also bad due to representativeness. Docsity.com Heuristics  Some implications of representativeness:  Stereotypes. If something you see is representative of a stereotype you are more likely to notice it and add it as evidence (especially with confirmation bias). Docsity.com Heuristics  Availability heuristic. When you decide how likely something is, think of an example, and base your estimate on how hard it is to do that.  Are there more words that begin with a k or have a k as the third letter? Docsity.com Heuristics  Availability is influenced by a lot of factors that should be unsurprising to people finishing a cognitive class:  Frequency: More frequent = more available.  Familiarity: More familiar = more available.  Vividness: More vivid = more available.  Recency: More recent = more available.  How could these influence people’s thinking that driving is safer than flying? Docsity.com Heuristics  Simulation heuristic. Ease of simulation influences people’s judgments.  Two men are on flights that leave at the same time and are riding in the same car. They arrive 1/2 hour late. Mr. Crane’s flight left on time, but Mr. Tee’s flight was delayed and only left five minutes ago. Who is more annoyed at missing their flight? Docsity.com Heuristics  Influences on simulation:  Undoing. People usually file down unusual details to make the sequence of events more typical (downhill change) rather than add details (uphill change) when simulating events. ○ George decides to leave work early. When he gets to the parking lot he has a flat tire and stops to change it. Still in a good mood, he decides to take the scenic drive home even though that will take a little longer. He stops at the store on the way. As he is nearing his house, his car is hit by a drunk driver running a red light and he is killed. If only… Docsity.com Heuristics  John S. is a supervisor in a local manufacturing firm. John is responsible for promoting the employees in his department. In the past he has been accused of being against equal rights and opportunities for women. There are 10 male and 90 females in his department who are potential candidates for promotion. John decides to give these employees a written examination to help with his decision. John grades these exams himself, and reports that the highest mark was obtained by a man, whom he promotes.  How suspicious are you that John’s grading of the exam was unfair? (Write 1-100, with 100 being very suspicious). Docsity.com Heuristics  Ease of simulation makes the first one sound more suspicious. Docsity.com Additional Influences  Anchoring and adjustment. People tend to start from the first part of the problem (the anchor) and then adjust from there. If you start with a high anchor people tend to go high and vice versa. Docsity.com Additional Influences  Anchoring and adjustment examples:  One half multiply 1 X 2 X 3 X 4 X 5 X 6 X 7 X 8 Docsity.com Additional Influences  Anchoring and adjustment examples:  One half multiply 8 X 7 X 6 X 5 X 4 X 3 X 2 X 1  One half multiply 1 X 2 X 3 X 4 X 5 X 6 X 7 X 8  The median for the first problem was 2250, for the second it was 512 (Tversky & Kahneman, 1974). The answer is 40320. Docsity.com Additional Influences  Anchoring and adjustment examples:  Two lotteries: ○ 50% red marbles in a bag, 50% white. You try to draw a red marble. ○ 90% red marbles, 10% white. You try to draw 7 red marbles in a row. ○ Which gives the best chance of winning? ○ They’re about equal. Why do people prefer one over the other? Docsity.com Additional Influences  Set/Fixedness examples.  Fixedness. Solve these jug problems: Problem: Jug A: Jug B: Jug C: Target: 1 21 127 3 100 2 14 46 5 22 3 18 43 10 5 4 15 39 3 18 Docsity.com Additional Influences  Set/Fixedness examples.  Functional fixedness. You’re in a plane crash in the desert. You have the following items: A parachute, a map, a compass, and a pocket mirror. What is your most valuable asset? Docsity.com Additional Influences  Confidence. Generally people are more confident in their answers to general knowledge questions than they are correct. The more confident, the more they are overestimating their ability.  Hirsute probably means either “really hairy” or “habitually late.” Pick one and rate your confidence. Docsity.com Additional Influences  Let’s look at the CogLab exercise on decision making… Docsity.com Probability  People have pretty poor comprehension of probability, and rarely use it well in their day-to-day reasoning (e.g., Deal or No Deal).  It’s only sort of related, but let’s look at the Monty Hall problem. Docsity.com Probability  Conjunction. How do people combine the probabilities of independent events?  Try the “causes of death” estimate…  Not being killed in a car accident may be 99% certain, not perishing in a household accident is 98% certain, not dying of lung disease is 95% certain, dementia 90%, cancer 80%, heart disease 75%. What is the chance of not dying from any of them? Docsity.com Probability  Conjunction fallacy.  Try the two conjunction exercises…  Linda is 31 years old, single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply concerned with discrimination and social justice, and participated in anti-nuclear demonstrations. Which is more likely: ○ Linda is a bank teller. ○ Linda is a bank teller and a feminist. Docsity.com Probability  Conjunction fallacy.  Try the two conjunction exercises…  Health survey of a random sample of adults, including Mr. F. Which is more probable: ○ Mr. F has had one or more heart attacks. ○ Mr. F has had one or more heart attacks and is over 55. Docsity.com Probability  Conjunction fallacy.  Try the two conjunction exercises…  Linda is 31 years old, single, outspoken, and very bright. She majored in philosophy. As a student, she was deeply concerned with discrimination and social justice, and participated in anti-nuclear demonstrations. Which is more likely: ○ Linda is a bank teller. ○ Linda is a bank teller and a feminist. Docsity.com Probability  Conjunction fallacy.  The probability of the conjunction of two independent events has to be smaller than the probability of either one of them. People usually get that wrong, influenced by typicality.  We can look at our CogLab exercise for typical reasoning… Docsity.com Probability  Perceptions of randomness.  “People think a sequence is more likely, and hence random, if there is some irregularity in order of appearance (e.g., HHTHTH vs. HTHTHT).  “People think a sequence is more likely, and hence random, if the equiprobable outcomes occur equally often.  “The outcome alternation rate (i.e., how often H switches to T and vice versa) that people consider to be random is higher than that associated with chance.” (Hahn & Warren, 2009, p. 454) Docsity.com Probability NUN NA WL AANA ARE H TH T HT 5 H 2 I 4 I 6 Z 8 9 10 12 13 14 1 16 Docsity.com Hla ar larren (2009, p. 455) Probability a © = 2 2 o s TO vu L U Vv Oo. x rea l l l l l l TTTH TTHT TTHH THTT THTH THHT THHH HITT HTTH HTHT HTHH HHTT HHTH HHHT HHHH H/T sequence Warren (2009, p. 456) Docsity.com Probability  The power of chance. People often underestimate the power of chance.  Out of 1000 stock picking professionals, person A has picked correctly which direction the market would move 10 weeks in a row. Is that an impressive record? Docsity.com Probability  Even if they’re only flipping a coin, we would expect by chance:  1000  500 (week one)  250  125  62  31  16  8  4  2  1 (week ten) Docsity.com Probability  Implications of chance:  Amazing coincidences: Twins. Wyatt, Posey, Welker, and Seamonds (1984) studied pairs of random people. You get lots of pairs like these. For any one thing the odds of a match might be low, but with an unlimited number of variables to choose from, the odds of some overlap are quite high. Docsity.com Probability  Implications of chance:  The power of prediction. With the biases we discussed above, it’s easy to interpret an outcome as a confirmation of a prediction. You must be careful of probability and coincidence. An example: ○ I predict two people in here have the same birthday. Let’s see. Docsity.com Probability  Implications of chance:  The power of prediction. Most people do the math wrong, and it seems amazing. In fact, with 23 people in the room the odds are 51% that two will share a birthday. In other words, I have an even chance of being right. In general, if I know more about probability than you do, I can come off as an amazing psychic. Docsity.com Probability  Certainty. Add these up:  1000  40  1000  30  1000  20  1000  10 Docsity.com Probability  Certainty. You got 5000. The actual sum was 4100. Almost 100% of the population makes this mistake, it’s not exciting. Docsity.com Probability  Certainty. I have some additional demonstrations if there’s time… Docsity.com