Decision Making

Question 1

Bayes’s Theorem

Answer 1

Prescriptive mathematical model that explains how human should act.
Calculates probability that hypothesis is true after considering evidence

Question 2

Conditional probability

Answer 2

Probability that evidence is true, if hypothesis is true (in Bayes’s Teorem

Question 3

Descriptive model

Answer 3

Explains how humans actually make decisions (with biases and common human errors)

Question 4

Framing effects

Answer 4

The way info is presented affects the decision making among equivalent alternatives (depicted as gain or loss, no clear basis for choice)

Question 5

Gambler’s fallacy

Answer 5

Likelihood of event increases with amount of time since it last occurred (“law of averages” is an illusion)

Question 6

Posterior Probability

Answer 6

Probability that hypothesis is true after considering evidence. Combination of prior- and conditional probability

Question 7

Prescriptive model

Answer 7

How people should act (norms, rational choices, ideal robot)

Question 8

Prior probability (base rate)

Answer 8

Probability that hypothesis is true, before taking evidence into account

Question 9

Probability matching

Answer 9

Choosing among alternatives with the perceptive of succes in previous choices

Question 10

Recognition heuristic / Availability heuristic

Answer 10

If people are presented with two choices and only recognize one item, they tend to give it higher value (eg. American students think the city they recognize the name of is biggest –> often correct!)

Question 11

Subjective probability

Answer 11

Personal estimate of how likely an event is to occur (based on intuition and experience, not calculations)

Eg. Many prefer 1% chance of $400 over 2% chance of $200, because 1% is subjectively represented as more than half of 2%.

Question 12

Subjective utility

Answer 12

Value of things are individual for each person (depends on risk preferences)

Question 13

Risk aversion

Answer 13

A person values certainty (common) –> Choose option A) 500 DKK for sure instead of option B) 50% to get 1000 DKK

Question 14

Risk seeking

Answer 14

A person likes risks and uncertainty (uncommon) –> Choose option B) 50% to get 1000 DKK instead of option B) 500 DKK for sure

Question 15

Ventromedial prefrontal cortex

Answer 15

Area in the very front of the brain that is important in succesful judgment via reflective processing: Probability of receiving reward?
It is also crucial for personality, motivation, emotional regulation, social sensitivity.

Damage in this area make you socially incompetent and unable to learn from mistakes and judge whether it is a good or bad decision

Question 16

Base-rate neglect

Answer 16

People often fail to take into account how likely the scenario is in general –> leads to fundamental error in probability estimate

Question 17

Conservatism

Answer 17

Tendency to underestimate the full force of available evidence

Question 18

EVT: Expected Value Theory

Answer 18

Prescriptive model of expected value calculated as probability of outcome times value of outcome. This model assumes that subjective value is equal to objective value (which often isn’t the case)

Question 19

EUT: Expected utility theory:

Answer 19

Descriptive model, taking risk preference into account. Described with the utility function of subjective utility (risk aversion/-seeking)

Question 20

Prospect theory

Answer 20

Descriptive model. Adding endowment and framing to EUT. Described with the value function with a steeper loss-line because of loss aversion(it hurts more to loose 100 kr than it feels good to win 100 kr)

Question 21

Endowment

Answer 21

When you owe something, you often give it more value than the actual price (eg. Marc’s beanie ;))

Question 22

Justification

Answer 22

When choosing among many similar alternatives we choose the one easiest to justify

Question 23

Similarity effect (Irrational effect of adding a third option)

Answer 23

If two options are similar to each other, we tend to choose the third one.
–> Difficult to justify choosing between the similar ones

Question 24

Attraction effect (Irrational effect of adding a third option)

Answer 24

If one option beats the other option in a specific factor, we are likely to value that factor highest all of a sudden and go with the option that is dominant in that factor.
–> Easy way to justify: 1-1 comparison

Question 25

Compromise effect (Irrational effect of adding a third option)

Answer 25

If 3rd option make one option look like a compromise between the other two (middle in more parameters), we tend to choose this. --> Easy to justify: better than both other options in one factor

Question 26

FRN: Feedback-related-negativity

Answer 26

Brain signal that reflects how we evaluate outcomes related to our expectations (if dopamine-reward is less than expected they measure increased negative response. If dopamine-reward is greater than expected, they measure increased positive response) It is typically more negative than positive (loss aversion)

Question 27

Nucleaus accumbens

Answer 27

Part of basal ganglia that is related to reflexive/automatic processing: How big is the reward?

Question 28

Reinforcement learning

Answer 28

Mechanism for learning through experience what action to take. Eg. in FRN-study: Late reaction to expected/unexpected information (even when told what is the correct answer), because of slow learning process. Anderson: "As if the minds knew, but the hearts had to learn".

Question 29

What is reference dependence?

Answer 29

The value function in prospect theory emphasizes that people evaluate outcomes relative to a reference point (current status), not absolute value.

Question 30

Kahneman & Tversky (1973) - Engineer-High/Engineer-Low Study

Answer 30

Setup: Two groups told person chosen from either 70 engineers/30 lawyers OR 30 engineers/70 lawyers Method: Given personality description of "Jack" (conservative, careful, enjoys carpentry and math puzzles) Results: Both groups estimated .90 probability Jack is engineer, ignoring different base rates Control condition: Uninformative description of "Dick" led both groups to estimate .50 (should have stayed with base rates) Demonstrates: Base-rate neglect

Question 31

In a win frame, people tend to be…?

Answer 31

Risk averse

Question 32

In a loss frame, people tend to be…?

Answer 32

Risk seeking

Question 33

Edwards (1968) - Poker Chip Experiment

Answer 33

Setup: Two bags with 100 chips each: Bag 1 has 70 red/30 blue, Bag 2 has 70 blue/30 red Task: One bag chosen randomly, participants sample chips with replacement to determine which bag Example: After drawing one red chip, Bayesian calculation shows posterior probability of red bag = .70 Results: Participants estimated only .60 (conservative revision) After 12 draws (8 red, 4 blue): Correct Bayesian probability = .97, participants estimated ≤.75 Demonstrates: Conservatism

Question 34

Which bias causes overreaction to new evidence?

Answer 34

Base rate neglect

Question 35

Which bias causes underreaction to new evidence?

Answer 35

Conservatism

Question 36

Gluck & Bower (1988) - Disease Diagnosis Study

Answer 36

Setup: 256 fictitious patients with 1-4 symptoms (bloody nose, stomach cramps, puffy eyes, discolored gums) Design: One disease had base rate 3x the other; conditional probabilities varied by disease Task: Participants diagnosed disease for each patient, received feedback Key manipulation: Participants NOT told base rates or conditional probabilities - had to learn from experience Results (implicit): Frequency of diagnosing rarer disease matched true Bayesian probabilities almost perfectly Results (explicit): When asked directly about symptom-disease associations, showed base-rate neglect (overestimated rarer disease) Demonstrates: Implicit Bayesian behavior through probability matching, even when explicit judgments show bias

Question 37

Which two brain areas are central to decision-making processes?

Answer 37

The Basal Ganglia and the Ventromedial Prefrontal Cortex

Question 38

What is the main function of the Basal Ganglia in decision making?

Answer 38

Tracks subjective utility - how valuable or rewarding something feels to the individual.

Question 39

What does activity in the nucleus accumbens represent in the Basal Ganglia?

Answer 39

The magnitude of reward - it responds more strongly to high magnitude rewards

Question 40

Tversky & Kahneman (1974) - Letter K Position Study

Answer 40

Task: Estimate proportion of English words beginning with K vs. having K in third position Method: Participants recall words from memory in each category Results: Estimated 2x as many words begin with K Reality: 2x as many words have K in THIRD position (opposite of estimate) Explanation: Words more strongly associated with first letter; spreading activation makes them more available in memory Demonstrates: Memory availability biases probability judgments; easier to recall = overestimated frequency

Question 41

Shuford (1961) - Visual Array Proportion Study

Answer 41

Stimulus: Arrays of vertical and horizontal bars shown for 1 second Proportions: Varied from 10% to 90% vertical or horizontal Task: Estimate proportion of each type of bar Results: Very accurate estimates close to true proportion Contrast: Accuracy when information directly visible vs. must rely on memory Demonstrates: People reasonably accurate at frequency judgments with direct perceptual access

Question 42

Which brain area tracks the probability of reward? (higher activation for higher likelihoods)

Answer 42

The Ventromedial Prefrontal Cortex

Question 43

Goldstein & Gigerenzer (1999, 2002) - German Cities Study

Answer 43

Participants: University of Chicago students Task: Judge which German city is larger (e.g., Bamberg vs. Heidelberg) Pattern: Students almost always picked recognized city when they knew one but not the other Key finding: MORE accurate when they recognized only one city than when they recognized both Explanation: Recognition correlates with city size → frequency of newspaper mentions → probability of recognition Comparison: American students better at German cities than German students (who recognize all cities); German students better at American cities Demonstrates: Recognition heuristic is adaptive strategy, not judgment error

Question 44

Richter & Späth (2006) - Animal Population Study

Answer 44

Questions: "Are there more Hainan partridges or arctic hares?" and "Are there more giant pandas or mottled umbers?" First case: Recognition heuristic leads to correct choice (arctic hares) Second case: People recognize pandas but correctly choose mottled umbers (moth) because they know pandas are endangered Demonstrates: People intelligently combine recognition heuristic with other relevant knowledge

Question 45

Kahneman & Tversky (1984) - Asian Disease Problem

Answer 45

Setup: 600 people expected to die from disease, choose between two programs Frame 1 (lives saved): A: 200 saved for certain vs. B: 1/3 chance 600 saved, 2/3 chance none saved Result 1: 72% chose A (certainty) Frame 2 (lives lost): C: 400 die for certain vs. D: 1/3 chance nobody dies, 2/3 chance 600 die Result 2: Only 22% chose C (equivalent to A) Explanation: Negatively accelerated utility function - 600 lives < 3x utility of 200 lives; 400 deaths > 2/3 utility of 600 deaths Demonstrates: Framing effects - equivalent choices judged differently based on gain vs. loss framing Real-world extension: McNeil et al. (1982) found doctors' treatment choices affected by odds-of-living vs. odds-of-dying framing

Question 46

Shafir (1993) - Child Custody Study

Answer 46

Setup: Jury decides sole custody after messy divorce Parent A: Average on all dimensions (income, health, work, social life, relationship with child) Parent B: Extreme positives (very close with child, above average income) AND negatives (health problems, lots of travel, extremely active social life) Award custody question: 64% chose Parent B, 36% chose Parent A Deny custody question: 55% chose to deny Parent B, 45% chose to deny Parent A Paradox: Same parent chosen when asked to award AND when asked to deny Explanation: Parent B's description offers reasons both FOR (close relationship) and AGAINST (travel); choice based on what's easiest to justify Demonstrates: Justification-based decision making when no clear basis for choice

Question 47

Greene, Sommerville, Nystrom, Darley, & Cohen (2001) - Trolley Dilemma fMRI

Answer 47

Dilemma 1 (impersonal): Hit switch to divert trolley, killing 1 person instead of 5 Dilemma 2 (personal): Push large stranger off bridge to stop trolley, killing 1 person instead of 5 Results: Most people willing to act in Dilemma 1, unwilling in Dilemma 2 Brain imaging - Impersonal: Parietal cortex regions associated with cold calculation active Brain imaging - Personal: Regions associated with emotion active (including ventromedial prefrontal cortex) Demonstrates: Different framings engage different brain systems; emotional vs. cognitive processing affects moral decisions

Question 48

Johnson, Steffel, & Goldstein (2005) - Opt-In/Opt-Out Studies

Answer 48

Domain 1 - Organ donation: Much higher donation rates in opt-out countries than opt-in countries Domain 2 - 401(k) plans (Choi et al., 2003): Higher enrollment when must opt-out than opt-in Domain 3 - Flu shots (Chapman et al., 2010): More likely to get immunization when must opt-out than opt-in Explanation: People avoid having to justify decision to opt-in or opt-out, so accept default Demonstrates: Powerful framing effect with real-world policy implications

Question 49

Knutson, Taylor, Kaufman, Peterson, & Glover (2005) - Reward Magnitude & Probability fMRI

Answer 49

Design: Participants presented with gambles varying in magnitude (e.g., 50% chance of $5 vs. $1) and probability (e.g., 80% vs. 20% chance of $5) Nucleus accumbens (basal ganglia): Activity reflected different MAGNITUDES of rewards Nucleus accumbens: Did NOT reflect different probabilities of rewards Ventromedial prefrontal cortex: Activity reflected different PROBABILITIES of rewards Ventromedial prefrontal cortex: Did NOT reflect different magnitudes of rewards Demonstrates: Dissociation in neural processing - nucleus accumbens codes reward magnitude, ventromedial prefrontal cortex integrates probability information

Question 50

Bechara, Damasio, Damasio, & Anderson (1994) - Iowa Gambling Task

Answer 50

Task: Choose from 4 decks; A & B give big rewards (+$100) but bigger penalties (-$1,250); C & D give smaller rewards (+$50) and smaller penalties (-$250) Normal participants: Learn to prefer C & D; show emotional response to risky decks Ventromedial PFC damage: Keep choosing A & B; no emotional response Shows: Ventromedial PFC essential for avoiding immediate rewards with poor long-term outcomes

Question 51

Olds & Milner (1954) - Rat Self-Stimulation

Answer 51

Setup: Electrodes near basal ganglia; rats press lever for stimulation Result: Pressed to exhaustion; caused dopamine release in nucleus accumbens Connection: Same system activated by heroin/cocaine Shows: Dopamine system is fundamental mammalian reward/motivation system

Question 52

Schultz (1998) - Monkey Dopamine Recording

Answer 52

Unexpected reward: Dopamine activity at reward delivery Predicted reward: Activity transfers to predictive stimulus (not reward) Omitted reward: Depressed activity at expected reward time Interpretation: Dopamine codes prediction error (difference between actual and expected) Example: Why second meal at new restaurant not as good Shows: Dopamine tracks expectation violations, not absolute reward

Question 53

Gigerenzer & Hoffrage (1995) - Mammography Framing

Answer 53

Probability format: "1% have cancer, 80% positive if cancer, 9.6% positive if no cancer" → <20% correct Frequency format: "10 of 1,000 have cancer, 8 of 10 with cancer test positive, 95 of 990 without test positive" → ~50% correct Correct answer: ~8% probability of cancer given positive test Shows: People reason better with frequencies than probabilities