Reason, Passion, & Cognition, Week 2 Notes
Reason, Passion, & Cognition, Week 2 Notes 88-120
Popular in Reason, Passion, and Cognition
Popular in Social & Decision Sciences
This 6 page Class Notes was uploaded by Monica Chang on Saturday September 17, 2016. The Class Notes belongs to 88-120 at Carnegie Mellon University taught by Julie Downs in Fall 2016. Since its upload, it has received 23 views. For similar materials see Reason, Passion, and Cognition in Social & Decision Sciences at Carnegie Mellon University.
Reviews for Reason, Passion, & Cognition, Week 2 Notes
Report this Material
What is Karma?
Karma is the currency of StudySoup.
You can buy or earn more Karma at anytime and redeem it for class notes, study guides, flashcards, and more!
Date Created: 09/17/16
Week 2 LECTURE 3: NORMATIVE (PRESCRIPTIVE) THEORIES IN DECISION MAKING Optimal Decision-making: - Rational Actor Model o Assumptions: rationality, selfishness, capacity - Ideally, we should make simple decisions using the Weighted Additive Rule: 1. When making a decision between different options, find the attributes of these options that have significance and give each attribute a relative weight 2. Give each specific attribute of every option a value 3. Multiply each attribute value by its respective weight 4. Add these products together for each option to get each option’s overall value (for an example think back to the xbox versus playstation comparison from lecture 3) Expected Value: - Expected value: we can’t always use the weighted additive rule when making rational decisions because we often have uncertain outcomes, so we instead weigh each potential outcome by its probability (often used in gambling b/c they’re games of chance). - Expected value is the sum of the value of each outcome multiplied by their respective probabilities. If the outcomes are x, y, and z, expectedvalue=P x ∗x+P y)∗y+P (z∗z - Expected value is the average in the long-run - When deciding whether or not to make a gamble, we should consider: o how much the expected value is o whether it is a one-time thing or a long-term gamble, since the ways of losing could be greater than those of winning even though the expected value is positive - In gambling for the long-term, you can maximize profits by taking a gamble when EV>0 and not taking a gamble when EV<0. - But sometimes we can’t follow expected value… o For example, car insurance: Option 1 is paying $120 every month for car insurance. EV= -$120 Option 2 is not buying car insurance and having a 1% chance of paying $10,000 in one month. EV= -$100 o Even though we may have to pay more overall, we do not want to risk having to pay $10,000 in just one month Expected Utility: - St. Petersburg Paradox: Paul keeps tossing a coin until it lands heads up and says that he’ll give Peter $1 if it’s heads on his first toss, $2 if it’s heads on the second, $4 if third, etc., doubling each time. What should Paul pay to gamble? o Even though the expected value is infinite, it would make sense to pay a finite sum. o Bernoulli’s solution is to change the value of money to that of psychological utility - Expected utility - the psychological/subjective value put on an item (emphasis on subjective, not objective), it’s a measure of goodness/badness o Utility function relates objective value of outcomes to subjective values, the function is concave and proportional to log(x), the units have diminishing marginal utility because each unit increase in wealth has less worth than the one before - Would you pay $1000 to play a game with 50% of winning $2000 and 50% of winning $0? o No, because the expected utility of a sure payment of $1000 is more than that of playing the gamble: EU = (0.5)*U($2000)+(0.5)*U($0) < U($1000) Decision Theories: - Prescriptive/normative theories: what we SHOULD do o expected value – used to maximize $$ o expected utility – used to maximize utility - Descriptive theories: what we ACTUALLY do o do we actually maximize? o how is rationality actually measured in decisions? o In reality, assumptions of the rational actor model (rationality, selfishness, capacity), do not hold selfishness: people aren’t always selfish – philanthropy, sacrifice, etc. capacity: we have a limited mental capacity, there is a lot of capacity required to use the weighted average rule, and even then, you don’t know if you’ve considered all the alternatives, it is not realistic so maybe we shouldn’t always put in full capacity to decisions Bounded rationality: - Perfect rationality is limited by our access to information, human computational capacity in reasonable amounts of time, human memory - Bounded rationality: concerned with the difference between perfect rationality assumed in economics and realistic rationality, we satisfice as opposed to maximize - Satisficing means that we find good enough solutions b/c it is more efficient - Analogous to how optical illusions (e.g. Muller-Lyer Illusion) can help us understand how our brains see the world and how we can become more effective, deviations from rational model can help us explain/understand decisions Key ideas: - Rational actor model: normative, economic model for optimal/maximizing decision-making - We can be rational with uncertain outcomes: o expected value to maximize $$ o expected utility to maximize psychological utility - Analogous to optical illusions, we deviate from the rational actor model in reality (bounded rationality) NOTES on optional reading from The New Yorker: “Select All” by Christopher Caldwell. Key ideas: - Paradox of choice, unlimited choice leads to suffering, mischoice, regretting a decision because you realize there were many other options - Contrary to classical economist theories, we are not close to being rational “utility maximizers” - Maximizers – those who tend to go through all options before deciding on the best one -- tend to be more miserable, a lot of firms have instead started to “satistfice” – going with a good enough choice - A lot of the time, more choices will lead people to not buy things, de-motivation, and the paradox of choice not only applies to physical purchases but also marriage, careers, etc. - So what are solutions? o Changing the way we choose (satisfice instead of maximize?) o Changing the number of choices (self-binding? Or have others choose for us?) - Some New Economy innovations are limiting choice - On a more positive note, there is this phenomenon called the “hedonic treadmill (hedonic adaptation),” which says that humans are resilient b/c they can go back to a relatively stable state of happiness regardless of negative or positive changes in life, so it is hard for choices to defeat us LECTURE 4: BEHAVIORAL SCIENCE The Scientific Method: - The Scientific Method: a process of techniques for correcting old knowledge and getting new knowledge that follow these steps: 1. Questions of interest 2. Make testable hypothesis 3. Choose research method & design study 4. Get data 5. Analyze & make conclusions from data 6. Report findings Formulating a hypothesis: - Theory: set of ideas that can explain a natural phenomenon - The Law of Parsimony: simplest theories tend to be best - *Hypothesis: a falsifiable prediction that describes a relationship between at least 2 variables that are stated/implied to be measurable o Hypothesis structure: “If our theory is true, then we should observe our predicted outcome.” e.g. If plants use photosynthesis to grow, then a lack of sunlight should decrease plant growth. Falsifiable: If plant growth is normal, our theory is wrong. - An operational definition will specify how a variable is measured with a precise measurement/unit. e.g. the operational definition for intelligence can be IQ, memory task, etc. - Steps: o You will have a question of interest o Formulate a hypothesis o Identify the variables and operationalize them (define how they’ll be measured) Research design: - CORRELATIONAL STUDIES: a way to find how variables are related, does not imply causality o Correlation coefficient is between -1 and 1 Perfect correlations are close to either -1 (negative association) or 1 (positive association) Weak correlations have values closer to 0 - EXPERIMENTAL STUDIES: a way to find causal relationships where you change one variable and measure another variable o Natural/observational experiments are where environmental changes are observed to observe effects on behavior o Laboratory experiments are where one variable is manipulated and its effect on another variable is tested Experimental group exposed to manipulation and control is not Independent variable is manipulated, dependent variable is measured, scientific method allows us to infer that changes in the independent variable cause changes in dependent We have to give random assignments when manipulating the independent variable, otherwise validity is questionable Data Interpretation: - Differentiate between the expected average and how each person is unique (central tendency and variability) o Use this info to determine whether independent variable truly caused changes in dependent variable or if it was just chance of variability (p-value) - When you have multiple independent variables… o One main effect: only one variable can have a significant effect o Two main effects: both variable may have their own significant effects o Interaction effect: effect of one variable depends on the other (no main effects) Key ideas: - The Scientific Method o Understand basic process o Ideas to evidence, allows us to reject wrong ideas (falsifiable) - Research design o Correlational studies: way to find how variables are associated (does not imply causality) o Experimental studies: way to find causal relationships where you change one variable and measure another variable - Interpreting Data: o Central tendency and variability o Main effects and interaction effects w/ independent variables
Are you sure you want to buy this material for
You're already Subscribed!
Looks like you've already subscribed to StudySoup, you won't need to purchase another subscription to get this material. To access this material simply click 'View Full Document'