CW Decisionmaking

SUBJECTIVE EXPECTED UTILITY THEORY (SEUT)

• Calculate the expected utility (value) of each option (outcome) x probability it will actually happen (likelihood)

  • Add each possible option/outcome into equation and calculate which has highest probability

• Example: Value of getting better grades x probability that staying home will improve grades = choosing option that has highest expected utility for you

  • Should I buy lottery tickets?

    • Cost of a ticket is low utility

    • Winning millions is high utility but also high uncertainty

• SEUT seems rational but people don't always make decisions that way

Kahneman and Tversky

• Presented people with gambles

• Published in economics and psychology papers

  • Changed the way we assumed economics and rationality worked

• Started behavioral economics

Loss aversion

• we feel approx. 2x stronger about losses than gains

  • Losses have greater weight than potential gains in making decisions

Prospect theory

•  describes how people value gains and losses

  • Function is concave for gains

    • Reflects diminishing marginal value and risk aversion

  • Function is convex for losses: reflects risk seeking

  • Loss aversion: implies losses have greater weight than potential gains in making decisions

    • Reflected in steeper curve for losses

Tversky and Kahneman (1981) risk aversive vs risk seeking

• Imagine US is prepping for an outbreak of a disease expected to kill 600 people

  • 2 alt programs to combat disease have been proposed

    • Program A: 200 saved

    • Program B: 33% chance that all 600 saved, 67% chance no one will be saved

  • Most people choose: A - risk aversive

• Same experiment, new programs

  • Program A: 400 saved

  • Program B: 67% that 600 die, 33% no one dies

  • Most people chose: B - risk seeking

• Same question framed different ways

why do people not maximize utility in decisionmaking

• We feel more strongly about losses > gains

• We opt for certainty and certain gains

• We prefer to take risks with losses

• The way a situation is framed can influence choices

Framing effects and preference reversals

• When 2 options are identical, people make difference decisions depending on how outcomes are framed

• People tend to be risk aversive when the outcome is framed as gain, and risk seeking when the outcome is framed as a loss

• Kahneman and Tversky (1984): violates the principle of invariance

  • Peoples choices should depend on the situation, not on the way its framed/described

Prospect theory

• gains and losses are calculated from a reference point

  • We think about changes at points on the curve rather than absolute values

• Predicts many aspects of human decision making

  • Such as why we often don't maximize utility

• Theory is primarily descriptive

  • Explains what decisions people make

  • Not why

• Doesn't explain ind differences in preference for risk

Endowment effect Kahneman et al (1990)

• Given a mug, how much do you sell it back to the giver for

  • Randomly assigned participants as buyers or sellers

    • Sellers demand more than buyers are prepared to give

    • Ex. Housing market

      • House prices fall, may fall more, homeowners reluctant to put their homes on the market

      • Homes stay on the market because there's a discrepancy between what people offer/accept

• Endowment effect and prospect theory

  • Gains/losses calculated using a reference point

  • People care about change from this reference point rather than absolute values

  • Sometimes reference point shifts (ex. We obtain something) and leads to different decisions

Disjunction effect and Sure thing principle

• Sure thing principle

  • If we prefer X to Y in any state of the world then we should prefer X to Y when the state of the world is uncertain

  • Disjunction effect shows that sometimes people violate this principle

• Disjunction effect

  • People look for reasons or arguments to support their decisions

  • People may make the same decision but for a different reason

Reason based choice

• Reason based choice suggests that we will be more likely to choose an opt when we have a compelling reason for that selection

• Search for alts should occur when a compelling reason for choice is not available

Heuristics

• - a short cut rule of thumb for making judgements

• Often produce the 'right' answer but sometimes lead to biases

  • A systematic error in judgement relative to some normative standard

Intuition vs. logic

• System 1 uses heuristics

  • Quickly produces intuitive answer

• System 2 requires time and working memory to slowly work out a logical answer

conjunction fallacy

• Most people think that its more likely that Linda is a 'feminist' bank teller than a bank teller because she is more similar to that category

  • The probability is smaller though

• People judge whehter someone/something belongs to a category by judging the similarity of that to a stereotypical member of that category

  • Linda is more representative of a feminist bank teller than a bank teller

    • So judged to be more likely

    • Unordered lottery number sequence more typical of a winning lottery ticket

Availability heuristic:

•  the likelihood of events is judged based on the ease with which instances come to mind

  • youre going on holiday to the beach

    • More likely to kill you:

      • Coconut falling

      • Shark attack

      • Shark attack is less likely but representations are more available in memory (media) thus the brain takes that a s cue for it to be believed to be more relevant

Gambler's fallacy:

• the belief that after a streak of events (ex heads) the opposite becomes more likely (ex tails)

• Representativeness in gambling

  • We expect a non random sequence to be as close to an alteration rate of 0% as possible

• Realistically, a random sequence is most likely to have an alt rate of 50%

making judgements

• We find formal probability hard

• We tend to use heuristics like representativeness, availability, and anchors

• We are also very good at recognizing patterns and using them to make predictions

Chapman and Johnson (2002) anchoring

• Anchoring effect: Exposure to irrelevant numerical anchors influences judgments and estimations.

• Anchoring involves starting from an initial value (anchor) and adjusting to reach a final estimate, but adjustments are often insufficient.

• "Anchoring" can refer to:

  • Presenting a salient but uninformative number to subjects (anchoring procedure).

  • Experimental results where the uninformative number influences judgments.

  • The psychological process by which the uninformative number affects judgments.

Function of reasoning

• Allows us to apply our knowledge to new situations

• Enables us to think hypothetically make plans for the future, consider the consequences of our actions

• Enables us to evaluate past events and possibilities - think counterfactually

• Enables us to understand the meaning of what others say

Inductive reasoning

• Reasoning from specific cases to general rule

• Find a pattern in some cases and use it to make a prediction

  • Ex. All the swans I see are white -> all swans are white

Abductive reasoning

• Generating an explanation

  • Ex. Theres glass on the floor -> someone dropped a glass

Deductive reasoning

• Occurs when the conclusion necessarily follows from the premises

  • If the premises are true the conclusion must be true

Propositional reasoning

• A proposition is a statement that can be true or false

• Involves drawing inferences based on the relation between propositions

  • Ex. Relations such as if and or

mental models theory (johnson-laird (1983), J-L and Byrne (1991))

• People construct models to represent the premises

• Due to working memory limitations, they often represent only a single model initially

• Evidence: the difficulty of drawing an inference is related to:

  • Whether an inference can be drawn from the initial models

  • The number of models required

Cheater detection (Cosmides (1989))

• We have evolved an evolutionary mechanism that sensitizes people to look for cheaters

  • General rule: if one takes a benefit then one must pay the cost

  • Cosmides argues that this evolutionary mechanism explains the facilitation observed on deontic selection tasks

What is counterfactual thinking (CFT)

• The imagination of alternatives to reality

• Simulate a sequence of causal events

• Expressed in the form of a conditional

  • Ex. "if the west hadnt supported ukraine, it couldn’t have fought russia"

    • "what if" "if only"

Individual differences in CFT generation Bacon walsh and martin (2013)

• people spontaneously generate more counterfactual when their mood is low

  • Cause or effect?

CFT and blame

• People attribute more blame when preceding actions are unusual and hence it is easy to imagine and alternative outcome

luck

 the chance happening of fortunate or adverse events

• People use luck (good/bad) to mean narrowly

• Avoiding a (neg/pos) outcome

CFT and the future

• Prep for a task induces upward counterfactual thinking

  • After performing a task, inds who believe they will have to repeat it generate more upward counterfactuals

• CFT thinking can influence future behavior

  • Generating upward CFT between two anagram tasks leads to improved performance in the second set

• They help people learn from mistakes, formulate effective plans, and mitigate the certainty of hindsight bias.

Deficits in CFT Knight and Grabowecky (1995)

• Difficulty imagining counterfactual alternatives to reality may underlie many of the problems experienced by individuals with brain lesions in the dorsolateral PFC

  • Inflexible behavior

  • Suppressed emotions

  • Excessively bound by environmental cues

  • Uncreative

  • Have difficulty making plans

problem

• Start state - current situation

• Goal state - desired situation

• Problem - not clear how to get from start to goal state (subjective)

• Well-defined problem

  • Initial state, goal state, and possible moves are well-defined (ex. Chess)

• Ill-defined problem

  • Start state, end state, and/or possible strategies may be unknown (most everyday problems ex. Exams)

theories of problem solving

• Behaviorist approach

  • Trial and error learning

    • Unsystematic behavior

    • Requires no knowledge

    • Slow

    • Doesn’t work for all problems

    • Risky

  • Thorndike's (1898) cat experiment

• Gestalt approach

  • Problem solving requires insight

    • "aha" moment

  • Gestalt approach evaluation:

    • Recognizes the role of insight

    • Mechanisms underlying insight are not specified

insight and prob solving

• Neurosci and insight

  • Activation in right anterior superior temporal gyrus

    • FMRI: insight relative to non-insight solutions

    • EEG: prior to insight

representational change theory

• Representational change theory (Ohlsson 1992)

  • Aims to explain the processes underlying insight

    • Construct a problem representation

    • Retrieve operators (moves/actions) from memory by spreading activation from the problem representation

    • Impasses occur when the problem representation does not cue the right operators

    • Impasses are broken by restructuring the problem representation

    • Once an impasse is broken a full or partial insight may occur

Allen Newell and Herb Simon (1972) info processing

• Computational modeling approach - general problem solver

  • Most problems don't require insight, focus instead on well defined knowledge lean problems

• General problem solver

  • Problem space: all possible states of a problem (all chess positons)

    • Initial state: starting position

    • Goal state: final position (ex. Checkmate)

    • Operators: allows moves or actions (ex. Chess moves)

    • Problem solving is a search through the problem space

  • We don’t have the WM capacity to think of all possible moves

    • Use general purpose heuristics instead

    • Objective measure of optimal performance and can test whether people make moves that are consistent with the heuristics

Analogical problem solving

•  how do we learn from past problems

  • Negative transfer

    • functional fixedness the perceived inability of someone to use an object for something other than its original intended purpose

  • Positive transfer

    • Near transfer to a similar context

    • Far transfer to a different context

• Retrieving analogies is hard unless the problems share similar surface features

• In real life, this may be even more difficult because the time and context might be more distant than in lab studies

• Individual differences are not well understood

Judgment vs. Decision-Making:

• Judgment vs. Decision-Making:

  • Decision-making involves choosing a course of action, while judgment focuses on estimating the probability of events.

  • Judgments are evaluated based on accuracy, while decisions are assessed based on consequences.

Support theory

proposes that subjective probability increases with more explicit and detailed event descriptions.

Tetlock's social functionalist approach

suggests biases in decision-making stem from the need to justify decisions to others.

Dijksterhuis's unconscious thought theory

proposes the superiority of unconscious over conscious thinking, but this is debated in research.

Incubation

aids problem solving by allowing irrelevant information or ineffective strategies to fade.

Individuals with high fluid intelligence

• excel in analogical reasoning, with the left rostrolateral prefrontal cortex playing a crucial role.

Chess-Playing Expertise:

• Expert chess players possess more cognitive ability and extensive knowledge, allowing rapid identification of good moves.

• Chess expertise relies on template-based knowledge, emphasizing fast search processes over complex strategies.

Medical Expertise:

• Medical experts rely on fast, automatic processes for diagnosis, with analytic processes enhancing performance, especially for experts.

• Diagnostic errors in medical expertise can stem from failures in detection, recognition, or judgment.

Functional Theory of Counterfactual Thinking

• Counterfactual thoughts are reflections on alternative outcomes of past events.

• The functional theory suggests that counterfactual thinking serves the purpose of behavior regulation and performance improvement.

Counterfactual thoughts can influence behavior through two pathways (Byrne 2002)

• Content-specific pathway: Directly affects behavioral intentions based on specific informational effects.

• Content-neutral pathway: Indirectly influences behavior through affect, mind-sets, or motivation.

emotions and CFT (Byrne 2002)

• Counterfactual thinking amplifies emotions such as guilt, regret, and relief by comparing actual outcomes with alternative possibilities.

• It influences social judgments regarding accountability, responsibility, and blame.

Evans' hypothetical thinking theory

based on dual processes, proposes that people typically consider only one hypothesis at a time, prioritize the most relevant hypothesis, and accept candidate explanations as long as they are satisfactory. This suggests a nuanced approach to how people engage in hypothesis testing.

confirmation bias,

where people tend to seek evidence that confirms their hypotheses rather than testing them rigorously. Later, this idea was modified to suggest a bias towards positive test strategies.