Decision Theory: Decision-Making Compass

Decision Theory

Part A - Introduction

• Decision theory focuses on systematic decision-making processes
• The "Decision-Making Compass" outlines the steps in decision-making

Steps in the "Decision-Making Compass"

• Define the decision question
• Determine criteria for decision-making
• List possible options
• Model the effects of each option on the goals
• Create utility functions to measure/quantify the results
• Weigh the goals against each other
• Evaluate and Refelect on decision

Crafting Good Decision Questions

• Reframe the question to avoid simple yes/no answers and promote real alternatives
• Specify timeframes
• Clarify change by setting optimization goals (maximize, minimize, etc.)

Intuitive vs. Analytical Decision-Making

• System 1: Intuitive, fast decisions (gut feeling) for simple choices
• System 2: Analytical, slower decisions for complex problems

System 1 vs. System 2: Key Differences

• System 1: Intuitive, taps into experience through associations; filters distorting factors
• System 2: Analytical, identifies objectives + alternatives, avoids bias, selects best alternatives based on criteria
• Choice of method depends on: personality, experience, problem complexity, time, and available resources

Intuitive Decisions

• Involves retrieving associations, filtering ideas, and simple decision rules (heuristics)

Effects that Impact Decisions

• Need for cognition biases
• Becoming-Famous-Overnight phenomenon is when someone who is unknown suddenly becomes famous or well-known in a very short amount of time
• Priming effects
• Belief-bias effects
• Wason selection task tests ability to plausibly validate statements

Homo Oeconomicus

• A model assuming rational, economically driven decision-making
• The concept of Homo Oeconomicus is limited
• The Homo Oeconomicus oversimplifies human behavior by assuming individuals are entirely rational, self-interested, and able to make optimal decisions

Core Motivations, Values, and Goals

• Norms influence behavior, like social norms for warmth and community
• Market norms exist for financial decisions

Fairness and Social Distance

• Ultimatum game: Person A offers a split, Person B accepts or rejects (both get nothing)
• Dictator game: Person A decides how much to give to Person B
• People want to be perceived as fair
• Social closeness promotes fairness and reciprocity

Reciprocity

• A willingness to be friendly or unfriendly that impacts decision making

Clarity of Goals

• Define goals as relevant aspects with direction, but no specific target
• Strive toward an attainable state
• Value Focused Thinking is characterized by repeated questioning, differentiation of fundamental goals, and the creative creation of new alternatives

Instrumental vs. Fundamental Goals

• Instrumental Goals are a means to an end
• Fundamental Goals hold intrinsic value

Goals in Social Interaction

• Loyalty and altruism influence behavior

Probabilities

• Understanding probabilities is critical in decision-making

Probability Interpretations

• Symmetry based (coin toss)
• Frequency based (observed events)
• Subjective interpretations matter

Terminology of Decision Theory

• Experiment (e.g., die roll)
• Sample space (all outcomes)
• Event (subset of outcomes)
• Elementary event (single outcome)
• Complementary event
• Intersection (A and B)
• Union (A or B)
• Conditional probability

Bayes' Theorem

• It is a method to compute probabilities
• Bayes' Theorem uses likelihoods to derive a-posteriori knowledge of occurrences

Probability Distributions

• Variables can be discrete or continuous

Probability Distribution Functions

• It can be discrete
• It can be continuous

Binomial Distribution

• Focuses on outcomes with two different characteristics
• Involves probabilities that are known
• The events order is unimportant

Normal Distribution

• Characterized by mean (μ) and standard deviation (σ)
• The standard deviation is a measure of the amount of variation or dispersion of a set of values

Exponential Distribution

• Incorporates rate parameter (λ)
• The exponential distribution is memoryless as estimates never change

Weibull Distribution

• Includes shape (β) and scale (α) parameters

Part B - Prescriptive Decision Theory

Ideal Rational Decision Process

• Requires Situation Analysis, developing impact models, evaluating options using utilities/weights

Continuous Modeling

• List influencing factors, probabilities, and target specifications

Monte Carlo Simulation

• Simulate outcome distributions based on uncertain factors

Dominance

• In decision-making, one option is sometimes better than the other
• An option can either be weakly dominating or strongly dominating

Rationality

• Select the alternative maximizing preference model-based value

Decisions Under Uncertainty

• Expected utility calculation is the approach, incorporating risk preferences

Risk Premium

• It is the tradeoff between expectation and safety/assurance
• Influenced by attitudes toward taking a risk

Assessing Utility

• Involves using scales, defining fundamental elements, and resolving conflicts

Methods for Scale Determination

• Halfing
• Fractile
• Using variable risks
• Lottery comparisons

Smooth Utility Functions

• Can be determined through aversion parameters

Recognizing Distortions

• Prescription determines how to think
• Description determines how to act

The influence of cognitive biases

Rules for Compatible Decisions

• Involve establishing certain functions that should meet certain criteria

Targeting Goals

• Involves calculating utilities of each goal to assess an alternative

Trade-off

• It aims at assessing the relative importance of two objectives
• Determining the tradeoff between two goals can involve doing more than one

Bandwidth Impacts Affecting Quality

• Changes in range have effects
• Effects of changing bandwidth can be mitigated during the procedure

Missing Information

• Sensitivity analysis is a method for testing the effects of inputs in unknown outcomes

Sensitivity Analysis

• Considers influence of uncertain variables on targets and critical values

Stochastic Dominance

• Involves comparing risk profiles
• Involves knowing what each grade is

Multi-Stage Problems

• Problem can be represented as decision trees

Making Decisions Requires Planning Ahead

• It is a backwards thinking process of determining how to act
• Involves assigning values and determining strategies based on each outcome

Roll-Back Algorithm

• It goes from the back of the tree to the starting point in a progressive manner
• It looks at expectation values and makes an evaluation that is risk neutral

Part C - Descriptive Decision Theory

Cognitive Information Processing

• It is a model assessing thought related decision making

Simplification and Preferences

• Involves breaking down elements based on intransitive data

Heuristics-Based Processing (3-step)

• Involves hypotheses, processing objects, and determining conclusions

Contrast Effect

• Heightened cognition occurs when differences are easily noted

Memory

• The accessibility effects short-term processing
• Long-term process
• Information that lingers is more important

Determinants of Availability

• Salience
• Vividness
• Focus

Narrow Thinking

• Narrow Thinking and heuristics limits cognition
• People conserve cognitive resources
• The mind oversimplies

Effect of Availability and Thought in Decision-Making

Overreaction

• Involves drawing from information and neglecting related information
• The reaction is a consequence of this information flow

Narrative Bias

• Stories take precedence over facts

Primacy

• Involves information early on that is important

Mental Accounting

• Involves allocating costs rather than addressing them as a unit

Anchoring

• It is anchoring to information to extrapolate

Representative thinking

• Involves a schema where events conform

Overconfidence

• Involves overestimating outcomes and ability compared to others

Irrationality due to

• Success reinforcement
• Poor self-assessment

Motives

• People want to protect rational actions

Attribution

• Involves separating factors

Cognition

• Involves maintaining an information free state
• Rationalization occurs when information violates decisions

Factors that Predispose Commitment

• Freedom
• Responsibility
• Norms

Sunk Costs

• Involves money we cannot recover
• Rational decisions can be righted through awareness

Review of Biases and Fallacies

• Considers the effects of emotional decisions

Relative Rewards

• Explains how choices compare to other choices

Averision for Loss

• Loss aversion factors strongly into mental accounting

Relativity

• This theory influences thought

Uncertainty

• The evaluation criteria is different

Commonality

• It assesses the effects of evaluating something

Discounting Models

• They consider real-world decision analysis

Assessing Value

• A future option is not valued the same at the present
• Hyperbolic discounting affects decisions for the worst as we approach

Part D - Applications of Decision Theory

Application Field

• Improve decisiveness in decisions

Application areas

• Helping consumers/companies by nudging their decisions
• Improving user happiness
• Self Awareness of happiness

Effects

• Bias
• Accounting
• Framing

Uniqueness in Decisions within Firms

• Requires a moderator to guide others fairly

Team Decision Making

• Requires a process
• Requires getting data from stakeholder

Handling Conflicts

• Getting groups to agree

Bias Factors

• They occur in groups
• Sharing of knowledge helps everyone do right