Unit 4 - Reasoning and Problem Solving

Judgement

Draw conclusions from evidence often from life experience

Tradition Belief

Emotions skew logic and everything is based on emotions

Idealized Belief

Decision uses formal logic

Inductive Reasoning

Bottom up reasoning using observation to make a theory

Deductive Reasoning

Top down reasoning using the general idea and observations to confirm

Belief Perseverance

Even with disconfirming evidence, people will still cling to belief and seek to confirm it

Even when aware of bogus praise,

People will still think they’re great because when given information, you search for evidence to confirm in memory

Balance

Aims to give attention to both sides but ends up making people think experts are stupid

Frequency Estimate

Assessment of how often various events have occured in the past

Attribute Substitution

Strategy relying on easier accessed information as a measure for information you need

Availability Heuristic

Using memory of frequency of event to judge how likely something is to occur

The way memory is organized

Changed judgment

Conjunction Fallacy

Assuming more specific scenario is more probable than a more statistical general one

What’s feared is

What’s perceived as uncontrollable

Imagery increases

The chance you think it happened

Kahneman

People guided by emotions rather than reson, easily swayed by detail

Representative Heuristics

Assumes homogeneity and each member in that category is representative of the category (usually when you lack experience)

People use similarity to

Prototypical members rather than probability for judgment

Types of Attribute Substitution

• Availability

• Representative

• Affect

• Effort

Availability Attribute Substitution helps with

Frequency but many factors other than frequency can influence from memory

Representative Attribute Substitution helps with

Probability of things have certain features

Affect Attribute Substitution helps with

Risks and benefits with an outcome, but often emotionally and other factors influenced with no connection

Effort Attribute Substitution helps with

Estimate value but value is independent of effort

Heuristics

Efficient strategy that usually leads to the answer

Von Restoff Effect

Memory is better for things that stand out, is dated, humorous, bizarre

When looking for evidence

Use event frequency influenced by memory

More likely to overestimate frequency of

Rare events, especially emotional one because well recorded in memory

Anchoring

Cognitive bias people rely on the 1st piece of information received when making decisions/judgment

Juries like to fall for

Conjunction fallacies because more specific

Sunk Cost Fallacy

To keep doing something because of previously invested resources have been spent even when it’s more rational to stop

Gambler’s Fallacy

Prior outcomes can influence the outcome of probabilistic events by assuming homogeneity for small data pools

Law of Large Numbers

Things tend to even out in large samples but not smallM

Man Who Argument

A view of something used to persuade someone as an assumption of homogeneity

Hindsight Bias

People thinking after that they would’ve known something before the fact

Illusory Superiority Effect

On average, people think they’re better than average

Most accurate calibrated people are

Clinically depressed

Covariation

Relation between two variable where presence of a variable can be predicted from the presence of the other

Illusions of Covariation

Prone to detect patterns that aren’t there

People are really good at learning

Statistics in real world with no bias and are systematically conservative

When people have pre-existing theories, they get

Inaccurate, variable overestimates

Confirmation Bias

Tendency to be more alert to evidence that confirms beliefs rather than challenging it

Memory Scheta

Evidence that fits in schemata stays and counter evidence doesn’t have space in schemata

Paranormist and Conspirators tends to

See more illusory patterns and more likely to narcissistic

Base Rate Information

Information about how frequently something occurs in general

Diagnostic Information

Information about a particular case

When given base rates and descriptive information,

Only use descriptive information

Dual Process Model

Any model of thinking that claims people have a distinct judgment process type 1 or 2

Type 1

Fast, effortless, prone to error reasoning, heuristicsT

Type 2

Slow, effortful reasoning

Type 2 triggered by

Certain cues, if the circumstances are right to focus attention on judgement

Type 1 more likely when

Under pressure but also considers accidents or flukes if chance is involved

Base Rate Neglect happens for

Probabilties or proportion, but not frequency

Overconfidence Effect

More confidence in judgments than justifiable in both amateur and expert, even though the expert you are, the less confident

Categorical Syllogisms

Form of deductive reasoning with a major and minor premise

Valid Syllogism

Conclusion follows premises

Invalid Syllogism

Conclusion is not logically following premises

People are bad at

Reasoning syllogism and often approach with a consistently bad, illogical strategy

Conditional Syllogism

Uses if, then formate with the 1st part giving a condition under which the 2nd part has to be true, but not reversed

Modus Ponens

Affirms antecedent

Modus Tollens

Denying consequent (If not p, then not q)

Inverse Fallacy

Assumes negation of antecedent implies negation of the 2nd part (if not p, then not q)

Converse Error

Incorrectly concluding antecedent of conditional statement must be true because consequent is true

Bidirectional Premise

Uses if and only if allowing to use a conclusion about both premises because there’s no other way

Belief Bias

Tendency to endorse a conclusion if they believe to be true, and ignore premises of logic and argument to rely on broader pattern of belief of what’s true and what isn’t

Atmosphere Errors

Some conclusions seem more appropriate given context

Conversion Errors

Incorrectly assuming promise that’s illogically valid

Pragmatic Reasoning Schemata

Generalized, context-sensitive set of rules like syllogism to reason situations related to permission/obligation that people are better at

Formal Logic uses

Sufficiency, but languages allows people to reason necessity or sufficiency

Reasoning Schemata can be improved by

Examples that trigger pragmatic reasoning and guiding interpretation to the right frame

Selection Task

A person is shown 4 cards with certain information on each card, and given a rule that may describe the cards, people decide which cards must be turned to find out if the rule is right or not

Utility Theory

People should make decisions on what you value and want based on cost and benefit of choice (subjective)

Expected Utility

Value determined on probabilites

Certainty Effect

People prefer sure gain

Risk Seeking

Dealing with possible losses

Risk Aversion

Dealing with gains made lost

Reasoning is based on

Percentages than numbers

Prospect Theory

Value potential losses more heavily than equivalent gains, risk-averse for gains but risk-seeking for losses

Loss Aversion

Given a choice between avoiding loss and gains, strong preference to keep gains

Utility Theory says

Sunk cost is stupid

Reasoning Based Choice

People guided by decision that they feel good about and thought to be reasonably justified

People rely on

Justification and shifting frame causes a change in factor relevant to justification

Regret Theory

Avoiding regret is a big motivation

Somatic Markers

A way of evaluating options using “gut feeling”

Affective Forecasting

Predicting future emotions, often inaccurate

People overestimate how long

Feelings last and underestimate ability to adjust to changes in the future

People prefer to endure

More pain and it to gradually get better because it’s remembered better

Game Theory

Defeating is dominant strategy because it has the highest payoff but best outcome is mutual cooperation

Pretrail Publicity is rarely neutral

Emphasizes police and prosecution POV, causing confirmation bias

Jury decisions are based on

Evaluation of evidence which is filtered with confirmation bias of information previously seen

Problem-Solving

Needing to figure a path to a goal, either a process of search or visually

Operations

Relevant things you can do

Problem Space

Set of all relevant things you can do

Intial State

Givens in a problem

Operators

Tools and actions that can change current state

Path Constraints

Limitations on what you can do

Brute Force Solution

Going through every single solution

Hill Climbing

Always have to move uphill but sometimes solving it requires you to step away

Means End Analysis

Comparing difference between current state with goal and seeing which means can be used to make them more alike

Working Backwards is helpful if

Number of ways backwards is less than choices forward

Analogies are useful but it’s

Hard to come up with them to solve your own problems, especially since they aren’t obvious