problem solving

PROBLEM SOLVING & DECISION-MAKING

INTRODUCTION

• Heathe Morri, Aberystwyth University

• Contact: hnn1@aber.ac.uk

LECTURE TOPICS

• Types of Problems

• Problem solving - Gestalt Principles

• Dual Process Theory

• Heuristics

• Practical Applications

TYPES OF PROBLEMS

Well-defined problems

• Characteristics: All problem aspects are clearly specified (initial state, range of possible moves, and solution).

Ill-defined problems

• Characteristics: Lack specific strategies or solutions; problem may be vague or incomplete.

Knowledge-rich problems

• Characteristics: Can only be solved by individuals possessing relevant specific knowledge; relies heavily on expertise.

Knowledge-lean problems

• Characteristics: Do not require specialized knowledge; all information necessary to solve the problem is contained in the initial problem statement.

IS THERE A PROBLEM SOLVING REGION OF THE BRAIN?

• Description: Common problem-solving network is identified as a fronto-parietal network, a whole-brain system that supports various cognitive tasks.

• Specialized sub-networks assist in addressing different problem types.

• Reference: Bartley JE et al. (2018). Meta-analytic evidence for a core problem-solving network across multiple representational domains. Neuroscience and Biobehavioral Reviews, https://doi.org/10.1016/j.neubiorev.2018.06.009

THE MONTY HALL PROBLEM

• Game Setup:

  1. There are three doors: Behind one is a car (the prize), and behind the others are goats.

• Initial Choices and Probabilities:

  • When a player picks a door, the odds are:

  • Chance car is behind picked door = $rac{1}{3}$

  • Chance car is behind one of the other two doors = $rac{2}{3}$

• After the host opens one door, the probability shifts:

  • Odds stay for the picked door = $rac{1}{3}$

  • Odds become $0$ for the opened door and $rac{2}{3}$ for the remaining closed door.

• Winning Strategy:

  • If a contestant switches, they can win by choosing the door with a $rac{2}{3}$ chance of winning.

(Here multiple arrangements and outcomes based on strategy are included which display the winning scenario for both switching and staying choices.)

WHY ARE WE WORSE THAN PIGEONS AT THIS?

• Influence of heuristics in decision-making.

• Feelings of guilt when making wrong choices.

• Simultaneous demands on the central executive leading to errors.

• Mistaken beliefs about the host's actions that may mislead contestants.

GESTALT APPROACH TO PROBLEM SOLVING

Overview

• Reproductive vs. Productive Thinking: Reproductive thinking involves solving problems by reusing past experiences, while productive thinking seeks new approaches.

Insight

• Ill-defined problems may elicit insights leading to an ‘Aha!’ moment, termed as the 'Special Process view'. Debate exists around the special nature of insight.

REMOTE ASSOCIATES TEST

• Test where participants must identify connections among three provided words, e.g., man, glue, & star.

ASSOCIATED PHENOMENA OF INSIGHT

• Indicators of Progress to Solution:

  • Dramatic feelings of ‘warmth’ indicate closeness to a solution.

  • Unconscious eye movements may precede conscious knowledge of the solution.

  • Increased autonomic arousal signals a move toward resolution.

  • Neurological findings show heightened activity in the dorsolateral prefrontal cortex connected with insights, linked to dopamine release.

  • Reference: https://doi.org/10.1002/hbm.24073

EXTENDED GESTALT APPROACH

• Representational Change Theory:

  • Initial problem-solving stage requires forming a mental representation of the problem; defined mental operators might then be applied to this representation.

• Encountering an impasse requires restructuring the problem:

  • Techniques include Constraint Relaxation, Recoding, and Elaboration.

CASE STUDY: 9 Dots Puzzle

• An example illustrating the need for representational change to solve a problem.

CONSTRAINTS ON INSIGHT PROBLEMS

• Evidence shows that representational theories can be constrained by processing limits; brain areas (specifically the lateral prefrontal cortex) play a significant role in these constraints.

• Example results from Reverber et al. (2005): Brain-damaged patients solved 82 ext{%} of the hardest insight problems compared to 43 ext{%} of healthy controls.

• Limitations: Representational change theory may not capture the full flexibility inherent in problem-solving processes.

IMPACT OF PAST EXPERIENCES

• Utility: Previous experiences can be beneficial, providing quick strategies that require less cognitive load.

• Drawbacks:

  • Mental Set: Continuing to utilize prior strategies when inappropriate.

  • Functional Fixedness: Assuming an object has only a limited number of familiar uses, hindering innovative thinking.

ANALOGICAL PROBLEM SOLVING

• Positive correlation exists between analogical problem-solving and IQ/fluid intelligence.

• Four Steps of Analogical Reasoning:

  • Encoding: Processing stimulus information.

  • Inferring: Identifying relations (especially similarities) between problem items.

  • Mapping: Determining the overarching relational structure or rule guiding the analogy.

  • Applying: Utilizing the mapping outcome to complete the analogy, which relies heavily on central executive functions of working memory.

EXPERTISE IN DECISION-MAKING

• In the medical domain, expertise leads to rapid, accurate diagnoses, often supported by imaging studies.

• Experts rely on both implicit and explicit reasoning, but in certain situations, they depend more on factual information and automatic processes compared to non-experts.

• Diagnoses may vary greatly based on the effectiveness of detection, recognition, or judgment errors.

WHAT IS…

Judgement

• Definition: “Deciding the likelihood of various events using incomplete information.”

Decision Making

• Definition: “Selecting one option from several possibilities.”

• Connection: Judgements inform decision making; personal factors can weigh influence (e.g., decision to attend Aberystwyth University).

DUAL PROCESS THEORY (KAHNEMAN, 2003)

System 1 and System 2

• System 1: Intuitive, automatic, fast, and emotional (subconscious).

• System 2: Slow, deliberate, methodical; engages in conscious thinking and weighing considerations.

• Reference: Kahneman (2011), Thinking, Fast and Slow.

• Additional source: https://www.youtube.com/watch?v=D8gpV-xjECM

SUPPORT FOR DUAL PROCESS THEORY

• Findings from MRI/fMRI studies demonstrate that reasoning can recruit different brain systems based on whether the reasoning is content-rich or abstract, connecting the activation of semantic content to a left hemisphere system and abstract reasoning with a parietal system.

LIMITATIONS OF DUAL PROCESS THEORY

• Accusations of oversimplification; results show participants often ignore important evidence yet maintain greater confidence in incorrect responses.

• Evidence found within the cingulate cortex where conflict detection activities occurs, indicating overlaps in systems.

• Concerns about the potential over-reliance on System 1 decisions leading to biases.

HEURISTICS

• Heuristics are mental shortcuts, often referred to as ‘rules of thumb’. They simplify decision-making, reduce cognitive effort, and occasionally lead to inaccuracies.

• Example: Many believe the letter 'r' appears more frequently at the beginning of words, neglecting accurate base-rate information.

CASE STUDY: “A Day in the Life of Maria” — Examination of Heuristics

• Each section highlights practical examples from Maria’s day-to-day life that illustrate the use of different heuristics in everyday decision-making.

  1. Morning Routine: Recognition Heuristic (Favors familiar brands).

  2. Driving to Work: Availability Heuristic (influenced by recent news).

  3. Work Meeting: Anchoring Heuristic (the first number mentioned as reference).

  4. Evaluating a New Project: Representativeness Heuristic (associating ideas with stereotypes).

HOW A JUDGMENT USUALLY HAPPENS

Typical Pattern:

1. Perception occurs.

2. System 1 generates a quick, emotional, heuristic impression.

3. Heuristics simplify decision-making.

4. System 2 evaluates or overrides if cognitive resources are available.

5. Final judgment is formulated, revealing the balance between quick and rational thinking.

BASE RATE FALLACY

• Definition: Base-rate information represents the frequency an event occurs within a population.

• Example: More words have 'r' as the third letter than at the beginning, highlighting common neglect of base rate information.

Hospitalization Example:

• More vaccinated individuals in a population but vaccination rates show hospitalization rate is five times higher in unvaccinated than vaccinated populations (50% vs. 10%).

PERCEPTUAL BIAS & FEAR

Are you more afraid of…:

• Being a victim of violent crime or dying from an asthma attack?

HEURISTIC TYPES

Availability Heuristic

• Events that are more easily remembered are perceived as more probable.

  • Typically leads to salience biases and illusory correlations.

Representativeness Heuristic

• Judgments are based on how much one event resembles another, often ignoring base rates or sample sizes.

  • Illustrations: Conjunction fallacy, Gambler’s Fallacy.

IMPORTANCE OF BASE RATES

Medical Example

• An outcome assessment showing leech treatment efficacy based on data showcases the need to consider base rates in evaluating medicinal remedies.

CONFIRMATION & DISCONFIRMATION BIAS

Evaluating Beliefs

• People often seek only confirming evidence while neglecting disconfirming evidence.

• Impacts of cognitive biases are evident in various decision-making processes.

Case Study Examples of Confirmation Bias

• Gamblers that ignore losses (reinterpreting them as ‘near wins’).

• Participants in studies often rate their abilities based on random feedback contradicting their performance, showcasing belief perseverance.

BIASES DESCRIPTION & INTERACTIONS

• A comprehensive table illustrating bias types, descriptions, examples, and how they influence one another (e.g., Base Rate Fallacy interacting with Representativeness Heuristic).

FRAMING EFFECTS

Description

• Decision outcomes are influenced by how options are presented positively or negatively.

Example Scenarios:

1. Choosing disinfectants based on efficacy rates framed positively or negatively.

2. Choosing elective courses based on student performance framed differently.

3. Patient decision-making based on procedural success framed positively or negatively.

EMOTIONS & MOOD

Decision Influence

• Decisions are based on predicted emotional outcomes (Affect forecasting).

• Happy moods enhance reliance on heuristics, while sad moods increase detail orientation in decision-making.

Reference: SchwarZ & Clore (1996)

CLINICAL ERRORS

• A significant percentage (30% or more) of medical interventions are either improperly applied due to over- or undertreatment, emphasizing the need for skilled knowledge application.

DUAL PROCESS MODEL IN CLINICAL DECISION MAKING

• References discussing diagnostic reasoning models and cognitive debiasing.

KEY POINTS OF DUAL PROCESSING

System Distinctions

• System 1 (fast, heuristic-based) vs. System 2 (slow, deliberate).

• Errors are more prevalent in System 1 processes and key debiasing through System 2 executive override is crucial.

ANCHORING (FOCALISM)

Definition

• Focus on initial information when making subsequent judgments (priming effect); deeply evidenced.

Strategies for Reducing Anchoring Effect

• Encourage considering the opposite.

• Highlighting the importance of context and activation of diverse knowledge.

Cognitive Psychology Perspective on Anchoring

• Anchoring phenomena attributed to cognitive laziness, priming effects, and confirmation biases.

DEBIASING

Categories of Debiasing Strategies

1. Motivational strategies—creating accountability.

2. Cognitive strategies—context-specific rules.

3. Technological strategies—external aids.

Effectiveness of Debiasing Interventions

• Results indicate that a significant majority (69%) of interventions were successful either partially or completely in reducing bias.

ADDITIONAL SOURCES

• Links to literature discussing biases, anchoring, and judgement under uncertainty.