Decision Making Under Uncertainty

Flashcards covering types of uncertainty, sensitivity analysis methods, structured challenge techniques, and individual/group cognitive biases in decision making.

Aleatory uncertainty

Inherent and unavoidable variation such as biological noise or random equipment failure that cannot be reduced.

Epistemic uncertainty

Uncertainty due to a lack of knowledge, such as limited sample sizes or early-stage data, which can be reduced with more information.

Sensitivity Analysis

A mapping from inputs to outputs used to determine how changes in parameters affect conclusions and identify which uncertainties constitute actual threats.

One-way (univariate) sensitivity analysis

An analysis where one parameter is varied at a time while holding others constant to identify which single parameters have the greatest influence.

Multi-way sensitivity analysis

An analysis where two or more parameters are varied simultaneously to capture interactions between them.

Scenario-based sensitivity analysis

An assessment that compares outcomes across different "worlds" by defines a small number of coherent bundles of parameter values, such as "optimistic" or "challenging."

Global sensitivity analysis

A systematic exploration of a wide range of parameter combinations, often utilizing sampling methods in complex simulation models.

Devil’s advocate

A thinking technique where an individual deliberately takes an opposing or critical position to test the strength of an idea and highlight edge cases.

Pre-mortem analysis

A technique where a team imagines a project has failed disastrously before it starts and works backwards to identify all plausible reasons for that hypothetical failure.

Post-mortem analysis

An analysis that occurs after a project has failed to diagnose what actually went wrong, differing from a pre-mortem which is used for prevention.

Red teaming

The deliberate use of an independent group to challenge a main team’s (the "blue team") plans, assumptions, and analysis using an adversarial stance with constructive intent.

Delphi Technique

A group estimation method where experts provide individual forecasts, share reasoning for outer quartile estimates, and repeat the process until a uniform estimate is reached.

PERT

Program Evaluation and Review Technique; a method that uses probabilistic estimates, originally developed for calculating task timelines.

Confirmation bias

The tendency to seek out and overweight information that confirms pre-existing beliefs while underweighting negative or conflicting data.

Overconfidence

The tendency to overestimate the accuracy of one's judgements and provide narrow performance ranges, such as predicting a sensitivity of $95-97\%$ when the evidence only supports $85-97\%.$

Anchoring

Relying too heavily on an initial value or first estimate, such as a first prototype's performance, when making later judgements.

Sampling bias

Drawing strong conclusions from unrepresentative or convenient samples and assuming their results are generalisable to other populations.

Availability bias

Judging the likelihood of an event based on how easily examples come to mind, often overestimating rare but memorable incidents.

Sunk cost fallacy

Continuing a course of action based on the amount of past investment rather than on the current evidence or likelihood of future success.

Groupthink

A decision trap where the desire for harmony leads to the suppression of dissent, rapid consensus with little debate, and under-exploration of failure modes.

Polarisation

A group-level bias where discussion pushes the team's position to become more extreme, leading to the minimization of risks and overly optimistic assumptions.

Social loafing

A phenomenon where individuals feel less personal responsibility for the quality of a decision when working in a group, leading to less effort in checking assumptions.

HiPPO effect

The Highest Paid Person’s Opinion; a situation where the opinion of the most senior person dominates the decision regardless of the evidence.

Psychological safety

The shared belief within a team that it is safe to ask questions, admit uncertainty, and offer dissenting views without being seen as negative.

Shadow Boards

Parallel decision forums mirroring a main decision-making body but composed of younger or more junior staff to provide alternative perspectives.