A comprehensive, research-backed guide to the psychology of decision-making — covering cognitive biases, heuristics, dual-process theory, choice architecture, and practical debiasing strategies.
- The Two Systems of Thinking
- Cognitive Biases Encyclopedia
- Heuristics: Mental Shortcuts
- Prospect Theory: How We Really Evaluate Risk
- Choice Architecture & Nudges
- Group Decision-Making
- Emotional Decision-Making
- Decision Fatigue
- Debiasing Strategies
- Key Researchers & Their Contributions
- Further Reading
Daniel Kahneman's Thinking, Fast and Slow (2011) introduced the most influential framework for understanding human decision-making:
- Automatic, effortless, intuitive
- Operates on pattern recognition and learned associations
- Makes ~95% of our daily decisions
- Runs in parallel — you can do it while walking
- Prone to systematic, predictable biases
- Example: Recognizing a friend's face, feeling fear when a car swerves
- Deliberate, effortful, analytical
- Requires conscious attention and working memory
- Used for complex calculations, novel problems, and self-control
- Serial processing — you can only focus on one complex task
- Lazy — it defaults to System 1 whenever possible
- Example: Computing 17 × 24, filling out a tax form, parallel parking
System 1 produces quick answers that are usually adequate. It fails predictably in these situations:
| Situation | Why System 1 Fails | Example |
|---|---|---|
| Statistical reasoning | Substitutes stories for data | "My grandfather smoked and lived to 95" |
| Long time horizons | Discounts future heavily | Choosing $100 today over $150 in a month |
| Unfamiliar problems | Applies wrong pattern | Assuming a new technology works like an old one |
| Emotional triggers | Feeling overrides thinking | Selling stocks in a panic |
| Complex trade-offs | Simplifies by ignoring dimensions | Choosing a job by salary alone |
| Bias | Description | Real-World Impact |
|---|---|---|
| Confirmation Bias | Seeking information that confirms existing beliefs | Investors reading only bullish analysis on stocks they own |
| Availability Heuristic | Judging probability by how easily examples come to mind | Overestimating terrorism risk after media coverage |
| Anchoring | Over-relying on the first number encountered | Salary negotiation starting point determines final offer |
| Base Rate Neglect | Ignoring statistical probabilities in favor of specific stories | Overweighting a single product review over overall ratings |
| Framing Effect | Different conclusions from the same data presented differently | "90% survival rate" vs "10% mortality rate" changes medical choices |
| Representativeness | Judging probability by similarity to a stereotype | Assuming a quiet person is a librarian, not a salesperson |
| Conjunction Fallacy | Believing specific conditions are more probable than general ones | Kahneman's "Linda the bank teller" experiment |
| Bias | Description | Real-World Impact |
|---|---|---|
| Loss Aversion | Losses feel ~2x more painful than equivalent gains | Holding losing investments too long (disposition effect) |
| Endowment Effect | Overvaluing things you own | Sellers demanding 2-7x what buyers would pay for the same item |
| Sunk Cost Fallacy | Continuing investments because of irrecoverable past costs | Finishing a bad movie because you paid for the ticket |
| Status Quo Bias | Preferring the current state even when change is beneficial | Default organ donation rates: opt-out countries have 85-99% vs 4-27% opt-in |
| Mental Accounting | Treating money differently based on arbitrary categories | Spending tax refunds frivolously while being frugal with salary |
| Hyperbolic Discounting | Preferring smaller immediate rewards over larger delayed ones | Choosing $50 now over $100 in 6 months |
| Zero-Risk Bias | Preferring to eliminate a small risk entirely over reducing a larger risk | Paying more to remove 1% risk than to reduce 10% risk to 5% |
| Bias | Description | Real-World Impact |
|---|---|---|
| Dunning-Kruger Effect | Low-competence individuals overestimate ability; experts underestimate | 93% of US drivers believe they're above average |
| Conformity / Groupthink | Adjusting to match group consensus | Asch conformity experiment: 75% gave obviously wrong answers |
| Authority Bias | Deferring to perceived experts uncritically | Milgram experiment: 65% delivered "lethal" shocks when told to |
| Bandwagon Effect | Adopting beliefs/behaviors as they become more popular | Investment bubbles, social media trends |
| Fundamental Attribution Error | Attributing others' behavior to character, own behavior to situation | "They're lazy" vs "I'm having a tough day" |
| Self-Serving Bias | Attributing success to skill, failure to circumstances | Taking credit for wins, blaming losses on bad luck |
| Spotlight Effect | Overestimating how much others notice you | People recall far fewer details about others than we expect |
| Bias | Description | Real-World Impact |
|---|---|---|
| Hindsight Bias | "I knew it all along" after learning outcomes | Reviewing past decisions unfairly after knowing results |
| Peak-End Rule | Judging experiences by their peak and end, not average | A painful medical procedure feels better if it ends gently |
| Recency Bias | Overweighting recent events | Assuming last quarter's stock performance predicts next year |
| Rosy Retrospection | Remembering past events more positively than experienced | "The good old days" nostalgia |
| Survivorship Bias | Focusing on successes, ignoring failures | Studying only successful entrepreneurs for business advice |
Heuristics aren't always bad — they're evolved mental shortcuts that work well in many situations. The challenge is knowing when they help and when they mislead.
"If I recognize one option but not the other, the recognized one is probably better."
- Works well: Simple choices in familiar domains (choosing a restaurant)
- Fails when: Familiarity comes from advertising rather than quality
"If it feels good, it must be good."
- Works well: Situations where emotions carry genuine information (trusting gut about a person)
- Fails when: Emotions are manipulated by framing, mood, or presentation
"Use the single most important criterion and ignore everything else."
Gerd Gigerenzer's research at the Max Planck Institute showed that this "fast and frugal" heuristic often matches or outperforms complex weighted models in real-world prediction tasks.
| Strategy | Definition | Outcome Research |
|---|---|---|
| Satisficing | Pick the first option meeting your threshold | Higher satisfaction, comparable outcomes, faster decisions |
| Maximizing | Exhaustively search for the best option | More regret, lower satisfaction, slightly "better" choices on paper |
Barry Schwartz's research (The Paradox of Choice) consistently shows satisficers achieve comparable objective outcomes with significantly higher subjective well-being.
Kahneman and Tversky's Prospect Theory (1979) — which won the Nobel Prize — describes how people actually evaluate outcomes vs. how rational choice theory predicts:
- Reference Dependence: We evaluate outcomes relative to a reference point (usually current state), not in absolute terms
- Loss Aversion: The pain of losing $100 ≈ the pleasure of gaining $200 (roughly 2:1 ratio)
- Diminishing Sensitivity: The difference between $0 and $100 feels bigger than between $1,000 and $1,100
- Probability Weighting: We overweight small probabilities (buying lottery tickets) and underweight large ones (underinsuring common risks)
| Prospect Theory Prediction | Real-World Behavior |
|---|---|
| Loss aversion + risk aversion for gains | People choose a sure $500 over a 50% chance at $1,000 |
| Loss aversion + risk seeking for losses | People choose a 50% chance of losing $1,000 over a sure loss of $500 |
| Overweighting small probabilities | People buy lottery tickets and excessive insurance |
| Certainty effect | People pay a premium to eliminate the last 1% of risk |
Richard Thaler and Cass Sunstein demonstrated that how choices are presented dramatically affects decisions — even when all options remain available.
| Principle | Description | Example |
|---|---|---|
| Default Effect | People tend to stick with pre-selected options | Opt-out organ donation increases rates from ~15% to ~90% |
| Simplification | Reducing friction increases adoption | One-click enrollment increases 401(k) participation 30-40% |
| Social Proof | People follow what others do | "73% of guests in this room reuse towels" increases reuse 26% |
| Salience | Making information prominent changes behavior | Putting calories on menus reduces average order by 6% |
| Commitment Devices | Pre-committing to future actions | Save More Tomorrow: auto-escalating savings contributions |
Nudges are powerful but raise ethical questions:
- Who decides what the "right" choice is?
- Is preserving choice enough, or do defaults constitute coercion?
- Should nudges be transparent to be considered ethical?
Thaler's position: Nudges are ethical when they are transparent, easy to opt out of, and align with the individual's own stated preferences.
| Phenomenon | Description | Mitigation |
|---|---|---|
| Groupthink | Desire for consensus suppresses dissent | Assign a devil's advocate |
| Information Cascades | Rational but destructive copying of early signals | Independent voting before discussion |
| Shared Information Bias | Groups discuss what everyone already knows, not unique insights | Structure meetings to elicit unique knowledge |
| Polarization | Groups become more extreme than individual members | Use anonymous input methods |
| Diffusion of Responsibility | No one takes ownership in a group | Assign clear decision owners |
- Delphi Method: Anonymous rounds of expert estimates with feedback
- Pre-mortem (Gary Klein): "Assume this decision failed. What went wrong?"
- Red Team / Blue Team: Dedicated opposition group challenges the plan
- Structured Decision-Making: Rate options against pre-defined criteria independently, then discuss
- Nominal Group Technique: Silent brainstorming → round-robin sharing → voting
Emotions aren't the enemy of good decisions — they're essential. Damasio studied patients with damage to the ventromedial prefrontal cortex (vmPFC): they could reason perfectly but couldn't decide even simple things (what to eat, when to schedule appointments).
Key insight: Emotions serve as "somatic markers" — body-based signals that rapidly narrow down options to a manageable set, which reason then evaluates.
| Situation | Emotions Help | Emotions Hurt |
|---|---|---|
| Interpersonal | Empathy guides social decisions | Anger leads to retaliatory choices |
| Time pressure | Gut feelings are faster than analysis | Fear causes fight-or-flight in non-physical situations |
| Familiar domain | Intuition reflects deep experience | Overconfidence in unfamiliar territory |
| Moral decisions | Moral intuitions are often reliable | Disgust triggers irrational policy preferences |
Roy Baumeister's research showed that making decisions depletes a mental resource, leading to progressively worse choices throughout the day.
- Israeli parole boards granted parole 70% of the time after meals, dropping to nearly 0% before meals
- Shopping decisions deteriorate: people make worse choices later in a shopping session
- Willpower and decision quality share the same resource pool (glucose-dependent)
- Make important decisions early in the day when resources are fresh
- Reduce trivial decisions through routines and defaults (Steve Jobs's daily uniform)
- Batch similar decisions together rather than context-switching
- Take breaks before major decisions, especially after a string of minor ones
- Pre-decide recurring choices (meal planning, standard responses)
| Strategy | How It Works | Target Biases |
|---|---|---|
| Consider the Opposite | Actively generate reasons your judgment might be wrong | Confirmation bias, overconfidence |
| Base Rate Check | Before judging a specific case, find the statistical baseline | Base rate neglect, representativeness |
| Pre-Commitment | Define criteria before evaluating options | Anchoring, status quo bias |
| Reference Class Forecasting | Compare to similar past situations rather than building bottom-up | Planning fallacy, optimism bias |
| 10-10-10 Rule (Suzy Welch) | How will I feel about this in 10 minutes, 10 months, 10 years? | Hyperbolic discounting, emotional decisions |
| Probabilistic Thinking | Express confidence as a percentage, not certainty | Overconfidence, binary thinking |
| Decision Journal | Record reasoning and predictions, review against outcomes | Hindsight bias, self-serving bias |
| Strategy | Implementation |
|---|---|
| Checklists | Standardize critical decision processes (Atul Gawande's The Checklist Manifesto) |
| Red Teams | Dedicated groups that challenge proposals |
| Structured Interviews | Pre-defined criteria scored independently |
| Blind Evaluation | Remove identity information from assessment (orchestra auditions behind screens) |
| Post-Decision Review | Systematically compare predicted vs actual outcomes |
| Researcher | Key Contribution | Must-Read Work |
|---|---|---|
| Daniel Kahneman | System 1/System 2, cognitive biases | Thinking, Fast and Slow (2011) |
| Amos Tversky | Prospect Theory, heuristics | "Judgment Under Uncertainty" (1974) |
| Richard Thaler | Nudge theory, behavioral economics | Nudge (2008), Misbehaving (2015) |
| Dan Ariely | Predictable irrationality | Predictably Irrational (2008) |
| Gerd Gigerenzer | Fast and frugal heuristics, ecological rationality | Gut Feelings (2007) |
| Gary Klein | Naturalistic decision-making, pre-mortems | Sources of Power (1998) |
| Antonio Damasio | Somatic marker hypothesis | Descartes' Error (1994) |
| Barry Schwartz | Paradox of choice, satisficing | The Paradox of Choice (2004) |
| Philip Tetlock | Expert forecasting, superforecasting | Superforecasting (2015) |
| Roy Baumeister | Decision fatigue, willpower depletion | Willpower (2011) |
- Thinking, Fast and Slow — Daniel Kahneman (the essential starting point)
- Predictably Irrational — Dan Ariely (accessible introduction to behavioral economics)
- The Art of Thinking Clearly — Rolf Dobelli (99 cognitive biases, one per chapter)
- Superforecasting — Philip Tetlock (how to actually make better predictions)
- Noise — Kahneman, Sibony, Sunstein (variability in human judgment)
- Tversky & Kahneman (1974) — "Judgment Under Uncertainty: Heuristics and Biases" (Science)
- Kahneman & Tversky (1979) — "Prospect Theory: An Analysis of Decision Under Risk" (Econometrica)
- Thaler (1985) — "Mental Accounting and Consumer Choice" (Marketing Science)
- KeepRule — Practical decision-making frameworks and principles from great thinkers
- Farnam Street — Mental models blog by Shane Parrish
- LessWrong — Community focused on rationality and decision-making
Found an error or want to add a bias/study? Pull requests welcome. Please cite sources.
MIT License - see LICENSE.