bias-detection

What this skill does

# Bias Detection Reference

**Purpose:** A comprehensive reference for identifying and correcting cognitive biases that undermine decision quality.

**When to Use:** Consult during any decision gate, especially during Calibration and Contrarian Analysis.

## Severity Levels

| Level | Description | Response |
|-------|-------------|----------|
| **Minor** | Bias present but unlikely to significantly affect outcome | Note and monitor; proceed with awareness |
| **Moderate** | Bias could meaningfully skew decision | Pause to address; apply specific correction technique |
| **Major** | Bias likely to cause serious decision error | Stop and remediate; consider restarting analysis from earlier gate |

## Common Biases

### 1. Confirmation Bias

**What it is:** The tendency to search for, interpret, and recall information that confirms pre-existing beliefs while ignoring contradictory evidence.

**Signs:**
- Evidence search focuses only on sources likely to support preferred conclusion
- Dismissing contradictory evidence as "outliers" or "unreliable" without rigorous evaluation
- Interpreting ambiguous information as supporting the favored option

**Correction:**
- Actively seek out the strongest arguments against your preferred position
- Assign someone the explicit role of devil's advocate with genuine authority to challenge
- Before concluding research, ask: "What evidence would change my mind?" and search specifically for it

**Severity:** Major - This bias is pervasive and can undermine the entire decision process

---

### 2. Anchoring

**What it is:** Over-relying on the first piece of information encountered (the "anchor") when making decisions, even when that information is arbitrary or irrelevant.

**Signs:**
- Initial estimates or numbers disproportionately influence final conclusions
- Adjustments from starting point feel insufficient even when evidence suggests larger changes
- First option presented becomes the implicit benchmark against which all others are judged

**Correction:**
- Generate estimates independently before seeing any reference numbers
- Consider multiple starting points and work from each separately, then compare conclusions
- Explicitly ask: "If we hadn't seen that first number/option, what would we think?"

**Severity:** Moderate - Common in negotiations and financial decisions; addressable with awareness

---

### 3. Overconfidence

**What it is:** Excessive certainty in one's own answers, judgments, and predictions; believing you know more than you actually do.

**Signs:**
- Confidence intervals are too narrow (e.g., 90% confidence ranges that are correct only 50% of the time)
- Dismissing uncertainty or risk factors as unlikely without rigorous probability assessment
- Treating assumptions as facts and projections as certainties

**Correction:**
- Explicitly state confidence levels and track calibration over time
- Require evidence proportional to certainty claims - extraordinary confidence requires extraordinary evidence
- Use pre-mortem technique: "Assume this failed - why did it fail?"

**Severity:** Major - Leads to inadequate risk management and surprise failures

---

### 4. Sunk Cost Fallacy

**What it is:** Continuing to invest in a decision because of previously invested resources (time, money, effort) rather than future value.

**Signs:**
- Arguments for continuing reference past investment rather than future returns
- Reluctance to abandon projects despite poor prospects because of "how far we've come"
- Framing quitting as "waste" rather than rational resource reallocation

**Correction:**
- Evaluate decisions based solely on future costs and benefits, ignoring past investment
- Ask: "If we were starting fresh today with no history, would we choose this path?"
- Set explicit kill criteria at the start of projects and honor them when triggered

**Severity:** Moderate - Common but correctable with discipline; watch for escalation of commitment

---

### 5. Availability Bias

**What it is:** Judging likelihood or frequency based on how easily examples come to mind, rather than actual probability.

**Signs:**
- Recent or vivid events disproportionately influence risk assessment
- Overweighting dramatic risks (plane crashes) while underweighting mundane ones (car accidents)
- Personal experience trumping statistical evidence

**Correction:**
- Seek base rate data before forming probability estimates
- Ask: "Is this example memorable because it's common or because it's unusual?"
- Deliberately search for examples that contradict easily recalled ones

**Severity:** Moderate - Particularly dangerous in risk assessment and resource allocation

---

### 6. Groupthink

**What it is:** Prioritizing group harmony and conformity over realistic evaluation of alternatives; suppressing dissent to maintain consensus.

**Signs:**
- Dissenting opinions quickly abandoned or not voiced at all
- Pressure (subtle or explicit) on members who question the group's direction
- Illusion of unanimity - assuming silence means agreement

**Correction:**
- Assign a formal devil's advocate role that rotates among members
- Collect opinions independently (written) before group discussion
- Leader speaks last; explicitly invite and reward constructive disagreement

**Severity:** Major - Can lead to catastrophic collective failures; requires structural safeguards

---

### 7. Framing Effects

**What it is:** Being influenced by how information is presented rather than the information itself; different descriptions of identical outcomes lead to different decisions.

**Signs:**
- Strong preference shifts depending on whether options are described as gains vs. losses
- Risk tolerance changes based on problem description rather than underlying probabilities
- Emotional reactions to presentation style override rational evaluation

**Correction:**
- Reframe the same decision multiple ways: as gain/loss, risk/opportunity, short-term/long-term
- Convert qualitative descriptions to quantitative terms where possible
- Ask: "Would my decision change if this were described differently?"

**Severity:** Moderate - Addressable through deliberate reframing exercises

---

### 8. Planning Fallacy

**What it is:** Systematically underestimating time, costs, and risks of future actions while overestimating benefits; believing your project is different from similar past projects that failed.

**Signs:**
- Estimates based on best-case scenarios rather than realistic or base-rate outcomes
- Ignoring or discounting data from comparable past projects
- Explanations for why "this time is different" without rigorous justification

**Correction:**
- Use reference class forecasting: find similar past projects and use their actual outcomes as baseline
- Apply systematic multipliers to initial estimates based on historical accuracy
- Conduct pre-mortem analysis before committing to timeline

**Severity:** Major - Causes chronic under-delivery and budget overruns; requires systematic correction

---

## Bias Audit Checklist

Use this checklist during Calibration and before finalizing decisions:

### Information Gathering
- [ ] **Confirmation Bias**: Have we genuinely sought evidence against our preferred option?
- [ ] **Availability Bias**: Are we relying on base rates rather than memorable examples?
- [ ] **Anchoring**: Have we considered this without reference to initial numbers or first options?

### Evaluation Process
- [ ] **Overconfidence**: Are our confidence levels calibrated to our actual evidence?
- [ ] **Framing Effects**: Have we reframed this decision multiple ways to test consistency?
- [ ] **Planning Fallacy**: Have we compared to reference class of similar past decisions?

### Group Dynamics
- [ ] **Groupthink**: Has genuine dissent been expressed and taken seriously?
- [ ] **Sunk Cost**: Are we evaluating based on future value, not past investment?

### Overall Assessment
- [ ] All Major-severity biases explicitly addressed
- [ ] Moderate-severity biases n