Cognitive Bias

Systematic errors in human perception, reasoning, and judgment caused by mental shortcuts (heuristics) and unconscious assumptions. The primary human-factors problem that Structured Analytic Techniques are designed to counter.

Theoretical root: System 2 — Daniel Kahneman and Amos Tversky’s dual-process framework explains why biases arise (System 1 heuristics substituting for System 2 reasoning) and why structured techniques counter them (by forcing System 2 engagement).

LLM-native failure modes are also tracked here: Sycophancy and Hallucination — these don’t map to single human biases but are the LLM equivalents that SAT-structured prompting most directly counters.

This page is a hub — individual bias pages with full references are listed below.

---

Bias Library

Foundational (Kahneman/Tversky Heuristics-and-Biases Program)

Bias	Short Description	Primary SAT Counter
Anchoring Bias	Over-reliance on first information; insufficient adjustment	ACH, Key Assumptions Check
Availability Heuristic	Probability judged by ease of recall; vivid/recent events overweighted	ACH, High-Impact/Low-Prob, Alt Futures
Overconfidence Bias	Confidence intervals too narrow; certainty exceeds warrant	Key Assumptions Check, Quality of Info, High-Impact/Low-Prob
Framing Effect	Logically equivalent presentations produce different judgments	Outside-In Thinking, Alt Futures, Brainstorming
Status Quo Bias	Preference for current state; departures perceived as losses	What If?, Alt Futures, Devil’s Advocacy

Social and Motivational

Bias	Short Description	Primary SAT Counter
Confirmation Bias	Seek/weight confirming evidence; discount disconfirming	ACH (disconfirmation focus), Devil’s Advocacy
Groupthink	Group cohesion overrides realistic appraisal	Devil’s Advocacy, Team A/Team B, Brainstorming
Motivated Reasoning	Reasoning works backward from desired conclusion	ACH, Devil’s Advocacy, Key Assumptions Check

Analytical/Perceptual

Bias	Short Description	Primary SAT Counter
Mirror Imaging	Project own decision logic onto adversaries	Red Team Analysis, Outside-In Thinking
Hindsight Bias	Outcomes perceived as predictable in retrospect	What If? (pre-mortem), Indicators/Signposts
Mind-Set	Fixed mental model filters/distorts incoming information	All SATs; Key Assumptions Check most directly

LLM-Native Failure Modes

Bias	Short Description	Primary SAT Counter
Sycophancy	Agreement with user preference overrides accuracy; RLHF artifact	Devil’s Advocacy, Team A/Team B, ACH
Hallucination	Confident fluent output with no factual grounding; no internal uncertainty signal	Quality of Info Check, Key Assumptions Check

---

Original Tradecraft Primer Taxonomy

CIA Tradecraft Primer (2009) organizes biases into four categories:

Perceptual Biases:

• Expectations → Mind-Set
• Resistance → Status Quo Bias
• Ambiguities → related to Framing Effect

Biases in Evaluating Evidence:

• Consistency → Confirmation Bias
• Missing Information → Overconfidence Bias (doesn’t know what it doesn’t know)
• Discredited Evidence persistence → Status Quo Bias, Motivated Reasoning

Biases in Estimating Probabilities:

• Availability → Availability Heuristic
• Anchoring → Anchoring Bias
• Overconfidence → Overconfidence Bias

Biases in Perceiving Causality:

• Rationality → Mirror Imaging
• Attribution → Mirror Imaging

---

Sources

• CIA Tradecraft Primer (2009)
• Riley: SATs in Cybersecurity (2024)
• Roberts: LLM SATs FTW (2025) (LLM-native failure modes)