Cognitive Bias in CTI: A Practical Review Checklist Before You Publish
Find confirmation bias, anchoring, mirror imaging, recency effects, and group pressure before they distort a consequential intelligence judgment.
Bias is not a character flaw reserved for careless analysts. It is a normal feature of reasoning under time pressure, incomplete evidence, deception, and organizational expectations. The practical response is to design moments where an assessment can be challenged.
Use this checklist before publishing a consequential judgment or during a rapid update. Spend review time in proportion to the cost of being wrong, not the length of the product.
Check the Frame
Ask whether the question presumes an answer. “Which state group did this?” excludes criminals, insiders, false flags, and unrelated events. Rewrite it as “Which explanation best accounts for the activity?”
Identify the decision, deadline, and consequence of error. Check for mirror imaging: the adversary may not value, organize, or calculate as your organization does. Name assumptions about motive, access, capability, and rationality.
Check What Received Attention
List evidence that contradicts the leading view before listing more support. Trace repeated reports to their original source. Ask whether recent, vivid, or famous incidents have crowded out base rates and older evidence. Check whether the first label anchored later analysis.
Distinguish “not observed” from “did not occur.” Review collection coverage before treating absence as evidence.
Check the Team and the Language
Ask the least-senior reviewer first, invite a written alternative, and prevent the drafter from defending the assessment until challengers finish. Separate source reputation from claim quality. If consensus arrived unusually quickly, test whether hierarchy or prior messaging constrained debate.
Review estimative terms. Remove certainty borrowed from vendor branding, actor fame, or precise-looking scores. The ACH guide provides a fuller alternative test.
Record the Check That Matters
Do not add a ceremonial checklist to every document. Record the leading alternative, strongest contradiction, critical assumption, dissent, and change trigger. Those entries show whether review influenced the judgment.
After the decision, compare forecast with outcome and examine process rather than blame. A learning record helps the team discover recurring blind spots and design better collection and review gates.
Frequently asked questions
Can a checklist eliminate cognitive bias?
No. It makes common failure modes more visible and creates opportunities to challenge them before a decision.
What is confirmation bias in CTI?
It is the tendency to seek or interpret evidence in ways that support an existing explanation while discounting contradictory evidence.
Do analytic tools prevent bias?
Tools can improve consistency and provenance, but automation can also scale biased assumptions or give weak scores an undeserved appearance of objectivity.
Is structured review practical during a fast incident?
Yes. Use a short version focused on alternatives, contradictory evidence, critical assumptions, and the consequence of being wrong.
How should dissent be handled?
Record the alternative judgment and evidence, resolve factual disagreements where possible, and show decision-makers material uncertainty rather than forcing false consensus.
