Principle of Independent Judgments

The Principle of Independent Judgments is the rule that multiple sources of evidence should be kept independent so that their errors do not become correlated. Its decisive significance is easy to miss: a group is not automatically wiser than its members. Collective accuracy depends on how judgments are produced before they are pooled. When observers make separate mistakes, those mistakes can cancel one another; when observers influence one another, the apparent size of the sample shrinks and a shared error may survive intact.

Definition and Logic

The principle appears in the instruction to “decorrelate error.” The underlying mechanism is illustrated by a large number of observers estimating the number of pennies in glass jars. Individuals perform poorly: some overestimate and others underestimate. Yet when their estimates are averaged, the result tends to be accurate because the observers have a common basis—the same jar—but their individual errors are independent. In the absence of a systematic bias, those errors tend to average toward zero.

Independence is therefore the condition that makes aggregation useful, not a decorative feature of group procedure. If observers share a bias, averaging cannot remove it. If they are allowed to influence one another, their judgments become more alike, effectively reducing the number of genuinely independent observations. The group may then sound more confident while containing less information.

The rule belongs to a wider account of judgment under uncertainty. Research on representativeness, availability, and anchoring showed that intuitive heuristics are economical and often effective, yet can produce systematic and predictable errors. Independent judgments do not eliminate those biases automatically, but they can prevent one person’s intuition from becoming the template that organizes everyone else’s response.

Applications to Evidence and Meetings

A practical application appears in police procedure. When several witnesses observe an event, they should not discuss what happened before giving their testimony. The purpose is not merely to prevent hostile witnesses from colluding. Even unbiased witnesses can contaminate one another’s recollections: after exchanging experiences, they tend to reproduce similar errors, reducing the total value of their evidence. Eliminating redundancy from the sources preserves the diversity that makes comparison informative.

The same reasoning applies to meetings. Before an issue is discussed, each committee member should write a very brief summary of their position. Written judgments collected before conversation preserve differences in knowledge and opinion. By contrast, open discussion gives disproportionate weight to people who speak early and assertively, encouraging others to line up behind them before their own views have been independently formed.

The principle also explains a change in grading practice. A grading method that evaluated a student’s answers in a fixed sequence allowed the first answer to influence judgments of later answers. That procedure felt coherent and produced cognitive ease, but its consistency was spurious: it encouraged an overall impression, or halo effect, to spread across separate evaluations. Grading each question across all students before moving to the next exposed greater inconsistency. The discomfort was evidence of real uncertainty in both the questions as measures of knowledge and the grader’s own judgments, rather than proof that the new procedure was inferior.

Relation to Decision Theory

Independent judgment is closely connected to the normative principles of rational choice, especially dominance and invariance. Dominance requires preferring an option that is at least as good in every respect and better in at least one. Invariance requires that equivalent prospects receive the same preference even when they are described differently. Yet experiments showed that people often prefer risk-averse choices when outcomes are framed as lives saved and risk-seeking choices when the same outcomes are framed as lives lost. The effect persisted among sophisticated respondents and remained even when the same people answered both versions within minutes.

These framing effects resemble perceptual illusions more than ordinary computational mistakes. Respondents could recognize that their answers conflicted and still experience each framing as compelling. The lesson is not that confidence guarantees coherent judgment, but that preferences should be tested by deliberately restating a decision in more than one form. Independent evaluation helps protect this test from premature convergence: decision makers must first examine the alternatives separately, then compare their conclusions rather than inherit a shared framing from the discussion.

The same discipline is required when evidence is combined with prior beliefs. Bayesian reasoning demands that base rates be combined with the diagnosticity of case-specific evidence, rather than allowing a vivid description to overwhelm the prior probability. For example, if 3 percent of graduate students study computer science and a description is four times more likely for a computer-science student than for students in other fields, the resulting probability is 11 percent, not whatever impression the description immediately evokes. Separating the prior, the evidence, and the final judgment is another way of resisting the coherence of a prematurely shared story.

Organizational Use and Consequences

Organizations are especially well placed to apply the principle because they can impose procedures that individuals cannot reliably impose on themselves. A decision can be treated as a production process with quality checks at three stages: framing the problem, collecting relevant information, and reflecting on the decision afterward. Checklists, reference-class forecasting, independent preliminary opinions, and the premortem can make these checks routine rather than reserving them for the aftermath of disaster.

The premortem, proposed by Gary Klein, asks knowledgeable participants to imagine that a plan has been implemented and that, one year later, the outcome was a disaster. They then write a brief history of what went wrong. Because the exercise is conducted before formal commitment, it gives dissent and competing explanations a protected place in the process. It complements independent judgments by making disagreement informative before group consensus has made errors harder to see.

The broader institutional implication is that an organization is a factory that manufactures judgments and decisions. Quality control must therefore operate in initial design, fabrication, and final inspection. Efficient meetings are part of that hygiene: collecting positions before discussion prevents authority and speaking order from determining the apparent consensus, while later review can distinguish random mistakes from systematic failures in the procedure.

Limits and Significance

The principle does not promise perfect accuracy. Independent judgments can still share a common bias, and aggregation cannot correct an error that everyone has inherited from the same misleading frame, base-rate neglect, or institutional assumption. Independence is a defense against correlated error, not a substitute for sound evidence or disciplined reasoning.

Its importance lies in making the social production of knowledge visible. Intuitive thinking can answer complex questions by substituting easier ones, and a coherent first impression can quietly organize later judgments. The Principle of Independent Judgments interrupts that sequence. It protects disagreement before consensus, preserves the informational value of multiple witnesses, and turns collective decision making from an exercise in mutual reinforcement into a controlled attempt to make errors cancel rather than multiply.