Evidence and Confidence

Evidence Is Support, Not Vibe

Evidence is whatever should rationally move confidence. It may be direct observation, measurement, testimony, memory, statistical frequency, background theory, expert consensus, or the absence of an expected signal. Evidence is not the same as emotional vividness. A story can feel powerful and still be weak evidence; a base rate can feel dull and still matter enormously.

The discipline of credencing asks whether confidence is scaled to support. The central question is not "Can I tell a story in favor of this?" but "How much should this information move a careful agent?"

Confidence Is a Response

Confidence is not a virtue by itself. Confidence is admirable only when it is well-calibrated. Underconfidence can prevent action, learning, and justified trust. Overconfidence can produce dogmatism, risk inflation, bad medicine, bad policy, and personal self-deception.

A calibrated agent is not always cautious. Sometimes the evidence really is strong. The aim is not permanent hesitation; the aim is proportion.

The Path From Evidence to Credence

The model divides the path into two stages. First, objective evidence becomes perceived evidence. This stage can fail through missing data, bad measurement, poor statistical tools, misleading testimony, fear, fatigue, propaganda, or a narrow frame. Second, perceived evidence becomes assigned credence. This stage can fail when the agent knows, at some level, what the evidence indicates but chooses a more comfortable or identity-preserving confidence instead.

This division is the heart of the site. It lets us ask whether a person needs better information, better tools, better incentives, greater courage, or some combination of all four.

Evidence Quality and Evidence Direction

Evidence has both quality and direction. Direction tells us which claim it supports. Quality tells us how much weight it deserves. A single anecdote, a controlled trial, a replicated meta-analysis, and a coerced confession all point in some direction, but they should not carry the same weight.

Good credencing therefore requires both sensitivity and restraint: sensitivity to the direction of support, restraint about how much confidence the support can bear.

Adjacent References

These are starting points rather than exhaustive citations. They connect this page to nearby work in probability, calibration, cognitive bias, and social epistemology.

• |Thomas Bayes, "An Essay towards solving a Problem in the Doctrine of Chances" (1763).
• |E. T. Jaynes, Probability Theory: The Logic of Science (2003).
• |Daniel Kahneman and Amos Tversky, "Judgment under Uncertainty: Heuristics and Biases" (1974).
• |Sarah Lichtenstein, Baruch Fischhoff, and Lawrence D. Phillips, "Calibration of Probabilities" (1982).