Why 99% Does Not Feel Like 100%: The Certainty Effect

A sure thing feels special, and the jump from 99% to 100% feels larger than the jump from 60% to 61%. Here is why, and what it means for decisions.

Imagine a clinic offers you a treatment that works for certain. Now imagine the same treatment, but it works 99 times out of 100. The one percent difference is small on paper, yet for many people the certain version feels far more attractive. That gap is the certainty effect, and it is one of the clearest ways that real choices depart from the standard model.

Start with the benchmark. Expected utility theory, the workhorse model of choice under risk, treats probabilities linearly. A move from 0.60 to 0.61 should matter exactly as much as a move from 0.99 to 1.00, because both are a one-point change in the chance of a good outcome. The model is clean and often useful, and it is the right place to begin.

Now the behaviour. People routinely treat certainty as special. Consider a simple pair of choices. First, a sure A$300 against an 80% chance of A$400. Many people take the sure A$300, even though the gamble is worth A$320 on average. Second, a 25% chance of A$300 against a 20% chance of A$400. Here many of the same people switch to the A$400 option. The two problems have the same ratio of probabilities; only the presence of a sure thing has changed. Under expected utility the ranking should not flip, but it does. The pull of certainty in the first problem, and its absence in the second, drives the switch.

This is descriptive, not normative. The point is not that preferring certainty is irrational, nor that people should be argued out of it. The point is that a model which assumes linear probabilities will mispredict these choices, and that the error is systematic rather than random. Prospect theory captures the pattern with a probability weighting function in which the very top of the scale, near certainty, is treated differently from the middle.

The practical implications are real. Framing an option as guaranteed, when it genuinely is, can change uptake more than a small change in the underlying odds. In health, “eliminates the risk” reads very differently from “reduces the risk to almost nothing,” even when the numbers are close. That is useful to know when communicating, and it is something to watch for when a product or message leans on the word “guaranteed” to do heavy lifting.

The limitation is worth stating plainly. The certainty effect describes a tendency, not a law, and its size varies with context, stakes, and how a choice is presented. It tells us to be careful with near-certain probabilities; it does not tell us that people ignore probabilities in general.

If you want the mechanics behind this, the ECON3111 resources on choice under risk and uncertainty work through the certainty effect with numbers, and the post on probability weighting shows the other side of the same function.

Key references. Kahneman, D. and Tversky, A. (1979). Prospect theory: an analysis of decision under risk. Econometrica. Tversky, A. and Kahneman, A. (1992). Advances in prospect theory: cumulative representation of uncertainty. Journal of Risk and Uncertainty.