Principle of Diminishing Sensitivity

The Principle of Diminishing Sensitivity is a central rule of prospect theory: subjective responses become less sensitive as gains or losses grow in magnitude. The first increment from a low baseline is felt sharply, while an equal increment added to an already large amount produces a smaller psychological difference. Applied to risky choice, the principle makes the value function curved rather than linear and helps explain why people do not evaluate outcomes as simple additions to total wealth.

Definition and Reference Point

The principle operates relative to a neutral reference point, often the status quo but sometimes an expected or felt-entitled outcome. Outcomes above that point are encoded as gains; outcomes below it are encoded as losses. The same physical result can therefore acquire a different psychological meaning when the reference point changes. The basic perceptual demonstration uses three bowls: ice water, warm water, and room-temperature water. After each hand is adapted to a different temperature, the same middle bowl feels hot to one hand and cold to the other. Financial outcomes follow the same logic: evaluation concerns changes relative to a starting condition, not merely final wealth.

The rule applies on both sides of the reference point. Turning on a weak light produces a large perceptual change in a dark room, whereas the same increment may be barely noticeable in a brightly illuminated room. Likewise, the subjective difference between $100 and $200 is greater than the difference between $900 and $1,000, even though both numerical changes equal $100. For losses, the difference between losing $100 and losing $200 is experienced as larger than the difference between losing $1,100 and losing $1,200. In the formal account associated with prospect theory, gains and losses are therefore the effective carriers of value rather than ultimate states of wealth.

The Value Function and Risk

The principle is represented by the S-shaped value function of prospect theory. The curve bends on the gain side and on the loss side, expressing diminishing sensitivity as departures from the reference point become larger. The shape is not symmetrical, however, because loss aversion makes losses loom larger than corresponding gains. Diminishing sensitivity describes the curvature within each domain; loss aversion describes the abrupt difference between the two domains at the reference point.

This curvature produces different risk attitudes for gains and losses. In the domain of gains, diminishing sensitivity contributes to risk aversion: a sure gain can feel preferable to a gamble with the same monetary expectation. In the domain of losses, the same principle can produce risk seeking. The pain of losing $900 is more than 90 percent of the pain of losing $1,000, so avoiding a sure loss may feel less attractive than accepting a gamble that offers a chance of escaping the loss altogether. A typical example compares an 85 percent chance of losing $1,000, with a 15 percent chance of losing nothing, to a sure loss of $800. Although the gamble has the worse expected outcome, many people prefer it.

The contrast is not a contradiction between two unrelated tendencies. In a mixed gamble containing both a possible gain and a possible loss, loss aversion strengthens the negative part of the prospect and can produce extreme caution. In a choice between two losses, diminishing sensitivity bends the loss curve in the opposite behavioral direction, making the gamble attractive relative to the sure loss. The distinction between these mechanisms is essential to understanding the fourfold pattern of risk preferences.

Connection with Regression and Prediction

The principle also belongs to the broader psychological problem of mistaking changes in performance for changes in underlying ability. In discussions of regression to the mean, an observed result is divided into shared factors and factors specific to that result. A highly unusual score contains both information about talent and an element of luck or circumstance; a later prediction should therefore move back toward the average rather than simply repeat the extreme observation.

The flight-instructor example shows how failure to recognize this structure distorts judgments about training. An instructor reported that cadets praised after an excellent aerobatic maneuver usually performed worse on their next attempt, whereas cadets reprimanded after a poor attempt often improved. The apparent success of punishment and failure of reward were largely consequences of regression: unusually good performances were likely to be followed by less extreme performances, and unusually bad ones by better performances. Diminishing sensitivity is not identical to regression to the mean, but both challenge the intuition that every visible change has a proportionate cause. The psychological response to an outcome and the prediction of its recurrence must each account for the unstable, nonshared components of observation.

The same logic appears in the example of intelligence and marriage. The claim that highly intelligent women tend to marry less intelligent men invites causal stories about competition, compromise, or preference. Yet if spouses' intelligence scores are less than perfectly correlated, and men and women do not differ on average, regression toward the mean mathematically entails that exceptionally intelligent women will have husbands who are, on average, less intelligent than they are, with the reverse pattern for exceptionally intelligent men. The interesting-looking social claim is therefore equivalent to a routine statistical statement about imperfect correlation.

Consequences for Rational Choice

The rule exposes a gap between descriptive psychology and normative decision theory. Rational-choice analysis invokes dominance and invariance: an option that is at least as good in every respect should be preferred, and equivalent prospects should receive the same preference regardless of wording. Human judgments frequently fail these requirements because reference points and curved value responses change how outcomes are mentally represented.

The classic disease problem illustrates the conflict. When programs are described as saving lives, respondents tend to prefer saving 200 people for certain over a one-third chance of saving all 600, displaying risk aversion. When the identical outcomes are described as 400 people dying for certain versus a one-third chance that nobody dies and a two-thirds chance that all 600 die, respondents become risk seeking and favor the gamble. The objective consequences are equivalent, but the first frame presents gains from a reference state in which 600 people die, while the second presents losses from a reference state in which nobody dies. The failure of invariance remains common even among sophisticated respondents, who can recognize the equivalence and still retain their conflicting preferences.

A related pair of decisions can generate a violation of dominance. Respondents often prefer a sure gain in one decision combined with a gamble in another, even though combining the rejected alternatives would produce a prospect that dominates the preferred combination. The S-shaped value function and susceptibility to framing jointly create this inconsistency. Since confidence in a choice does not guarantee that the same choice would survive a change of frame, a practical safeguard is to restate a decision in more than one way and test whether the preference remains stable.

Significance in Prospect Theory

The Principle of Diminishing Sensitivity supplies one of the three major operating features of prospect theory's automatic evaluation system, alongside reference-point dependence and loss aversion. It replaces the simple assumption that equal monetary increments have equal psychological value with a richer account in which the location of an outcome matters. A gain of $100 is not psychologically the same event when it moves a person from $100 to $200 as when it moves that person from $900 to $1,000; the numerical increment is constant, but its experienced impact is not.

The principle also explains why a theory based only on total wealth is psychologically incomplete. Economic theory traditionally described a small bet as a movement from current wealth to a higher or lower wealth state, while prospect theory describes it as a gain, loss, or neutral outcome relative to a reference point. That shift makes room for the curvature of subjective value, the contrast between risk aversion and risk seeking, and the frame-dependent reversals that rational-choice principles cannot reliably prevent. Its importance lies precisely in showing how a simple perceptual regularity can organize apparently contradictory decisions across money, physical sensation, performance evaluation, and public-health choices.