Separate evaluation and joint evaluation

Single evaluation and joint evaluation are contrasting modes of judgment in which the same option can acquire a different value depending on whether it is considered alone or alongside an alternative. The distinction is central to the study of heuristics and biases because it shows that judgment is shaped not only by the facts of a case but also by which comparisons the presentation makes available. In isolated evaluation, System 1 often substitutes an easier emotional or salient question for the intended one; in joint evaluation, comparison can bring neglected attributes into view and engage System 2.

Definition and Cognitive Contrast

Single evaluation occurs when a person sees only one option or scenario and assigns it a value without a directly presented comparison. Joint evaluation occurs when two or more options are considered together, allowing their differences to be inspected explicitly. The contrast fits the broader account of question substitution: when a difficult target question does not receive a satisfactory answer quickly, System 1 answers a related heuristic question instead. Salience, emotional intensity, and the ease with which a story comes to mind can then stand in for a more appropriate standard of value or probability.

This contrast also belongs to the research program developed by Daniel Kahneman and Amos Tversky. Their work documented how representativeness, availability, and adjustment from an anchor provide economical shortcuts that are often useful but can generate systematic and predictable errors. The same program connected psychological description with normative standards. Bayesian reasoning requires prior beliefs to be combined with the diagnosticity of evidence, yet intuition often exaggerates the evidence and neglects base rates. Decision theory likewise requires dominance and invariance, although the way a choice is presented can make those principles psychologically difficult to follow.

Moral Judgment and Emotional Salience

The difference becomes vivid in the compensation scenarios designed by Dale Miller and Cathy McFarland. A man loses the use of his right arm after being shot while entering a robbery at a convenience store. In one version, he is shot in his regular store; in another, his regular store is closed for a funeral, so he shops at a store he rarely visits and is shot there instead. When both scenarios are presented together, nearly everyone agrees that the compensation should be identical. The injury is the same, and the store’s identity is not regarded as a morally relevant factor.

Joint evaluation makes that principle inspectable. Comparing the scenarios allows people to ask directly whether the location should affect compensation, and the deliberate comparison brings System 2 into the judgment. The moral rule is not hidden: compensation is for the crippling injury, not for the poignancy of the circumstances.

The result changes when each scenario is evaluated by a different person. In the between-subjects experiment, participants assigned a substantially larger dollar amount when the victim was shot in the rarely visited store than when he was shot in his regular store. The difference was driven by poignancy, a counterfactual feeling akin to regret. The thought that the victim might have avoided the shooting by shopping at his regular store comes readily to mind, and the emotional intensity of that thought is translated onto a monetary scale through substitution and intensity matching.

The comparison between the two modes reveals an unsettling instability. Nearly everyone rejects poignancy as a legitimate basis for compensation when the alternatives are visible together, yet the same principle can shape an isolated judgment when the contrasting case is absent. In ordinary life, people generally encounter situations one at a time. WYSIATI therefore leaves the emotionally available facts to govern the response, while moral principles that seem compelling in reflection may fail to regulate immediate emotional reactions.

Preference Reversals in Risky Choice

The same structure appears in the preference-reversal experiments associated with Sarah Lichtenstein and Paul Slovic. Participants first choose between two roulette-wheel bets. Bet A offers an 11/36 chance to win $160 and a 25/36 chance to lose $15. Bet B offers a 35/36 chance to win $40 and a 1/36 chance to lose $10. When choosing between them, participants generally favor the safer Bet B: its almost certain modest gain is more attractive than Bet A’s small chance of a larger prize combined with a high probability of losing.

The judgment changes in single evaluation. When participants imagine owning only one bet and state the lowest price at which they would sell it, the prize associated with Bet A becomes salient. Selling prices are higher for Bet A than for Bet B, even though participants had preferred Bet B in the direct choice. Thus a person can choose B over A while valuing A more highly when each bet is considered separately.

Joint evaluation focuses attention on the difference in safety, whereas single evaluation permits anchoring on the prize. The feature that drives the isolated valuation is therefore not the feature that governs the comparison. System 1 is more likely to control the single evaluation through emotional response and salience; joint evaluation requires a more effortful assessment of the alternatives and consequently calls for System 2.

The reversal can be exposed in a within-subject design in which participants price both bets and also choose between them. Participants are often unaware of the inconsistency until it is pointed out. In the classic interview described in the research, a bewildered participant chose one bet, then was willing to pay to exchange it for the rejected bet, and repeatedly moved through the same cycle. The conflict is not simply a failure to calculate; it is produced by the different mental operations invited by choosing between options and valuing an option in isolation.

Relation to Framing and Rational Principles

Single and joint evaluation illuminate the broader failure of invariance. Rational choice requires that equivalent prospects receive the same preference order even when their descriptions differ. In the disease-treatment problems, respondents preferred a sure saving of 200 lives over a one-third chance of saving 600 when outcomes were framed as lives saved. When the objectively equivalent outcomes were described as deaths, they preferred the gamble: a one-third probability that nobody would die rather than a sure loss of 400 lives. The same respondents could be risk averse in the gain frame and risk seeking in the loss frame.

The failure was pervasive and robust. It appeared among sophisticated as well as naive respondents and remained even when the same people answered the two versions within minutes. Participants were puzzled by their conflicting answers, wishing simultaneously to preserve both their frame-dependent preferences and the principle of invariance. The effect resembles a perceptual illusion more than a simple computational mistake: once a frame establishes a reference point, the psychological meaning of the outcomes changes.

Joint evaluation can expose inconsistencies that single evaluation conceals, but it cannot guarantee rationality in every setting. In concurrent decisions involving gains and losses, respondents favored a combination that was actually dominated by the rejected combination. Their choices reflected the S-shaped value function and the tendency toward risk aversion for gains and risk seeking for losses. When the separate choices were combined, the apparently preferred conjunction produced a prospect no better than the alternative and worse in at least one respect, violating dominance.

The normative difficulty is therefore structural. Invariance is essential and intuitively compelling, yet psychologically unfeasible in many ordinary decisions. One possible remedy is to transform equivalent problems into a canonical representation based on total assets rather than gains and losses, but intuitive computation struggles to compound the outcomes of concurrent decisions. An actuarial evaluation may help with some questions involving human lives, but it is inadequate for financial choices and cannot be applied to outcomes without an objective metric. Deliberately presenting a decision in more than one frame is consequently recommended as a practical test of preference robustness.

Experience, Memory, and Choice

The distinction extends beyond monetary and moral judgment to the evaluation of experiences. In the colonoscopy study designed by Daniel Kahneman and Don Redelmeier, 154 patients reported their momentary pain every 60 seconds on a scale from zero to 10. One illustrated procedure lasted eight minutes and another 24 minutes; the longer procedure had at least as much pain as the shorter one at every level, so its momentary pain total, understood as the area under the curve, was clearly greater.

Retrospective evaluation did not reproduce those momentary reports. Global ratings were well predicted by the average of the worst pain and the pain at the end, a pattern known as the peak-end rule. Duration neglect meant that the length of the procedure had no effect on the remembered total pain. Patient A and patient B both reached a peak of 8, but the final reading was 7 for A and only 1 for B. Their peak-end averages were therefore 7.5 and 4.5, so the shorter procedure left the worse memory merely because it ended badly.

The cold-hand experiment made the conflict between experience and retrospective choice explicit. Participants immersed one hand for 60 seconds in water at 14 degrees Celsius. In the 90-second trial, the first 60 seconds were identical, but the water then became roughly one degree warmer during an additional 30 seconds, producing a slight reduction in pain. From the perspective of the experiencing self, the longer trial was plainly worse because it contained all of the short trial plus 30 seconds of continued immersion.

When asked which experience they would repeat, however, 80 percent of participants who reported a reduction in pain during the final phase chose the longer trial. The peak-end rule made its memory more favorable, while duration neglect muted the extra half-minute. The participants were not seeking pain and understood which exposure had lasted longer; they simply chose the episode they remembered as less aversive. Here, single and comparative evaluation operate through remembered rather than momentary utility: the option preferred for repetition is the one that feels better in memory, not necessarily the one that was better while being experienced.

Significance for Judgment

Single evaluation and joint evaluation show why preferences cannot be treated as fixed properties of options. A lonely scenario invites substitution, emotional intensity matching, anchoring, and the neglect of absent alternatives. A comparison can suppress those influences by making a relevant dimension explicit, but it can also introduce a new frame or reference point that changes the psychological meaning of the outcomes. The resulting reversals are not isolated curiosities; they recur in compensation, betting, public-health decisions, and the remembered evaluation of pain.

The concept therefore joins the descriptive and normative strands of the book’s argument. Psychological judgment is economical and often effective, yet its shortcuts create predictable departures from Bayesian reasoning, dominance, and invariance. Joint evaluation offers a practical corrective because it forces alternatives into the same mental field, where overlooked differences and contradictions can be inspected. But the deeper lesson is more sobering: even confidence in a preference does not establish that the same choice would survive another presentation. Deliberate comparison and reframing are not guarantees of rationality, but they are among the few available ways to discover when a judgment belongs less to the object being evaluated than to the mode of evaluation itself.