Reversals

Kahneman examines preference reversals — cases where the same person ranks two options one way when they are judged separately and the opposite way when they are judged side by side. The mode of evaluation, single or joint, changes the answer, which is impossible under the classical assumption that preferences are stable and internally consistent. The reversals reveal that our judgments are constructed in the moment and depend heavily on what is available for comparison.

The mechanism is the 'evaluability' of attributes. In single evaluation, System 1 responds to whatever is emotionally vivid and easy to assess, while attributes that mean little in isolation are ignored; in joint evaluation, System 2 can compare the options on dimensions that only become meaningful through contrast. An attribute that is hard to judge alone — the number of entries in a dictionary, the scale of a harm — barely registers in single evaluation but carries real weight once two options sit beside each other for comparison.

Kahneman illustrates with charitable causes. Asked in isolation, people may feel more for endangered dolphins than for farmworkers exposed to a skin-cancer risk, because the dolphins evoke a vivid emotional response while the human risk is abstract. But evaluated jointly, the comparison shifts judgment toward the humans, whose plight now seems more deserving when set against the animals. The cause people 'prefer' depends entirely on whether they consider it alone or alongside an alternative.

He extends the analysis to the law, where preference reversals produce real injustice. Juries assessing punitive damages in isolated cases return wildly inconsistent awards, because each is anchored to the emotional intensity of its own facts with no comparison available; the same cases, evaluated jointly against one another, yield far more coherent and proportionate awards. The single-case format of trials, Kahneman notes, builds incoherence into the justice system by denying decision-makers the comparisons that would discipline their judgments.

The applied takeaway is to seek joint evaluation for important comparative decisions. Judging options one at a time invites the vivid, emotional, single-attribute responses of System 1 and produces inconsistent choices; lining up alternatives side by side engages comparison and surfaces the attributes that matter but are hard to assess alone. When consistency and fairness count — in hiring, purchasing, or judging — evaluate the candidates together rather than in isolation.

Kahneman's deeper observation is that preference reversals are decisive evidence that preferences are not pre-formed and retrieved but constructed on the spot, shaped by the framing and the comparison set. There is often no single 'true' preference waiting to be measured; the answer depends on how the question is posed. This undermines the bedrock economic assumption of stable, coherent preferences and shows that rationality, far from being automatic, requires the deliberate adoption of a broad, comparative frame that the mind does not supply on its own.