Consider a puzzle from Christopher Hsee's research. When evaluated separately, people will pay more for a dinnerware set with 24 intact pieces than for a 31-piece set that includes 7 broken items. The larger set is objectively better—it contains more usable pieces—yet it commands a lower price.

This anomaly violates a basic principle of rational choice: adding options or items should never make something worse. Yet across domains—from product pricing to medical decisions to job offers—we observe that more information, more features, and more choices frequently produce worse evaluations and outcomes.

Understanding when and why less becomes more requires moving beyond the assumption that people are unboundedly rational information processors. The less-is-more effect emerges from predictable interactions between human cognitive constraints and the structure of decision environments. Mapping these conditions offers practical guidance for anyone designing choices—whether for consumers, employees, patients, or themselves.

The evaluability hypothesis, developed by Christopher Hsee, explains a counterintuitive feature of judgment: the same attribute can be weighted differently depending on whether options are evaluated separately or jointly. Attributes that are difficult to assess in isolation get systematically underweighted when only one option is visible.

Consider a music dictionary with 10,000 entries and a torn cover versus one with 20,000 entries and a pristine cover. Evaluated separately, consumers pay more for the smaller intact dictionary because cover condition is easy to assess while entry count is meaningless without comparison. Place the two side by side, however, and the entry difference becomes evaluable—prices reverse.

This produces the less-is-more effect through a specific mechanism. When an option contains a hard-to-evaluate strength (like 31 pieces of dinnerware) alongside an easy-to-evaluate weakness (broken items), separate evaluation amplifies the weakness while neutralizing the strength. Removing the broken pieces—reducing the objective set—improves perceived value.

The implication extends beyond pricing experiments. Job candidates evaluated one at a time are judged on superficial cues; resumes reviewed in batches reveal genuine quality differences. The mode of evaluation, not just the information itself, determines what counts.

Takeaway

What people weight depends on what they can evaluate. An attribute with no comparison context tends to vanish from judgment, regardless of its objective importance.

Decision quality does not scale linearly with information volume. Research on information overload demonstrates an inverted-U relationship: accuracy improves with information up to a threshold, then deteriorates as additional inputs introduce noise, increase cognitive load, and dilute attention to diagnostic features.

Paul Slovic's classic studies of horse race handicappers showed that providing experts with 40 pieces of information produced no better predictions than providing 5—but it dramatically increased their confidence. This dissociation between accuracy and confidence is dangerous: more information can make us worse decision-makers while convincing us we are better ones.

The mechanism involves signal-to-noise ratios. Each additional variable that lacks predictive validity competes with diagnostic variables for attention and weight in the judgment process. When weak cues are added to strong ones, decision-makers often dilute the weight placed on genuinely informative features—a phenomenon called the dilution effect.

Medical decision-making illustrates the stakes. Studies of diagnostic accuracy show that physicians presented with extensive but largely irrelevant patient history sometimes perform worse than those given only key symptoms. The relevant question is not whether information is true but whether it changes the optimal action.

Takeaway

More data is not more knowledge. Information that does not change a decision adds cognitive cost without benefit—and may actively obscure the signals that should drive your judgment.

If less information sometimes produces better decisions, the practical question becomes: how do we determine optimal information levels and presentation? Three principles emerge from the behavioral evidence.

First, match information to evaluability conditions. When users will compare options jointly, complete attribute information helps them weight differences correctly. When evaluation will be separate or sequential, focus on attributes whose meaning is self-evident, or provide reference points (averages, benchmarks, percentile ranks) that make abstract numbers interpretable.

Second, distinguish decision-relevant from decision-irrelevant information. Before adding a feature, attribute, or data point, ask whether different values would change the recommended action. Information that fails this test is decoration at best and dilution at worst. Choice architects in fields from retirement plans to pharmaceutical disclosures have improved outcomes by stripping away non-diagnostic content.

Third, sequence information to reduce simultaneous load. Progressive disclosure—revealing details only when relevant to the current decision stage—respects working memory limits while preserving access to depth. The goal is not minimalism but precision: every piece of information should earn its place by improving the decision it enables.

Takeaway

Designing information is itself a decision. The default of 'include everything' assumes recipients have unlimited processing capacity—an assumption that no behavioral evidence supports.

The less-is-more effect is not a paradox once we abandon the fiction of unlimited cognitive capacity. Judgment operates within evaluability constraints, attention budgets, and working memory limits. Information interacts with these constraints, sometimes productively and sometimes destructively.

For practitioners, the insight reframes the design challenge. The question is not 'what could I provide?' but 'what will improve the decision I am supporting?' Subtraction is a tool, not a deficiency.

The deeper lesson concerns humility about complexity. Adding options and information often feels like generosity. Frequently, it is the opposite—offloading cognitive burden onto people whose attention is the scarcest resource of all.