Why an option looks better alone than beside a rival

A dictionary is offered in two versions. One has ten thousand entries and a torn cover; the other has twenty thousand entries and a perfect cover. Shown one at a time, most people choose the smaller, damaged one, because a torn cover is easy to judge and the entry count is not. Shown side by side, the same people switch to the larger one. The evaluability hypothesis is the reason: an attribute with no obvious reference point is nearly invisible until a rival supplies one.

What everyone sees

What everyone sees is a contradiction, or a fickle shopper. The same person who chose the damaged dictionary alone reverses the moment the two are compared, and the reversal is read as indecision or as a trick of the salesperson. Each choice looks reasonable in its own setting, so nobody suspects that the setting did the work. The preference is filed as unstable, and instability is treated as a personal flaw rather than a property of how the options were presented.

What is actually happening

Christopher Hsee set out the evaluability hypothesis in “The Evaluability Hypothesis: An Explanation for Preference Reversals between Joint and Separate Evaluations of Alternatives” (Organizational Behavior and Human Decision Processes, 1996), using the dictionary case and others to show that people lean on attributes with a natural reference point and neglect attributes without one. In a separate-evaluation condition the hard-to-evaluate attribute carries almost no weight; in a joint evaluation it suddenly does, and the preference reverses. Hsee, Loewenstein, Blount, and Bazerman reviewed the wider pattern in “Preference Reversals Between Joint and Separate Evaluations of Options” (Psychological Bulletin, 1999). Hsee’s related paper “Less Is Better” (Journal of Behavioral Decision Making, 1998) showed the extreme case, where a low-value option is chosen over a high-value one because only the low option’s strength is easy to evaluate.

Why it stays hidden

It stays hidden because each evaluation feels complete. A person judging one option alone believes they have weighed everything, when in fact the unmeasurable attribute was never given a number. The bias also hides because joint evaluation is the exception: shops, forms, and offers usually present one option at a time, so the reversal is rarely triggered and rarely seen. And when it does appear, it is attributed to the chooser rather than to the frame, which is why the same person keeps making the same reversal without ever noticing the pattern.

How the evaluability hypothesis changes a choice

The rule is simple: attributes with an obvious scale, such as a torn cover or a known price, dominate when an option is judged alone. Attributes without a scale, such as the number of dictionary entries or the durability of a material, are ignored because there is nothing to compare them to.

Put two options together and the missing scale appears. The larger entry count now has a reference point, so it becomes legible and can outweigh the cover. The preference does not change because the person changed; it changes because the attribute became visible.

The evidence in numbers and field studies

Hsee demonstrated the reversal across several goods, including dictionaries, ice cream, and used cars, and found it robust enough to be treated as a general feature of judgement rather than a quirk of one product.

The 1999 Psychological Bulletin review by Hsee and colleagues placed the effect within a broader family of joint-versus-separate reversals, showing that many preference shifts share the same cause: an attribute that is hard to evaluate alone becomes decisive when a comparison is supplied.

When the effect is weakest

The reversal shrinks when the hard attribute is made easy to evaluate on its own. Give the dictionary a benchmark, such as the entry count of a well-known rival, and the separate judgement begins to match the joint one.

It also weakens for experts, who carry their own internal scales and can evaluate attributes without a side-by-side comparison. That is why the effect is strongest for unfamiliar goods and first-time buyers.

Questions readers ask

What is the evaluability hypothesis?

It is the idea that people rely on attributes with a natural reference point and neglect attributes without one. Judging an option alone, the hard-to-evaluate attribute carries little weight; comparing two options makes it legible and can reverse the preference.

Why do I choose differently when options are side by side?

Because comparison supplies a scale for attributes that had none. The option itself has not changed; the previously invisible attribute has simply become measurable, so it can now outweigh the easy-to-judge ones.

When is the evaluability effect strongest?

For unfamiliar goods and first-time buyers, who have no internal scale to fall back on. Experts, who carry their own benchmarks, show the reversal far less often.

How can I avoid being swayed by it?

Give the hard-to-evaluate attribute a reference point before deciding, such as a benchmark from a known rival. Once the attribute has a scale, the separate and joint judgements tend to agree.