Evaluability Hypothesis (Joint vs. Single Evaluation)
The evaluability hypothesis explains preference reversals between assessing options separately and comparing them together. An attribute that is hard to assess alone, such as the number of entries in a dictionary, can carry more weight when another option provides a comparison.
THE IDEA
Consider the question, "How tall is John?" If John is 5' tall, your answer depends on his age; he is very tall if 6 years old, very short if 16. Your System 1 automatically retrieves the relevant norm, adjusting the scale of tallness. You can also match intensities across categories, e.g., "How expensive is a restaurant meal that matches John's height?"
Now look at this: John is 6, 5' tall. Jim is 16, 5'1" tall. In single evaluations, everyone agrees John is very tall and Jim is not, because they are compared to different norms. If asked directly, "Is John as tall as Jim?" you answer no. No surprise. But in other situations, this process leads to incoherent choices.
Our world is broken into categories with norms (six-year-old boys, tables). Judgments are coherent within categories but potentially incoherent across categories. Example: Which do you like more, apples or peaches? Steak or stew? Apples or steak? The first two compare items within a familiar category; the cross-category comparison may be harder to answer consistently.
Imagine an email requesting a contribution to save dolphins from pollution. You automatically compare dolphins to other endangered species; dolphins rank high (more charming than ferrets or carp). You translate your liking into a dollar amount via intensity matching, using your personal scale of contributions. Later, a different appeal for farmworkers with skin cancer evokes a public-health category; the farmworkers' issue may rank lower among health issues, so you might contribute less than to dolphins. But in joint evaluation, the human vs. animal feature becomes salient, and people contribute more to farmworkers.
Christopher Hsee's example: secondhand music dictionaries – Dictionary A (1993, 10,000 entries, like new) vs. Dictionary B (1993, 20,000 entries, cover torn). In single evaluation, A is valued higher; in joint evaluation, B is preferred because the number of entries (20,000 vs. 10,000) is easily evaluable when compared, while condition is less important. This illustrates the evaluability hypothesis: attributes that are not evaluable in isolation gain weight in joint evaluation.
When to use it
Comparing product features: evaluating a single product online vs. comparing multiple products side-by-side in a store.
Charitable donations: choosing between causes presented individually in separate emails vs. together on a list.
Job candidate evaluation: reviewing resumes one by one vs. interviewing candidates in a group or side-by-side comparison.
Investment decisions: assessing a single stock's prospects vs. comparing several stocks in a portfolio context.
FROM THEORY TO EVERYDAY LIFE
In practice
John's Height Example
John (6 years old, 5' tall) and Jim (16 years old, 5'1" tall). In single evaluation, John is considered very tall and Jim is not, because each is compared to age-appropriate norms. In joint evaluation, it's clear Jim is taller.
Dolphins vs. Farmworkers
Single evaluation: People donate more to dolphins (charming species) than to farmworkers (skin cancer from sun exposure) because dolphins rank high among endangered species while farmworkers rank lower among health issues. Joint evaluation: People donate more to farmworkers because the 'human vs. animal' feature becomes salient.
Music Dictionaries
Dictionary A (1993, 10,000 entries, like new) vs. Dictionary B (1993, 20,000 entries, cover torn). Single evaluation: A is valued higher. Joint evaluation: B is preferred because the number of entries is easily comparable and more important than the condition.
A MOMENT TO REFLECT
Put it to work.
When making a decision, are you evaluating options alone or together? How might changing the evaluation context alter your preference?
Think of a time you changed your mind between single and joint evaluation of options. What attribute became salient that you had overlooked?
How can you guard against preference reversals when comparing options from different categories (e.g., job offers, purchases, donations)?
References
- Evaluability hypothesis — Christopher Hsee
CONTINUE THE CONVERSATION
Source texts
- Thinking, Fast and Slow (Daniel Kahneman)