Worse-than-average effect

The worse-than-average effect is the tendency to place yourself below comparable peers when a task feels difficult or when you expect low absolute performance. Researchers also call it the below-average effect or, in work on overconfidence, underplacement.

The comparison is essential. “I may score poorly” is an absolute prediction. “I will score worse than the average person in this group” is a relative one. The second judgment can be distorted when your own struggle is vivid and everyone else's struggle is mostly unknown.

A fictional sound-map exercise

The museum, volunteer, sound-map task, group size, predictions, and outcome in this scene are invented. Nothing here is study evidence or a description of an actual person.

During an imagined after-hours training session, ten first-time museum guides hear recordings of gears, pipes, and lifts, then mark their likely locations on a building plan. Lena cannot place the opening sound. She immediately predicts that her result will sit near the bottom.

Lena knows how difficult the clip felt to her. She does not yet know how the other nine volunteers interpreted it, and they are also using the system for the first time. The difficult experience may justify a low prediction for her own score. It does not, by itself, tell her where that score will fall among the group.

Before turning discomfort into rank, Lena writes two separate questions: “How many will I identify?” and “What position will that score occupy among these ten volunteers?” She waits for the fictional practice distribution before answering the second.

What the original research found

Justin Kruger's three-study program challenged the idea that people invariably see themselves as above average. In the first two studies, domains associated with high absolute skill produced better-than-average judgments, while domains associated with low absolute skill produced reliable below-average judgments. Under cognitive load in Study 3, both directional biases grew stronger. Kruger interpreted the pattern as excessive weight on one's own ability and insufficient adjustment for the comparison group's ability. (Kruger, 1999)

That evidence supports a narrow version of the catalog claim: people can give lower comparative self-ratings on difficult tasks. It does not establish that every unfamiliar or complex task causes underplacement, or that each person who reports it is objectively equal to the peers they name.

A four-experiment program by Don Moore and Tai Gyu Kim made the shared-difficulty point especially clear. Participants wagered on beating an opponent in either a simple or difficult trivia quiz. Difficulty did not change the average participant's probability of winning, yet people wagered more on the simple quiz in the first three experiments. In the fourth, shifting attention toward the opponent reversed the effect. (Moore and Kim, 2003)

The important distinction is mathematical as well as psychological. If two randomly paired people face the same task, making the task harder can lower both absolute scores without automatically changing the average person's chance of winning the pair. Shared difficulty changes expected performance; it does not make everyone lower-ranked.

Why difficulty can reverse the comparison

Comparative judgment requires information about two sides: the focal person and the reference group. We usually know more about our own confusion, preparation, mistakes, and effort than we know about an average peer. When a task feels easy, that self-focused evidence can support a better-than-average judgment. When it feels hard, the same weighting can pull the judgment below average.

Paul Windschitl, Kruger, and Ericka Nus Simms tested this issue across six experiments involving hypothetical and real competitions. Shared advantages and disadvantages influenced judgments through both egocentrism, extra weight on self-relevant information, and focalism, extra weight on whichever competitor the question made central. (Windschitl, Kruger, and Simms, 2003)

This is why the worse-than-average effect is better understood as one possible direction of comparative misjudgment, not a permanent pessimistic trait. Change the task, the available information, the focal target, or the reference group, and the direction can change.

Motivation is possible, but it is not the settled cause

One proposed account says difficult tasks shift self-evaluation from self-enhancement toward self-improvement motives. Modest estimates might protect against failure or encourage preparation. The checked evidence does not establish that motivational switch as necessary.

John Chambers and Windschitl reviewed research on comparative judgments and showed how information-processing limits and features of the judgment itself can produce bias without a self-enhancement motive. They treated the relationship between motivated and nonmotivated causes as unresolved. (Chambers and Windschitl, 2004)

So “I ranked myself low because I wanted to improve” may sometimes be true, but the rating alone cannot reveal that motive. Difficulty, missing information about peers, focal attention, and noisy judgment offer competing explanations.

Skill level does not create a simple mirror image

The worse-than-average effect is sometimes mistaken for the opposite of the Dunning-Kruger effect. That shortcut loses an important result.

Katherine Burson, Richard Larrick, and Joshua Klayman studied 12 tasks across three published studies. On moderately difficult tasks, the best and worst performers differed little in comparative-judgment accuracy. On harder tasks, the best performers were less accurate than the worst performers in judging their relative standing. Difficulty could create negative comparative bias across the skill distribution, including among high performers. (Burson, Larrick, and Klayman, 2006)

Actual skill still matters. A participant near the bottom of a valid performance distribution can accurately judge themselves below average. A high performer can underplace themselves. Without measured performance and a defined comparison group, the direction of a self-rating does not tell us whether it is calibrated.

Underplacement is not the same as low confidence

Moore and Paul Healy separate three judgments that are often bundled together:

• Absolute estimation: How well will I perform?
• Placement: How will I perform relative to others?
• Precision: How certain am I that my estimate is right?

Their analysis shows why hard tasks can produce overestimation of an absolute score alongside underplacement relative to other people. Easy tasks can produce the reverse combination. (Moore and Healy, 2008)

Someone can therefore feel confident about a predicted score and still think peers will score higher. Another person can be unsure of an absolute estimate while placing themselves above average. Calling both “confidence” hides the decision you need to check.

The reference group can change the answer

“Average compared with whom?” is not a minor wording question. A first-time runner might rank differently against other beginners, everyone in a local club, elite competitors, or adults in general. Each group defines a different distribution.

Ethan Zell and Mark Alicke's review of the local-dominance effect found that comparisons with a few concrete people can influence self-assessment more than comparisons with larger groups. The review also identifies physical proximity and self-enhancement as moderators and considers alternative explanations. (Zell and Alicke, 2010)

A valid comparative estimate needs a named group, a comparable task, a consistent measure, and enough information about the distribution. Without those pieces, “below average” may describe a feeling more than a rank.

What the effect does not prove

The label does not show that a person has low self-esteem, impostor syndrome, or a stable tendency to underrate everything. It also does not prove that the person is wrong. Some people are below a specified average on a measured task.

Nor is it the same as:

• The better-than-average effect: placement above peers, often observed on easier or high-success tasks.
• The Dunning-Kruger effect: a proposed relationship between skill and self-assessment error, not every low comparative estimate.
• Underconfidence: confidence lower than accuracy warrants. Underplacement concerns relative rank.
• Negativity bias: extra weight on negative information, which does not require a peer comparison.
• Outcome bias: judging a decision from its result. A worse-than-average judgment can occur before the task begins.