# The attraction effect in perceptual decision-making: a case of dominance asymmetry

## TL;DR

The study shows that the attraction effect in decision-making is reliable when stimuli create clear dominance asymmetry, resolving past inconsistencies in perceptual choices.

## Contribution

Demonstrates that the attraction effect is robust in perceptual decision-making when decoys induce strong item-level dominance asymmetry.

## Key findings

- Star-shaped stimuli reliably produced strong target-decoy dominance over competitor-decoy dominance.
- Perceptual stimuli in triangular layouts showed a positive attraction effect when dominance asymmetry was present.
- Inconsistencies in previous studies were linked to the absence of asymmetric dominance in stimulus design.

## Abstract

The attraction effect (AE), or asymmetric dominance effect, occurs when the presence of a clearly inferior decoy increases the choice of a target over a competitor. While robust in value-based domains, findings with perceptual stimuli have been inconsistent, with some studies even reporting reversals in triangular arrangements of stimuli.

Across four experiments and a reanalysis of prior data, we investigated whether these inconsistencies are attributable to the presence or absence of genuine item-wise dominance asymmetry. We utilized novel star-shaped stimuli and traditional rectangle stimuli to test the dominance asymmetry.

Experiment 1 established that the star-shaped stimuli reliably produced strong target-decoy (TD) dominance over competitor-decoy (CD) dominance, whereas traditional rectangles showed weaker but still positive asymmetry. Experiment 2 provided the first robust demonstration of a positive AE with perceptual stimuli in a triangular layout using stars, while Experiment 3 showed that rectangle stimuli, when presented in a triangular layout, produced an aggregate null rather than a negative effect. Similarly, the reanalysis of data from a previous triplet experiment involving bars stimuli pointed toward a null effect. Experiment 4 again linked inconsistent findings from linear vs. triangular alignment of triplet rectangles to the presence of asymmetric dominance, while also demonstrating an interaction between the differential ease of comparison in pairs and presentation format.

Together, these results demonstrate that AE is a robust phenomenon that emerges whenever decoys create strong item-level dominance asymmetry. Apparent inconsistencies with perceptual stimuli reflect stimulus-specific dominance structures. This work clarifies the boundary conditions of the AE, reinforces its domain generality, and provides methodological guidance for future research on context effects in perceptual decision-making.