# Exposure to subtle dominance cues activates the stress response and affects decision-making

## TL;DR

Exposure to subtle dominance cues can trigger stress and influence decisions, even when the cues are not overtly aggressive.

## Contribution

The study shows how low-intensity emotions from social cues affect decision-making, relevant for human-like AI design.

## Key findings

- Most participants found subtle dominance cues offensive, regardless of text or image presentation.
- Combining condescending text with a dominant face influenced medication-taking decisions.
- Arousal levels increased with dominance cues but decreased when making decisions after exposure.

## Abstract

Dominance cues may offend and elicit anger. Based on theories of affect-as-information, we tested whether subtle cues in words or images indicative of dominance could activate the stress response and impact decision-making.

Participants asked to imagine being patients were exposed to subtle dominance cues of a doctor. By measuring the skin conductance levels and through self-reported assessments, we examined whether participants would be offended when exposed to dominance cues in text alone or when combined with facial images. Participants assessed the probability of a medication’s side effects and chose to take the medication prescribed after reading a doctor’s advice that was worded to sound either condescending or neutral. The doctor’s statements were shown alone or matched with a photo of either a dominant-looking or a trustworthy-looking face.

Most found dominance cues presented in text, with or without a facial image, offensive. No differences were found in probability assessment but the choice to take the medication was affected when the condescendingly worded text was combined with a photo of a dominant face. Arousal levels increased while viewing the dominance cues, but the levels decreased when decisions were made following exposure to a condescendingly worded text and its matching facial expression of dominance.

The study contributes to understanding the impact of lower-intensity emotions sensed during social interactions on decision-making, which could be important for designing computer programs that mimic human social interactions.