Can two short nights make you look less attractive?

A tired face can speak before you do

First impressions do not wait for a full biography. A face can carry traces of what happened over the preceding nights—long before personality, ability or intent has entered the conversation.

The narrow research question is not whether sleep determines beauty. It is whether a short period of restricted sleep changes how strangers read the same person when almost everything else in the image is held constant.

One controlled experiment says yes, but only slightly.

The experiment changed sleep, not the person

Researchers at Karolinska Institutet and Stockholm University photographed 25 healthy adults twice. Before one session, participants had eight hours in bed for two consecutive nights. Before the other, they had four hours in bed for two nights. The order was counterbalanced and the sessions were at least one week apart.

Activity monitors estimated an average of 7 hours 35 minutes of sleep per night in the normal-sleep condition and 4 hours 15 minutes in the restricted condition—about seven fewer hours across the two nights.

Photography was deliberately standardized: same time of day, camera setup, clothing, hairstyle rules, minimal jewellery, no makeup and a relaxed expression. A person who did not know the sleep condition selected one image from each session.

Then 122 adults recruited from the Stockholm public rated the 50 photographs on seven-point scales. They reported how willing they would be to socialize with the person and how attractive, healthy, sleepy and trustworthy the person appeared.

Strangers noticed small differences

Compared with the well-rested photographs, the sleep-restricted photographs received the following adjusted mean changes:

The first four differences were statistically significant in the authors’ mixed-effects models. Trustworthiness was not.

Every rating ran from one to seven. The authors described the condition effects as small. Two short nights did not transform a person’s appearance; strangers detected a subtle average difference in standardized images of the same faces.

Sleepiness carried the strongest signal

Sleepiness showed the largest direct condition difference. Faces rated as sleepier also tended to receive lower ratings for health, attractiveness and willingness to socialize.

The authors’ exploratory mediation model estimated that the four measured impressions together accounted for 34.7% of the sleep-condition effect on willingness to socialize. That analysis is consistent with sleepiness and related appearance judgments carrying part of the response. It does not prove why any one rater made a judgment, and it does not convert a photograph rating into real social behavior.

A photograph cannot predict a real encounter

Only 25 people were photographed. Most were healthy students in their early to mid-20s, and the authors described the photographed group as mainly Caucasian. The raters came from the same metropolitan setting.

Each outcome was a six-second rating of a neutral photograph of a stranger—not an actual conversation, date, interview or access decision. Real interactions add voice, movement, familiarity, expression, makeup, caffeine, stress and many other cues that the experiment held constant or did not test.

The sleepiness scale also changed direction after early raters reported confusion, so those first 40 raters were removed from that outcome. Additional raters with very little variation in their answers were excluded separately for each measure.

Treat sleep as preparation, not judgment

The experiment supports a specific causal conclusion: two nights of restricted sleep produced small changes in how this sample of faces was rated under standardized conditions. It does not show that sleep determines beauty, trustworthiness, social access or human value.

Sleep is therefore best understood here as a temporary, potentially modifiable condition before a first meeting. The unresolved question is whether the social response depends only on subtle appearance—or whether an explicit, voluntary signal of openness can change what strangers actually do.

Mean rating-point change · zero at center

Sleepiness

+0.25

Willingness to socialize

−0.15

Perceived health

−0.11

Attractiveness

−0.09

TrustworthinessNot statistically significant

−0.04

Sundelin et al. (2017), mixed-effects estimates. Each outcome used a seven-point scale.