Stress and poor sleep were the strongest independent predictors of emotional eating in a nationwide study of 1,297 Greek university students, outranking sex, smoking, physical activity, and weight status. The findings treat emotional eating as a measurable health pattern with identifiable drivers, and the drivers are not what diet campaigns usually target.
The behavior in question. Emotional eating means reaching for food when emotions churn rather than when the body actually needs fuel. The food chosen is rarely a carrot. Sweet, fatty, calorie-dense options dominate, because they deliver fast comfort: they activate reward pathways and, through cortisol-driven mechanisms tied to the stress response, briefly quiet negative emotion. The relief is short. Over time the pattern is linked to weight gain, abdominal adiposity, poorer diet quality, metabolic disturbances, and a higher risk of disordered eating, along with deepening anxiety and depressive symptoms.
Why students are a pressure cooker. Higher education is a developmental bottleneck. Young adults leave structured home routines and gain near-total autonomy over what and when they eat, often with no budget, no time, and no kitchen skills. Academic workload, worries about employment, financial limits, and the effort of social integration stack on top, producing a population with chronically elevated stress and chronically poor sleep.
The study. The team enrolled 1,297 students between 18 and 24 (average age 20.5) across ten regions of Greece, from Athens and Thessaloniki to the Aegean islands, between March 2021 and October 2024. The response rate was 80.5%. Recruitment was voluntary, conducted through in-person contact and department announcements, with no incentives offered; students with chronic medical or psychiatric conditions were excluded. Each of the three emotional eating tiers ended up holding roughly 432 students.
How the measurements worked. Stress was measured with the widely used 10-item Perceived Stress Scale (PSS), which scores the preceding month on a 0 to 40 scale, with cutoffs of 0 to 13 for low, 14 to 26 for moderate, and 27 to 40 for high stress. Sleep was scored on the PSQI, the Pittsburgh Sleep Quality Index, a 19-item questionnaire whose global score runs 0 to 21; below 5 counts as adequate, at or above 5 as inadequate. Emotional eating was captured through the three emotion-driven items of the Revised Three-Factor Eating Questionnaire, converted to a 0 to 100 scale and divided into tertiles labeled low, moderate, and high. Both instruments performed well in this sample, with Cronbach’s alpha of 0.85 for the PSS and 0.84 for the PSQI. Because the outcome was ordered rather than binary, the analysis used ordinal logistic regression, and a test of the proportional odds assumption confirmed the model was valid (p = 0.46).
The findings. Stress showed a clear dose-response relationship with emotional eating. Highly stressed students had nearly triple the odds of sitting in a higher emotional eating tier than their low-stress peers (odds ratio 2.86, 95% confidence interval 1.68 to 4.85, p = 0.0001), and moderately stressed students had roughly double the odds (odds ratio 2.01, confidence interval 1.41 to 2.87, p = 0.0001). Poor sleep proved just as potent: students sleeping badly carried triple the odds of higher emotional eating relative to adequate sleepers (odds ratio 3.08, confidence interval 1.75 to 5.42, p = 0.0001). Both links survived simultaneous adjustment for a broad battery of sociodemographic, academic, lifestyle, and body-composition variables.
The other significant predictors were familiar but weaker: female sex (OR = 1.43), rural residence (OR = 1.53), smoking (OR = 1.68), low versus high physical activity (OR = 2.12), overweight (OR = 1.95), and obesity (OR = 2.10). Stress and sleep outperformed nearly all of them. In the study’s framing, psychological and behavioral regulation processes may sit closer to the center of emotional eating than demographic or body-composition factors.
The results survived robustness checks. The team reran the analysis with continuous scores. Emotional eating correlated with perceived stress (Spearman’s rho = 0.46) and with sleep quality (rho = 0.41), and a linear regression model accounted for roughly 38% of the variance in emotional eating, with both stress and sleep remaining independent predictors.
Two levers, one outcome. Stress and sleep are deeply entangled, and the two scores correlated with each other (rho = 0.43). Yet both retained independent significance in the same model, which implies partially distinct pathways. Stress is thought to operate through the HPA axis, cortisol, and emotion-focused coping; poor sleep may act through appetite-regulating hormones such as leptin and ghrelin, impaired executive function, weakened inhibitory control, and heightened reactivity to food cues. Students carrying both burdens appear especially vulnerable, which is why the study concludes that interventions pairing stress management with sleep hygiene are needed rather than dietary education alone.
Caveats. The study is cross-sectional, so it cannot establish cause or temporal order; emotional eating may deepen stress and disturb sleep as much as the reverse. All measures were self-reported and vulnerable to recall and social desirability bias. Recruitment was non-probability based, so self-selection cannot be excluded, and the sample cannot be assumed representative of all Greek university students, let alone students elsewhere. Unmeasured confounders such as chronotype, social support, personality, and dietary intake could not be fully ruled out. The study also acknowledges that the non-Greek subgroup was small and heterogeneous, making the nationality finding provisional.
The takeaway. Emotional eating is not a discipline failure; it is a stress and sleep problem wearing a food costume. The strongest levers this study identifies, stress reduction and sleep improvement, are also the ones most campus health programs underinvest in. Longitudinal and intervention studies are the next step, but the signal from 1,297 students is clear: fix the pressure, and the midnight kitchen trip may lose its grip.
Source
Alexatou O, Papadopoulou SK, Angelakou EP, Migdanis A, Mentzelou M, Migdanis I, Louka A, Serdari A, Giaginis C. Associations of Emotional Eating with Perceived Stress and Sleep Quality Among University Students: A Nationwide Cross-Sectional Study. Nutrients. 2026;18(15):2434. doi:10.3390/nu18152434. PMID 42588057.