War Leaves Most Ethiopian University Students in Psychological Distress, Bayesian Study Finds

More than eight in ten university students surveyed at Debre Markos University in Ethiopia reported psychological distress linked to war, according to a cross-sectional study published in BMC Psychology. The finding, drawn from 714 students selected by simple random sampling, paints a stark picture of the mental health burden carried by young people pursuing higher education in a region scarred by armed conflict. Researchers Yenesew Fentahun Gebrie and Kebadu Tadesse Cherie, both statisticians at Debre Markos University, set out to quantify how widespread distress is among students from conflict-affected areas and to identify which experiences most strongly predict it.

The scale of the result is difficult to overstate. A prevalence of 81.4 percent for self-reported psychological distress means that roughly five out of every six students in the sample were struggling with war-related emotional strain at the time of the survey. For comparison, the researchers had based their sample size calculation on an assumed prevalence of about 45 percent, a figure drawn from earlier studies. The observed rate turned out to be far higher than that planning assumption, suggesting that the psychological toll of conflict on this population had been substantially underestimated.

Armed conflict can push students toward psychological distress through several intertwined pathways. The study’s authors point to forced displacement, disruption of education, direct exposure to violence, worry about personal and family safety, bereavement, economic instability, and the erosion of social support networks. University students occupy a particularly vulnerable position in this landscape: they are old enough to be directly involved in or targeted by conflict, yet their studies depend on a stability that war destroys. Many also carry responsibility for supporting families whose livelihoods have been shattered.

Methodologically, the study stands out for its use of Bayesian logistic regression, a statistical approach that differs from the classical frequentist methods more common in public health research. Instead of producing a single point estimate and a p-value, Bayesian analysis yields a posterior probability distribution for each effect, summarized here as odds ratios with 95 percent credible intervals. A credible interval has a more direct interpretation than a confidence interval: given the data and the model, there is a 95 percent probability that the true effect lies within that range. This framing is especially useful in settings like conflict research, where prior knowledge from earlier studies can be formally incorporated and where uncertainty is often large.

Before fitting the Bayesian model, the researchers used descriptive statistics and chi-square tests to screen for associations between distress and a set of candidate factors: gender, region, war zone exposure, family death, physical injury, direct participation in the war, and family support roles. Several of these survived into the multivariable analysis with strong effects. Students who had been physically injured in the conflict faced the steepest odds, at 5.13 times those of uninjured peers (95 percent credible interval: 2.26 to 11.29). Physical injury, the data suggest, is one of the most powerful markers of lasting psychological harm in this population.

Indirect participation in the war carried an odds ratio of 4.43 (credible interval: 2.70 to 7.18), while direct participation was associated with an odds ratio of 2.03 (credible interval: 1.06 to 4.17). The finding that indirect involvement predicted distress more strongly than direct involvement may seem counterintuitive, but the authors do not overinterpret the mechanism; both forms of exposure clearly elevated risk. Losing a family member to the war nearly tripled the odds of distress (odds ratio 3.03, credible interval: 1.59 to 5.56), and having spent time in a war zone more than doubled them (odds ratio 2.66, credible interval: 1.64 to 4.41). Loss of family property raised the odds by a smaller but still meaningful margin (odds ratio 1.68, credible interval: 1.02 to 2.81).

Two further factors emerged as significant. Female students had roughly twice the odds of distress compared with male students (odds ratio 2.06, credible interval: 1.23 to 3.56), a pattern consistent with broader literature on gender differences in the psychological consequences of conflict and displacement. Students’ work status was also associated with distress, with an odds ratio of 2.73 (credible interval: 1.23 to 7.18), although the wide credible interval here reflects considerable uncertainty, likely because relatively few students fell into the relevant category. Taken together, the model indicates that distress in this population is not randomly distributed but concentrated among students with specific, identifiable war exposures.

The practical implications are direct. The authors recommend that universities in conflict-affected settings strengthen mental health screening so that students at highest risk, including those who were injured, bereaved, displaced, or exposed to war zones, can be identified early. They also call for accessible psychosocial support services and functioning referral systems that can connect students to appropriate care. With more than four-fifths of students reporting distress, the need is not marginal; it touches nearly the entire student body, and campus counseling capacity in such settings is typically far below what the numbers demand.

The study does have limitations that shape how its findings should be read. The authors themselves note that further research using validated mental health instruments is needed to assess the severity of psychological distress, since the current survey relied on self-reported measures rather than clinical diagnostic tools. The cross-sectional design captures a single moment in time and cannot establish that war exposures caused the distress, even though the temporal logic of the exposures makes the association highly plausible. The sample comes from a single university, so generalization to other conflict-affected regions requires caution, although the study’s focus on students from conflict-affected areas gives it particular relevance for northern Ethiopia.

Even with those caveats, the study contributes something the field has lacked: a quantified, statistically rigorous estimate of psychological distress among university students in an Ethiopian conflict zone, along with a ranked set of risk factors that intervention programs can act on. The Bayesian framework adds transparency about uncertainty, giving decision-makers not just best guesses but plausible ranges for every effect. As conflicts continue to disrupt higher education in various parts of the world, the message from Debre Markos is clear: the mental health of students in war-affected regions is a large-scale, measurable problem, and the first step toward addressing it is systematic screening backed by services that can respond to what screening finds.

Subject of Research: Prevalence and risk factors of war-related psychological distress among university students in conflict-affected Ethiopia

Article Title: Psychological distress due to war and associated factors among university students in a conflict-affected region of Ethiopia: a Bayesian logistic analysis

Article References: Gebrie, Y. F., & Cherie, K. T. (2026). Psychological distress due to war and associated factors among university students in a conflict-affected region of Ethiopia: a Bayesian logistic analysis. BMC Psychology. https://doi.org/10.1186/s40359-026-05777-8

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05777-8

Keywords: psychological distress, war exposure, Ethiopia, university students, Bayesian logistic regression, mental health, armed conflict, cross-sectional study, physical injury, bereavement, Debre Markos University, odds ratio