Psychological distress, sleep, coping, and perceived cognitive difficulties among architecture students in Jordan: an observational network analysis

Architecture design education involves open-ended problem solving, iterative critique, sustained project work and compressed deadlines. Psychological distress, coping, sleep, recovery, burnout and perceived cognitive difficulty may be interrelated, but they are often examined separately. This study used a network approach to describe their conditional organisation in architecture students. Linked questionnaire records from 139 architecture undergraduates at one Jordanian university were analysed. Scores were audited and reconstructed using explicit missing-item rules. Nine nodes comprised psychological distress, burnout, recovery, poor sleep quality, insomnia severity, perceived cognitive difficulty and three study-defined coping aggregates. A regularised Gaussian graphical model followed bounded iterative imputation using age, sex and academic year as auxiliaries. Three-level burnout and recovery indicators were treated approximately as continuous. Assessments covered bootstrap uncertainty, case-dropping and community stability, exploratory cognitive psychometrics including parallel analysis, and sensitivities to missing-data handling, rank transformation, ordered-node exclusion, penalty variation, demographic adjustment, sleep scoring and timing quality. The network contained 25 of 36 possible edges. The strongest linked problem-focused with emotion-focused coping (partial r = 0.506, 95% bootstrap interval 0.380–0.599) and poor sleep quality with insomnia severity ( r = 0.373, interval 0.155–0.481). Three author-labelled communities emerged in a secondary descriptive analysis, with modest modularity (Q = 0.230) and limited sampling stability (median adjusted Rand index = 0.556; exact primary partition = 12.6%). Emotion-focused coping and psychological distress had high observed connectivity, but their bootstrap centrality intervals overlapped. Case-dropping stability was 0.30 for strength and expected influence. Complete cases ( n = 112), posterior imputations and demographic adjustment reproduced broad edge patterns; timing-quality checks changed community assignments. Psychological distress, coping, sleep, recovery, burnout, and perceived cognitive difficulty formed an interconnected between-person network. The findings are best interpreted as a structural map of cross-domain associations that can guide what future longitudinal studies measure together, rather than as evidence of causal pathways, within-person processes, or intervention priorities.