General Regression Modeling of the Impact of Physical Activity on Stress-Related States in Higher Education Students During Military Conflict

Abstract

Understanding the factors influencing stress-related states in higher education students during military conflict is crucial for developing interventions to mitigate stress in young people. Purpose: This study aimed to identify significant factors related to the impact of physical activity on stress-related states in Ukrainian university students and to analyze the relationships between these factors. Materials and methods: Sociological and psychodiagnostic approaches, along with statistical analysis, were applied. The study included 1,901 students from Ukraine. Results: Using general regression modeling (GRM), the study identified relationships between stress-related states, psychophysiological indicators, and various factors such as gender, lifestyle, physical activity (PA), bad habits, and the use of sedatives. Statistically significant (p<0.05) models were built that can predict the number of stress-associated conditions based on mood and well-being scores (Model 1) and on the presence of bad habits, use of sedatives, PA, and active lifestyle depending on gender (Model 4); the total score of students’ psychophysiological indicators based on the presence of stress-related states (Model 2) and on the basis of data on gender, bad habits, use of sedatives, PA, and active lifestyle (Model 3). It was proved that an active lifestyle and systematic PA partially affect the psychophysiological indicators of students. According to Model 3, a sedentary lifestyle leads to a decrease in the total score of the indicators by 5.8% from 21.16 to 19.94 points. Moreover, a sedentary lifestyle combined with the absence of regular PA causes a decrease in this parameter by 8.5% to 19.36 points. And, according to Model 4, the predicted number of stress-related states increases by 7.8% due to a sedentary lifestyle. Furthermore, students with bad habits can be predicted to have the highest number of stress-related states, especially in the absence of regular physical activity. Conclusions. The developed GRM-models can predict: the number of stress-related states based on the assessment of well-being and mood; the total score of psychophysiological indicators based on the presence of stress-related states and taking into account the gender of students; the probability of stress-related states in this population group. Furthermore, we obtained new information about the factors that affect the stress-related states in higher education students under military conflict conditions.