The COVID-19 global health emergency has greatly impacted the educational field. Faced with unprecedented stress situations, professors, students, and families have employed various coping and resilience strategies throughout the confinement period. High and persistent stress levels are associated with other pathologies; hence, their detection and prevention are needed. Consequently, this study aimed to design a predictive model of stress in the educational field based on artificial intelligence that included certain sociodemographic variables, coping strategies, and resilience capacity, and to study the (...) relationship between them. The non-probabilistic snowball sampling method was used, involving 337 people from the university education community in south-eastern Spain. The Perceived Stress Scale, Stress Management Questionnaire, and Brief Resilience Scale were administered. The Statistical Package for the Social Sciences was used to design the architecture of artificial neural networks. The results found that stress levels could be predicted by the synaptic weights of coping strategies and timing of the epidemic, with a predictive capacity of over 80% found in the neural network model. Additionally, direct and significant associations were identified between the use of certain coping strategies, stress levels, and resilience. The conclusions of this research are essential for effective stress detection, and therefore, early intervention in the field of educational psychology, by discussing the influence of resilience or lack thereof on the prediction of stress levels. Identifying the variables that maintain a greater predictive power in stress levels is an effective strategy to design more adjusted prevention programs and to anticipate the needs of the community. (shrink)

Little has been studied on the relationship between affect and school problems related with attendance. This study aims to identify different affective profiles and to determine whether these profiles differ from each other based on the four functional conditions of school refusal behavior. Participants comprised 1,816 Spanish adolescents aged 15–18 years. The Positive and Negative Affect Schedule for Children-Short Form and the School Refusal Assessment Scale-Revised for Children were administered. Latent profile analysis revealed five affective profiles: low affective profile, self-fulfilling (...) profile, low positive affect profile, self-destructive profile, and high affective profile. The self-destructive profile revealed the highest average scores in the first three factors of the SRAS-R-C, whereas the high affective profile reached the highest average score in the fourth factor. On the contrary, the self-fulfilling profile obtained the lowest average scores in the first two factors of the SRAS-R-C, whereas the low affective profile revealed the lowest average scores in the last two factors. Findings suggest the relevance of developing more adaptative affective profiles, such as the self-fulfilling profile, which would contribute to diminishing school attendance problems. (shrink)

Artificial intelligence is a useful predictive tool for a wide variety of fields of knowledge. Despite this, the educational field is still an environment that lacks a variety of studies that use this type of predictive tools. In parallel, it is postulated that the levels of self-esteem in the university environment may be related to the strategies implemented to solve problems. For these reasons, the aim of this study was to analyze the levels of self-esteem presented by teaching staff and (...) students at university and to design an algorithm capable of predicting these levels on the basis of their coping strategies, resilience, and sociodemographic variables. For this purpose, the Rosenberg Self-Esteem Scale, the Perceived Stress Scale, and the Brief Resilience Scale were administered. The results showed a relevant role of resilience and stress perceived in predicting participants’ self-esteem levels. The findings highlight the usefulness of artificial neural networks for predicting psychological variables in education. (shrink)