Academic resilience is evident in students who are living in vulnerable environments, yet achieve success in academic outcomes. As a result, substantial attention has been devoted to identifying the factors associated with academic resilience and supporting students to be resilient. This study used the Classification and Regression Tree and Multilevel Logistic Regression modeling to identify the potential factors related to students’ academic resilience. Using these tools, the study analyzed the B-S-J-G (China) sample in PISA 2015. The variables that significantly predicted whether a student is disadvantaged and resilient (DRS) or not resilient (DNRS) were shown to be: Proportion of teachers in school with master’s degrees, Proportion of teachers in school with bachelor’s degrees, Environmental awareness, Science learning time per week, Number of learning domains with additional instruction, and Students’ expected occupational status. These findings may enlighten governments, teachers, and parents on ways to assist students to be resilient.
Digital problem-solving competence is widely recognized as one of the core skills of the 21st century. A number of important factors influence this competence; some are task-specific pertaining to the problem-solving processes while others are non-task-specific related to knowledge, skills, attitudes and beliefs of the problem solvers, as well as the student learning environment. This study sought to determine important factors that classify student problem-solver as “high-performing expert” versus “low-performing novice”, using computer-generated log files of an exemplary digital problem task assessed in Organization for Economic Co-operation and Development (OECD)’s Programme for International Student Assessment (PISA) 2012 Study. The participants comprise 11,599 fifteen-year-old students from 42 economies. Apart from multilevel logistic regression of problem-solving process and student questionnaire data, the secondary data analysis employed was a data-mining approach involving classification and regression trees. Five important factors were identified that are key to the discrimination of the “expert vs novice” dichotomy.