The Intersection of AI Educational Psychology and Learning Analytics Predicting Student Dropout Risk through Behavioural Indicators

This study investigated the predictive power of behavioral and psychological risk factors in predicting dropout rates in online and hybrid higher education programs. 250 undergraduate students enrolled in Pakistani public and private universities were the subjects of data collection using the quantitative, correlational, and predictive research methodology. Simple linear regression, multiple linear regression analysis, and the Pearson correlation were used for sample analysis and simple random sampling. The result showed that the behavioral predictors, namely digital engagement patterns, showed significant predictive capability regarding the dropout risk and its explanation of 37.4 percent of the variance. The influence of the psychological aspects (academic self-efficacy and emotion control) on the student retention was positive and moderate. Incorporation of both behavioral and psychological factors into a multiple regressive model enhanced the relationship of regression between the outcomes of dropout and variance of 50.7. These results confirm that there is the need to combine the applied learning analytics based on AI with the lectures on educational psychology to arrive at comprehensive, time-bound, morally correct intervention strategies to curb dropout in the online learning platforms. This study is interdisciplinary and is a contribution to the already built body of literature in the study area, and it contains some practical implications for those universities interested in further improving student retention even in what has become a virtual learning environment. Provisionally, it also presents the need for ethically specific early-warning systems that can be applied so as to attain academic and emotional success of students.