Аңдатпа
The article discusses the experimental use of different models and methods of the university’s digital ecosystem in the process of modernization of higher education. The digital ecosystem is an environment that integrates the educational, scientific, and managerial digital resources. The study also responds to the lack of focus in previous studies, which usually only use individual indicators to report the effectiveness of digital learning models, while not using indicators that are comparable with the results from previous studies. This study aims at quantifying and comparing traditional, hybrid and digital learning models and evaluating their effect on academic achievement, motivation and engagement. The quasi-experimental method was performed on 90 students and 30 teachers (N = 120) in the control and experimental groups for a semester-long period. The instrument for measuring motivation was an adaptation of the Motivated Strategies for Learning Questionnaire (MSLQ), and the instrument for measuring engagement with a learning-analytics activity index was a learning-analytics activity index. Pre- and post-tests were used to measure academic performance. Paired and independent-samples ttests, effect sizes (Cohen’s d) and a proposed integral effectiveness indicator that combines the three sub-indicators together into one score were used for data analysis. Results showed significant improvements in academic performance and motivation in the experimental group, with higher post-test values than in the control group. The findings showed a higher reported integral composite for the digital learning model (68.4%) than for the traditional model (58.1%). The experimental group also showed statistically significant gains in academic performance and motivation, while platform activity was analyzed as a behavioral digital-use indicator rather than as an independent engagement measure. The value of the work is that universities have a replicable approach to designing, comparing, and evaluating digital ecosystem learning models.