Abstract:
Self-regulated learning (SRL) has become one of the most important theoretical concepts in educational research. In light of contemporary educational challenges, including the widespread use of information technology in educational settings, the growing focus on enabling students to become lifelong learners, and the increased emphasis on learner-controlled learning activities, SRL further shows a significant practical importance. The ability to effectively regulate learning processes is a key skill for learners to meet the aforementioned challenges. Typically, SRL is referred to as the regulation and control of cognitive, metacognitive, motivational, as well as affective states and processes in service of learning goals. Following this definition, a broad body of literature investigating SRL from different theoretical backgrounds and perspectives has shown that SRL is key factor for students’ academic success throughout all stages of education. However, the diversity in approaches to investigate SRL has also led to lack of clarity what SRL is and how it can be most effectively fostered. This issues becomes even more apparent when SRL is investigated in the context of other, more general research traditions on self-regulation (SR). The present dissertation addressed this research issue by integrating four areas of research on (SR) in education. These were derived from the mechanisms through which self- regulatory variables affect learning and include learning activities (e.g., cognitive and metacognitive strategies), driving forces (e.g., motivation and affect), personal dispositions (e.g., personality), and limited resources (e.g., working memory and executive functions). Specifically, based on research that has strongly linked each of these areas of research to learning and academic achievement, an integrative framework that situates SRL as part of SR in education has been proposed. To test this framework, the present dissertation tested the predictive value of key constructs representing all areas of proposed framework across different contexts (e.g., learning in school and laboratory learning task). Through this approach, this dissertation is the first study that empirically integrated the aforementioned research traditions on self- regulation in education.
Study I aimed at identifying the best predictors of learning in school and laboratory learning tasks from a comprehensive set of self-regulatory constructs that reflect the four areas of research on self-regulation proposed in the framework (i.e.,
learning activities, driving forces, personal dispositions, and limited resources). Specifically, robust machine learning predictions were used to predict performance in school and laboratory learning task across five academic domains (i.e., math, physics, biology, art, and history). Results showed that predictors from all areas of the proposed framework are required to optimally predict learning in both settings. However, the specific variables that optimally predicted learning in school and laboratory learning tasks varied. While measures of driving forces (i.e., motivation) and limited resources (i.e., working memory capacity) predicted learning in both settings, predictors representing learning activities (e.g., effort-related vs rehearsal strategies) and personality (e.g., openness) only showed predictive value for one of the outcomes.
Study II investigated if and how self-regulatory requirements in a computer- based learning task differed depending on the way participants interacted with the learning environment. In detail, participants used either mouse-based or touch-based interaction to work with the learning materials. Robust machine learning models predicting learning outcomes in both conditions were developed. Specifically, these models used measures that represent the four core areas of the proposed framework similar to Study I. Results showed that self-regulatory requirements were higher when learning with tablets. Specifically, beyond the predictive value of prior knowledge, learning on tablet was determined by critical evaluation (learning activity), motivational cost (driving force), openness (personal disposition), and switching (limited resource). Differences in performance using mouse-based interactions on the other hand were only related to control measures (reading comprehension and prior knowledge) but not related to self-regulatory constructs.
Study III extended the scope of the first two studies to a detailed, process-oriented investigation of one key area of the proposed framework. In this study the emotional experience of participants (driving force) and its temporal unfolding throughout a learning activity was related to learning. Results showed that a group of students with primarily negative emotional experiences learned the least. Moreover, these students showed an increase in negative emotionality during learning that was predictive of lower learning outcomes. Lastly, additional analyses demonstrated that these emotional processes are related to stable personal dispositions (i.e., trait emotion regulation and neuroticism).
Overall, across all three studies this dissertation has shown that SR shares a common underlying structure across contexts. However, the specific SR processes
required to achieve optimal learning outcomes differ depending on the learning task, context and environment. Through these findings, this dissertation provides a theoretically derived and empirically supported theoretical framework, that situates self-regulated learning within the larger context of self-regulation in education. The findings of the studies are discussed in light of the proposed framework and the added value of a broader conceptualization of SR in education. Key steps for future research programs to extend upon this framework and integrate research traditions on self- regulation in education are derived.
Print-on-Demand