On day 2 of the Intelligent Tutoring System Conferenre 2012, the first talk was held by Jennifer Sabourin from the North Carolina State University, and she elaborated on learner self regulated strategies in game-based learning environments. In the domain of game-based learning, game technology fosters engagement, promotes agency, interactivity and increases motivation and interest. Especially, open-ended environments promote exploration, agency and self-guided learning. What can we learn from self regulated learning in open gaming environments?
In game-based learning environments, success is gained when the learner identifies his or her learning objectives, utilizes the learning Â recourses given and guides his or her own learning. Without these skills, learners may flounder and not receive the full educational value of the game. The question that rises here is formulated as follows; can we identify which learners need additional support, without compromising their game experience?Â
To answer this question, Sabouring reflected on the concept of Self Regulated Learning in which students take an active role in setting and achieving learning goals. Typical behavior includes goal setting, monitoring and measuring progress and adjusting behavior based on success and failure. SRL skills can be improved with practice and instruction, therefore, a game-based learning environment was constructed that focussed on compelling features such as narrative and interactivity. However, open-ended learning environments pose additional issues, such as, no clear indication of game progress and uncertainty of how to set or accomplish goals.Â
With these issues in mind, students were situated in an open game environment. The environment represented a 3D island in which the learner would walk around and typically talk to agents, and answer questions, based on the information given by these agents. During the game, students were prompted to update their mood and status. Their status was then send to an in-game ‘social network’.Â
The status updates included evidence of self regulation. Status updates that were tagged for evidence of Self Regulated Learning included; goals setting and reflection. So, these 2 types of behavior needed to be included in the status updated, in order to tag it as Self Regulated Learning. Then, SRL was ranked according to 4 classifications; specific reflection, general reflection, non-reflective statements and unrelated statements.
Specific reflection is reflection in which the learner evaluates progress towards a specific goal or area (“I’m trying to find X in order to see what caused Y”) General reflection is a state in which the learner evaluates progress or knowledge, but does no reference a specific goal (“I think I’m getting it”). Non-reflective statements include statements that do no provide reflective evaluation (“I am doing this task”). Unrelated statements are statements that do not fall into a previous mentioned category.Â Then, all statements were tagged and students were even tertiary split into three categories; High, Medium, and Low SRL students.Â
It turned out that High and Medium SRL students had significantly higher learning gains than Low SRL students. Sabouring explained that Low SRL students may have not used the game recourses effectively and that student SRL skills need to be identified early in the game if you want to provide adaptive scaffolding. Undirected prompts (the status updates the learners had to make) could therefore have the ability to uncover the evidence of self regulation.Â
Self Regulated Learners tend to make better use of curricular recourses and, as said, achieved higher learning gains. A machine learning model that predicts SRL looks promising in this case. In this way, predicted SRL can be compared agains more validated measures.