Personalized Learning with Experience API (#xAPI)
There are already a lot of research on Personalized Learning. When talking about this topic, relevant terms that might be mentioned include Self-Regulated Learning, Self-Directed Learning, Student-Centered Learning, Universal Design for Learning, or even Adaptive Learning. All of these aim to break the assembly-line factory model of learning and customize learning experiences for individuals, not only “offering” personalized activities (push), but also emphasizing the learner’s agency and ownership of learning (pull).
Learners Should Own Their Data and Their Learning
Experience API (xAPI) is praised for being capable of recording granular data (every action, time spent of each event, etc.) and context information (learning goal, content hierarchy, tag, media file uploaded, group activity, device used, location, environmental noise, temperature) and integrating all records from disparate sources in real time. But tracking for the sake of tracking is wrong in terms of privacy concerns, and a waste of resources. We believe that necessary tracking with justified purposes is helpful to help stakeholders, and the first priority should be returning data to learners themselves.
In today’s economy, content knowledge alone is not the critical commodity. The true assets in today’s workplaces also include the skills and dispositions to respond effectively to changes and to problem-solve. Learners need to be able to use tools and data to reflect on learning, to communicate with clarity, to find help from peers and mentors, to be creative in problem-solving, to understand how to find and use information, and to motivate themselves to engage in work persistently. This is also true in preparing K12 learners for college and their careers.
We had a case demo called A Teacher with AcrossX Solution Enabled by #xAPI, in which we simplified feedback to learners by categorizing xAPI statements into the following gamified scores:
- XP: efforts, engagement
- PowerA: competency in subject knowledge
- PowerX: study skills and dispositions, with benchmark and how-to-improve information to guide learners
- Force: soft skills such as cooperative learning and social intelligence
We can see the great potential of xAPI statement for capturing a learner’s multifaceted capabilities beyond just content knowledge (which is the only result that traditional tests can assess). Returning data to learners can help them take actions. This is just a first try; further iterations from real case studies can help refine the design, including modeling higher order skills with learning analytics techniques.
LEAP Learning Framework for Personalized Learning
We believe all learning should be focused on, demonstrated by, and led with the learner, connected meaningfully with mentors, peers and the community.
- Personalized learning is learner focused, providing opportunities that reflect an understanding of each individual’s needs, strengths, interests.
- Personalized learning is learner demonstrated, prioritizing mastery over seat time, and allowing learners to progress at their own speed based on demonstrated competencies.
- Personalized learning is learner led, empowering learners to take ownership of their learning, adjusting dynamically to their skills, curiosity and goals.
Instruction Cycle for Student-Centered Learning from iNACOL
A student-centered instructional cycle is different from a traditional instructional cycle. The central elements described in that cycle form a logical relationship for student learning, as represented in this figure. It’s noted that the learner as co-developer of their own personalized learning plan is crucial. (We might discuss xAPI recipe design for learning plan actions in our next post.)
And the key functional components of student-centered learning is represented as:
xAPI and related dataviz analytics, like Visca, offer the component “Evidence of Learning Functions”, and can adapt to other tools to be integrated. It’s important to facilitate seeking help or peer feedback across applications. The evidence of learning includes not just formative assessment scores, but also learning process data, so that instructors and learners can gain actionable insights of how to improve learning. Returning process data to learners in real time will help them learn “How to Learn” and build a growth mindset.
Major Elements of Personalized Learning System
- Instruction, activities, and resources are aligned to standards or the social and emotional skills students need. What’s needed: a reference framework of competencies and alignment of resources with the framework.
- Instruction is customized, allowing each learner to design learning experiences aligned to their interests. What’s needed: a platform to access myriad learning content, resources, and learning opportunities beyond school, or to report their learning (resource links or written records) outside the platform, that enables instructors and learners to collaborate on learning plans.
- The pace of learning is varied based on individual student needs, allowing students to accelerate or take additional time based on their level of mastery. What’s needed: integration of multiple systems and data flows with interoperability, computer-based assessments and machine intelligence for real time feedback and recommendation.
- Instructors use data from embedded assessments, formative assessments, and student feedback in real-time to differentiate instruction, support, and interventions. Learners have multiple ways of demonstrating their learning. Frequent student progress checks and dynamic teacher responses to student data are necessary. What’s needed: frequent and diversified assessments aligned to objectives and dashboards that reveal real time learning progress, engagement, and struggle, and pinpoint at-risk learners.
- Students (and parents) have access to clear learning objectives, learning process data, and assessment results compared with cohort or larger group. What’s needed: customized learner dashboard and profile that combine data from disparate systems and input from peers, parents, educators, and others who work with the learner.
xAPI’s role is to enable the data capture, flow, and interoperability needed in the whole picture. As learning technologies advance or diversify, xAPI can still serve the distributed experiences. It doesn’t matter where contents are or where activities happen, data flow and learning design are both freed. Ideal personalized, learner-centered learning is anywhere anytime learning. Even we can try to integrate applications by Single Signed On (SSO) and let users launch disparate resources with iframe within the integrated scope. Many learning activities cannot be launched from a web-based system, such as mobile applications and Augmented Reality(AR) and Virtual Reality(VR). xAPI was born to solve the data integration issue.
Learners and teachers can read xAPI data within minutes by leveraging dataviz or use Machine Learning algorithms to provide recommendations to learners. The goal of using xAPI is to build a Total Learning Architecture (TLA) and Personal Assistant for Learning (PAL). PAL could be an assistant navigator for the personal learning journey. The statements are in the form of “Actor”+”Verb”+”Object”+…. But, we don’t think the “Actor” is only a puppet in a given story. They are characters in their own stories and in the stories of their peers. Learners should own their data and their learning, participate in creating their own learning stories, and build their portable ePortfolios.
“Learning is most effective when it’s personalised; it means something to the learner. That happens when people feel they are participants and investors in their own learning, shaping what and how they learn, and able to articulate its value to them.” — Leadbeater, Charles
 Glowa, L. and Goodell, J. (2016) Student-Centered Learning: Functional Requirements for Integrated Systems to Optimize Learning Vienna, VA.: International Association for K-12 Online Learning (iNACOL).