Personalized learning is an umbrella term for a related set of concerns around tailoring learning opportunities and instruction to meet the needs and interests of learners. Many characteristics of learners are posited as being ripe for tailoring, including cultural backgrounds, language, knowledge, skills, abilities, interests, learning style, common conceptions, etc. In our research, we focus on cognitive concerns: adapting instruction to support differences in learners' knowledge, skills, abilities, and conceptions. We also focus on intentional learning environments; i.e. settings where there are explicit learning objectives. This describes most K-12 environments, where teachers are accountable to state and local standards. In these settings, there is a growing recognition that while there may be shared learning objectives, due to the "extreme diversity" in many classrooms there needs to be multiple methods for students to approach shared objectives and multiple ways for them to apply and demonstrate their knowledge.
Many of my colleagues equate "personalized learning" with "individualized learning"; i.e., the locus of personalization being the individual learner. This is a somewhat extreme perspective as many social and collaborative models offer excellent opportunities for personalizing of teaching and learning. In our work, we often talk about "customizing learning" in an effort to shift the locus beyond the individual to include groups of learners.
The broad range of concerns covered under this umbrella term is exemplified in the diverse range of approaches to supporting personalized learning:
- Pedagogical approaches: This includes approaches to supporting student inquiry, collaborative learning, and self-directed learning. Tomlinson offers practical advice and best practices for teachers in her books.
- Formative assessments: This includes a range of work on developing questions, instruments, and methods to diagnose current student understanding; the idea being that it is challenging to personalize learning without first getting a handle on the learner! The work that Bill Penuel and colleagues at SRI are doing in this area on "contingent pedagogies" is a very interesting, technology-supported approach.
- Student-centered constructivist environments: These efforts focus on creating a specific kind of learning experience where students can create their own objectives and research questions, construct their own knowledge, and often learn through collaboration with their peers. The work of the MIT Media Lab exemplifies this approach, as does the work of Stahl and colleagues from Drexel with their Virtual Math Tutors project.
- Tutoring environments: These environments try to emulate the experience of one-on-one tutoring between and individual student and a knowledgeable instructor. Typically, these systems are underpinned by detailed models of the domain, common student misconceptions, and cognitive models of learning. The algebra tutors from Carnegie Mellon being the most prominent examples.
- Open educational resources: Open educational resources (Oer) are teaching and learning resources that reside in the public domain or have been released under licensing schemes that allow their free use or customization by others. This "movement" has fueled the rapid expansion of free educational content available on the web. Central to the vision of Oer is the assumption of local adaptation; namely, that scientists, educators, and learners will find useful Oer available on the World Wide Web and in specialized repositories (reuse), adapt and/or combine them to better meet their specific needs (remix), and share their new or revised resources with others. The Hewlett Foundation and the National Science Foundation have been big supporters through programs such as the National Science Digital Library and OER Commons. The work that Mimi Recker from Utah State has done over the years in developing tools supporting teachers' peer-production processes is noteworthy in this area. In our research, we have contributed to the design and evaluation of educational digital libraries (DLESE, NSDL) and we have developed software tools and web services, such as the Strand Map Service, which enable the creation of standards-based visualizations and conceptual browsing interfaces around Oer.
- Personalization services: There is a growing area of research-driven activity where researchers are developing educationally-focused recommendation engines and sophisticated statistical models that help to address "long tail" concerns arising from living in a content-rich world; namely, how to match people to content and match people to people (to support peer learning). Much of our basic research lies in this area. In one project we are studying conceptual personalization: how to select resources that address learners' specific knowledge gaps and misconceptions. We are developing natural language processing algorithms that analyze student work to detect potential knowledge gaps and misconceptions (currently working with essays) and recommendation engines that take these diagnoses and make personalized resource selections. In another project, we are trying to support the meta-cognitive skills of teachers and learners through the development of automatic quality assessment models. These models use sophisticated algorithms combining machine learning and natural language processing methodologies to automatically analyze Oer along a number of dimensions important to teaching and learning, such as "organized around learning goals", "age appropriateness", and "effective use of representations." The goal is not to produce a single thumbs-up-or-down decision on the overall quality of a resource; rather the goal is to produce a rich profile characterizing the different strengths and weaknesses of a resource to aid human cognition and judgments. We envision these models helping teachers and learners to make better choices about resources to support their personalized learning and instructional needs, and help users to make more strategic use of OER in their own peer-productions; i.e., leveraging the strengths of particular resources and, conversely, compensating for the weaknesses of resources they incorporate in their own products. In both of these projects, our goal is not to create a specific learning or teaching environment per se, rather our goal is to embody these capabilities into web services that can be embedded in a rich variety of teaching and learning applications.