In Chapter 4 in Volume IV of the Green Book ("Instructional Design Theories and Models"), Bill and Sunny Watson offer five universal principles (1. Personalized instructional goals, 2. Personalized task environment, 3. Personalized scaffolding of instruction, 4. Personalized assessment of performance and learning, and 5. Personalized reflection - with detailed guidance/principles under each) and four situational principles (1. Personalized instruction in time-based systems, 2. Personalized instruction without supporting technology, 3. Personalized instruction for traditional students, and 4. Personalized instruction for online teaching - again with detailed guidance/principles under each).

Thanks, Charlie! I have found that chapter helpful.

The issue I'm focused on for now is identifying and defining the various bases for personalization. Bloom's mastery learning made the decisions based on achievement tests, and varied instructional time (and tutoring, though that was never defined). Some cognitive tutors use a content domain map, but there are a variety of ways of doing those maps (e.g., Anderson, Scandura, etc.). Some systems use personal interests, some use some kind of learning style, some use some kind of personality profile, etc. etc. Some use no systematic criterion, and simply crowdsource either behaviors such as success, time on task, etc., and some use crowdsourced tags with no defined structure. Some systems have a recommender/prescription engine, some do not. Lately I've been working with one of my students to see if there is any definition of personalization underlying MTSS (the California successor to RTI). Recommender/prescription systems, and their underlying analyses, also vary by granularity of analysis, which controls the granularity of prescriptions made by the system. Classroom/teacher systems (without tech) tend to be undefined, with coarse granularity, in my experience.

What I've found is that many case reports and projects don't even mention the basis of their personalization decisions, and rarely is any definition offered. I'd like to see systematic discussion of the dimensions of personalization as a routine part of any project plan in this space.

Roger Schank did an online book as an ASK system with links in the form of questions (http://www.engines4ed.org/hyperbook/). The questions were designed to provide curated pathways. The questions, if I remember correctly, became a bank that could be applied selected during design for any subject area. Following the links generates data about the personalization into standard question buckets that can be used to refine question types to avoid too much expansion of design. The basic argument is that we as learners need both the content and the ability to ask the right questions about it.

From my research on learning by performing, questions related to "how do I..." lead to the performances people want, and performances are supported by learning & deliberate practice. This provides not only the opportunity for explicit knowledge resources linked to interest, but also implicit learning through performance experimentation.

I hope this helps!

Thanks, David!

The issue of knowledge type certainly underlies some schemes of personalization — whether one uses Schank or the procedural vs declarative knowledge you mention.

very helpful!

rob