Sharebar?

Learning Impact Blog


Community leadership for more effective use of technology in service to education

#5 of a series in preparation for May 2016 Learning Impact Leadership Institute

After a 1-week pause to focus on the release of the IMS Annual Report, I’m back at the topic of the anatomy of an educational or learning app.

ASU GSV why can't we be friends: interoperability in school districts

In the previous post of the series I provided a high level overview of the interoperable inputs that an educational app might utilize.  Now it’s time to look at the outputs.  

It is interesting to note that last week I was on a panel at the well-known ASU GSV Summit entitled, “Why Can’t We Be Friends: Vendor Interoperability in Districts.” The salient takeaway for this post is that there was great agreement that while the sector is making some pretty clear progress on the input side of interoperability, the output side is pretty early in its development. So, get ready for the harder part!

As with my discussion on the inputs I am going to refer to the Anatomy of an Interoperable Learning App figure shown here.  And, I am going to focus on what makes these outputs “next generation” in comparison to today.

Anatomy of an Interoperable Learning AppFor next gen learning apps the objective is to enable better information outputs to understand progress, further personalize the learning experience and just in general understand the usage of various digital resources. Shown in the figure are some categories of potential outputs (on the right hand side) that are discussed further here:

Activity: Some IMS members call this aspect of digital resources “adoption.” As we evolve into the next generation of scaling the number and types of digital educational resources and applications one of the first questions that needs to be answered is, “Are teachers and students using these resources?” One function of this is for administrators, teachers and suppliers to get an understanding of what is being used, and perhaps be able to infer why or why not?  One of the major deficits of research into use of digital resources is the dependence on user surveys. It is much better to have ways to directly measure usage. And, once usage can be measured accurately it enables the potential that use of specific resources or combinations of resources may be statistically correlated with outcomes via various assessment mechanisms.  Activity information is also very important to the content creators/publishers because it enables a better understanding of how digital resources are being used and thus provide clues for improvement.

Outcomes:  The most effective educational platforms and apps will be those that can indicate to the student, parent and teacher how well the learning is progressing. Of course, the term “outcomes” can have a very broad definition in education.  However, the thought here is that the educational application will have some way to report progress, including scores to measure success on a specific activity, “gradebooks” that are compilations of scores, and even comparisons to the progress of other users.  The challenge is that different applications will measure progress differently, and thus be difficult to compare and analyze.  From the perspective of interoperability, one potential big win is a usable way to collect all the progress information on a single student on a single screen or other usable interface. Today, many tools may have various progress reports or dashboards, but they are resident “inside the application” and thus not readily looked at in conjunction with progress information from other apps.  Another big win is agreement on some common summary information across a wide variety of learning applications.  The holy grail perhaps, depending on your view of the viability of common learning standards, is an agreed upon approach for all apps and platforms to report on progress by specific learning standards.  Finally, another very important type of outcome data in education is assessment “item data,” which is the collected responses to assessment items.

Credentials: There’s a lot of talk these days about the potential advantages of micro-credentials in education.  Micro-credentials are ways to recognize granular achievements or cross-curriculum competencies that are highly valuable and typically not inferred by a higher-level construct such as a course grade or GPA. Such achievements or competencies may be extremely valuable in matching a person to a job or career.  In the educational context, micro-credentials might do a much better job than grades in determining if a student is “college ready” in terms of meeting specific prerequisites.  In the context of next generation digital learning there is the distinct possibility that a learning application may provide evidence toward a defined micro-credential or even a validation of such a credential.  The interoperability that will be required will be a way to combine or mash-up such credentials. IMS is working to apply open badge extensions and modular electronic competency and transcripts to enable credential validation and mash-up. The current state of the market is quite early and exploratory. So, we have a long way to go – but this is very important work for the future that may find useful application in those institutions that are doing a better job at micro-credentialing  and competencies.

Artifacts: This very important area of next generation learning is potentially related to credentials. It is about how to capture, in an effective manner, what the learner may have created during their interactions with the application.  While an artifact is not a credential, it can provide evidence of competency, creativity, agency, etc.  The ability to capture meaningful artifacts can help a learner reflect on and assess their own learning, reinforce past learning experiences and display their work to others.   An e-portfolio is a type of learning app that is designed for this purpose.  But, artifacts can be created in many other types of learning applications and platforms. Next generation learning requires not only interoperable publishing mechanisms, but also interoperable metadata to enable classifying, sorting and searching artifacts.

Ramifications on Next Gen Architectures

From the above, one can see that there are a broad range of possible outputs from learning applications and platforms.  As previously noted, interoperability of the outputs of learning applications is a complex topic that needs a lot or work.   It is also much more variable across different educational levels and types than the inputs. Highly variable areas that are early in their maturity are usually not good topics to attempt to get industry agreement on.  Therefore, we can reasonably make some conclusion regarding interoperability of outputs for next generation learning

  1. Interoperability of activity, outcomes and credentials will need to be structured to enable affordable collection of data that supports research on potential correlations. 
  2. A very high-level of cooperation will be required among suppliers of a wide range of products to establish the interoperability standards and enable the data exchange to achieve #1.
  3. For #1 and #2 to accelerate understanding of the impact of learning apps and platforms, industry must set up and sustain funding models for cross-industry research.
  4. There will also need to be some “quick win” areas for interoperability of outputs.  There are some ideas for such quick wins in the descriptions of the categories of outputs above.  By “wins” I mean providing value to institutions, users and suppliers.  Another way to state this is that “the road of establishing the interoperability for next generation learning must provide current generation benefits to all market participants.”

Next up in the series:  What education sector realities are telling us about the likely evolution toward next generation digital learning architectures?

With the release this week of the IMS Global Annual Report 2015, I wanted to attempt to summarize some ways in which “our little engine of change” is helping to enable some potentially pretty profound stuff.

A little more than two years ago (December 2013) I was asked by Diana Oblinger (CEO of EDUCAUSE), to present, along with Jack Suess (CIO at University of Maryland Baltimore County),  a “game changer” webinar discussing the recently published EDUCAUSE Review article, “A New Architecture for Learning.”  We decided to entitle the webinar:  “Building the Connected Learning Platform – One Brick at a Time.”

Of course, it was flattering to be considered a “game changer” – whatever that term really means. Most people, including technology veterans, would probably think of interoperability standards (the work of IMS Global and important basis for above paper and webinar) as being “part of the game” rather than necessarily “changing the game.”  I would include myself in that group when thinking about the large majority of standards-setting activities. 

However, I did think it possible that the work of IMS Global could be a “game changer” when I became the CEO a little over 10 years ago.  Now I believe in the “game-changing” nature of IMS even more.  I’d like to enumerate here some of the specific ways that IMS has been and is changing the “edtech game” (albeit never fast enough for me!).

#1 Providing a foundation for collaboration that fits where the edtech sector wants to go in terms of diversity of platforms, resources, apps and tools. The rules of the game are different when an industry is dominated by proprietary walled-gardens then if there is plug and play interoperability of tools among a wide variety of platforms.  As consumers, we live all day long in proprietary walled gardens created by a few dominant platform suppliers with app stores and mobile devices. However, the education sector has figured out that the idea of just a few dominant platforms just does not make sense for the sector. For instance, most K-12 districts are compelled to support both Microsoft and Google. IMS standards, and more importantly the critical mass of IMS members leading the charge, don’t see a few dominate platforms as the future of edtech (to the MBA people out there that will say that every market moves in the direction of a handful of dominate providers, I just want to say that I know those arguments well, but the education sector is very, very different). I’m very pleased to report that the IMS community is leading a better way to build an ecosystem based on what the education sector requires, namely products that can actually work together to improve teaching and learning while keeping lock-in and barriers to entry low to spur innovation.

#2 Providing an interoperability development process that is based on engaging directly with suppliers and institutions. Much of the work on standards in EdTech has been very visionary and developed by some really brilliant people. And, my hats off to the great work.  However, the missing link to make this work payoff for all parties concerned is a necessary reality check of compelling sector needs and market adoption. IMS has added this extremely important element in the last 10 years.  For those familiar with venture capital or market development approaches, the IMS process is very much a funnel where lots of good ideas go in that are then processed through market validation.  So, to get out (become a published IMS standards) there has to be an acceptable combination of market, financial, people and technical risk (which are typically the four dimensions of risk that VC firms look at).

#3 Focusing on enabling the plug and play ecosystem that results in more investment on innovation. IMS is creating a major culture shift in EdTech standards, namely that the measure of a successful standard is that it helps to lift up the entire sector.  IMS sees our job as enabling an interoperable ecosystem – what has already become hundreds and in the future thousands of products that work with each other “out of the box” (or as close as possible).  Time and cost of integration goes to zero.  When that ecosystem happens it focuses investment on the innovation of the individual products themselves and also of what can occur when products work together. To anyone familiar with some of the more mainstream standards in the tech world that were huge successes, like Ethernet, the Worldwide Web, or wireless standards, you would most likely say, “isn’t that the point of most standards?” A good standard reduces cost while increasing connection into a large ecosystem, which means that suppliers are naturally motivated to adopt.

#4 Accelerating interoperability progress through synergies across K-12 and higher education.  I’ve gone on record many times that I believe there are many differences between the needs of K-12, higher education and corporate training. I know from actual experience that if you try to build a product for all of them you are unlikely to succeed with any of them. Thus, when the U.S. K-12 sector came knocking on the door of IMS about 5 years back I had an open mind, but was cautious about how it would work out. Well, it has worked out great! What we have found is that one sector can take work coming from the other sector, get use out of it, and then, amazingly, take it further in ways that have then gone back the other way to help the originating sector.  It has taken some very good market development and project filtering skills (see #2 above) to encourage the right focus at the right times. But, it is very clear that IMS is now a 50/50 partnership between higher education and K-12 – which of course means that each sector is leveraging the investment and ecosystem of the other. The bigger the ecosystem the more valuable the standards.

#5 Creating an EdTech ecosystem leadership community that will lift the sector in the sort term and is likely to endure in the long term.  This is really the most important impact that IMS is having. As I mentioned in a recent EDUCAUSE Review article, “Foundations Past and Future,” IMS is really all about education sector participants willing to lead a critical aspect of cooperation that will accelerate the innovation we need to improve education. There are, of course, many other critically important areas where cooperation and communities are developing with complementary foci. But, I do believe that interoperability is an underlying foundation for much of the rest that requires strong leadership collaboration.  The beautiful thing is that we are finding that those organizations, institutions or suppliers, that are investing time in leadership in IMS, are doing very well. It turns out that to be technically interoperable also requires organizational leadership that is interoperable. It’s a big claim to make, but I will make it. Namely that we are seeing IT, curriculum and instructional leadership and strategies getting to the next level when an organization works on and “gets” interoperability.  We are seeing a focus that goes way beyond the historical desire to just have the latest technology to being able to leverage technology toward personalized learning and other strategic goals.   As indicated in the figure, the net result is a financially strong (while still small compared to many education associations such as EDUCAUSE or ISTE) IMS organization that can play a substantial ongoing role in developing the EdTech ecosystem.

As a result of the above innovations, led by the IMS members, I do think the trajectory of EdTech is changing in a very positive way. I’m very proud of the IMS members stepping up to take on the challenge of working together on building an open and innovative EdTech ecosystem.

 

#4 of a series in preparation for May 2016 Learning Impact Leadership Institute

In the previous post of the series I laid out a perspective on interoperability for next generation digital learning environments from the perspective of an application being able to fit into a configurable constellation of educational apps.  I would now like to elucidate the key features that make said app and environment “next generation.” I will once again refer to the “Anatomy of an Interoperable Learning App” figure introduced in the previous post.

What might next gen learning apps know about the learner? What should they be allowed to know? What can be personalized in the user experience from that knowledge?

For next gen learning apps the objective is to enable better information inputs to allow the app to personalize the learning experience.  What is the information and what does the app do with it? Shown in the figure are some categories of potential inputs (on the left hand side) that are discussed further here:

Institutional context: This is the type of information that is typically found in a student information system (students, student grade level, courses, sections or other groupings). There are potential common languages for this from interoperability standards like OneRoster and Learning Information Services, as well as the Rostering (aka Membership) Services of Learning Tools Interoperability. While exchange of this sort of information is not especially “next gen,” enabling this exchange without the need for custom integrations and code writing is. The information is typically used to authorize use of the app as well as group users together (such as in a collaborative app). Information about who the user is can also be used to bring the student back to the place where they left off, store results in progress specific to the user, etc.

User context: This is information about the learning activities that the student is currently engaged in (reading, assessment, discussion, assignment, media, etc) when the app is launched.  User context is also information that the app can generate and output in the form of an activity description.  User context is an area that requires more work to develop a common language.  Standards work applicable to this is Caliper Analytics and Learning Design.  User context information could be used by the app to personalize how the app behaves. For instance, an assessment app might act differently if it is launched in conjunction with a reading activity, versus if it is launched as a separate assessment.  Again, this is a more advanced topic.  It is very “next gen” in that it enables an app to respond to the circumstances (activity sequence) in which it is launched.

User Preferences:  We all know that apps typically have some ability to allow the user to specify preferences.  For “next gen” digital learning environments to be personalized in a scalable way, there needs to be a way for the user to specify preferences one time and have that information propagated to all relevant apps.  These are not “app specific” preferences, but rather preferences that could be pertinent to most apps.  For instance, when a user interacts with apps on their smart phone they may have certain preferences for size of fonts, use of media, etc.  There is a rich body of interoperability work called Access for All that has developed a rich framework for describing personal needs and preferences (PNP).  Access for All is applicable to all users, but also has a rich foundation in accessibility (and has even been published as an ISO/IEC standard). PNP has been applied to high stakes assessment accessibility via the Accessible Portable Item Protocol (APIP) standard.  One of the most important preference areas in education is privacy.  While the “setting privacy preferences in every app” model of consumer app stores may make sense for that world, it does not make sense for education. Students or parents need to be able to set privacy preferences once in the context of their institutional experience and then have those privacy options set as defaults for every app.  Therefore, for “next gen” learning environments, our expectation is that the privacy preferences will be selected from an interoperable privacy framework that is provided as an input to each app for each user.

Learner Profiles:  One of the potentially most valuable interoperable inputs for next gen learning apps is a learner profile. This is information that lets the app know where the student is in their learning experience and progression. It can be used to provide a personalized experience for the student. In some ways, the learner profile is the “holy grail” of next gen learning in that the better a learner’s “state” can be described the more personalized the learning experience could be. The problem of course is that no one knows exactly how to describe what a person knows or doesn’t know.  A learner “state” must be generated and kept for every adaptive learning or assessment app.  Is that state record standardized or interoperable? No, not at this point. In fact, such a description is probably considered to be the “secret sauce” of such products.  If the products are tracking progress toward agreed upon competencies or learning standards (like the U.S. K-12 Common Core) then it makes the possibilities of exchanging learning profiles greater.  But there are still a lot of nuances.  More light will potentially emerge at the end of that tunnel as adaptive summative testing becomes more mature.  Such testing will require well-defined “levels” of mastery to agreed upon learning standards.  However, even in the short-term there are some interesting possibilities.  For instance, in IMS right now we see tow more tractable paths to building and exchanging learner profiles.  In higher education the IMS community is developing ways to exchange competency frameworks among cooperating products in the educational enterprise.  In K-12 the IMS community is looking at taking small, but potentially powerful steps looking at profile items like reading level.  IMS does have some past work on learner profiles and competencies as well that may become applicable: Learner Information Package, ePortfolio and Reusable Definition of Competency or Educational Objective.

Ramifications on Next Gen Architectures

Getting the architecture “right” for next generation digital learning presents numerous opportunities and challenges.  Here are some questions that the education community needs to think through:

  1. Is there a software application that is best suited to be the keeper and provider (in terms of interoperability) of learner profiles? Should the learning management system or student information system do this?  Or, perhaps a competency management product? What about the new category that the Gates Foundation has been encouraging, Integrated Planning and Advising Systems (IPAS). Yet another possibility is a competency-oriented assessment product.
  2. Same question for entering, storing and sharing personal needs and preferences, including the all important privacy preferences?
  3. What pragmatic steps do we take in moving our institutional or product architecture forward as we build toward this future?

Next up in the series:  Deeper dive on next generation digital learning environment (NGDLE) interoperability: part b: interoperable app outputs

 

#3 of a series in preparation for May 2016 Learning Impact Leadership Institute

Now we begin to get into some of the meatier topics when it comes to evolving from the world of today to next generation digital learning environments.

One of the challenges of even envisioning, much less implementing, a configurable constellation of educational apps (see previous post in this series on metaphors) is the notion that potentially “everything is connected to everything else.” How does that scale? I will consider the scale issues in a later post.  But, for now, let’s just consider the interoperability needed to enable next generation digital learning.

It is very tempting here to elucidate the various types of products, functions and services that must come together in the delivery of an institutional education experience. Indeed I have often gone down that road for the sake of making the discussion very tangible.  Shown here is a figure from the precursor paper to the NGDLE, A New Architecture for Learning, along those lines.

Another, more detailed treatment along these lines, has been published by IMS member SURF of the Netherlands. The paper, entitled, “A Flexible and Personal Learning Environment,” is a really great survey of ideas and realities from a functional (“components”) perspective.

For this post I am going to take a different perspective, namely looking at the anatomy of the interoperable educational or learning app.  If you’re into legos than this is akin to thinking about how the legos connect to enable personalized learning environments.

For this “app-centric” view I offer the following figure depicting the potential anatomy of an interoperable educational or learning app:

The view shown here is informed by numerous insights from many years of IMS members working on many, if not all, aspects of this challenge. However, it is also flavored by the “dimensions” enumerated in the Next Generation Digital Learning Environment research paper, namely:

  • Personalization
  • Analytics, Advising, and Learning Assessment
  • Collaboration
  • Accessibility and Universal Design

Ignoring the detailed information on inputs and outputs for a moment, the following overview statements apply to the figure:

  • An “app” is different than “content” in that content does not process inputs or outputs, while an “app” does
  • While not required, an app can “mash-up” other educational apps and content using the same structures shown (note the reference to mash-up per the NGDLE research paper:  “If the paradigm for the NGDLE is a digital confederation of components, the model for the NGDLE architecture may be the mash-up. The confederation-based NGDLE will be mashed up at both the individual and the institutional levels, as opposed to consortia forming to create open enterprise applications.”)
  • An app may also produce content or app launch links that can be mashed-up into other apps
  • An app does not necessarily support all of the inputs, outputs or management interfaces shown – only those required for what it supports
  • An app typically has a web back-end as well as a web or mobile client, however, some apps may be mobile only

What can be mashed up?  Again, mash-up support is not required but if supported enables an app to leverage the ecosystem of interoperable apps and content.  Based on IMS experience we see three categories of mash-up:

  • App interoperability: For example an app launched and integrated using LTI.
  • Assessment interoperability: For example assessment items using QTI.  Assessment items and tests are a very important and specialized category of content in education and learning that merits a specialized language for interoperability.  
  • Content & app manifest interoperability: For example common cartridge or thin common cartridge. A “manifest” is a list of the assets, some of which may be available locally and some of which are web accessible. Note that both varieties of common cartridge support LTI app launch links.

It is also worth noting the reference in the figure to “IMS App SDK and APIs.” While it is not the official or even necessarily majority view of IMS members that IMS will ever have a production-worthy developers kit, it is my own personal view that this is the direction that makes most sense for enabling high levels of interoperability and cost reduction. It is also an extrapolation of what is occurring in most IMS spec areas, namely development of APIs, code libraries, sample code, etc.

Again, the above is my attempt an interoperable app definition that we are building towards to enable the NGDLE.   We have seen over and over again that the evolution will take time and respond to market needs.

Next up in the series I will explore more explicitly how the interoperability features shown in the Interoperable App Anatomy support the NGDLE dimensions, including some potential ramifications on the enterprise apps required.