Last fall, ACTNext, the innovation unit of ACT, unveiled its new tool for learning assessment systems (LASs), the Recommendations and Diagnostics (RAD) application programming interface (API).

Before I get into how the tool is helping students and its potential for benefiting EdTech companies, let’s cover a few definitions to clarify the alphabet soup:

A learning assessment system (LAS) contains information on specific learning goals or standards aligned to curriculum, learning and assessment. This provides students and educators with information to navigate the next steps of the learning process to achieve learning goals or standards.

Next, an application programming interface (API) allows different software systems to talk to each other. APIs act as connectors between customers and companies.

So, back to the learning assessment system tool ACTNext built: the RAD API.

RAD delivers on ACT’s promise to transform learning through the development of holistic education technology, utilizing 60 years of ACT research and item response data. It’s a clever approach that integrates assessment with personalized and adaptive learning.

And it’s a reality today because of ACTNext’s pioneering work to merge artificial intelligence, machine learning and algorithmic development with computational psychometrics, the science of measuring mental capacities.

How does RAD help learners?

RAD tracks test results and learner evidence over time to build a diagnostic model of the skills students have mastered and ones still requiring review. Building those diagnostic models then drives a “recommendations engine” that provides specific sets of personalized learning resources to students.

Once the diagnostic model identifies a subject area, like math or science, where learners can benefit from review, then they’re connected to resources. This gives specific practice and prep work, such as providing learning resources for linear equations within algebra, for example.

When learners navigate to the resources tab, they can choose an area of study by going to a subject category. RAD builds personalized lists of recommendations based on identified areas of need. For example, complex number operations have floated to the top for this learner's number and quantity area in math.

How does it work?

RAD gathers learner data in real time and applies results immediately.

Here‘s how it works: A student takes a quiz. Then, the learning assessment system relays the results of the quiz (in what’s known as a standard data format using the Caliper standard),
to the API. RAD then uses that knowledge to update the learner’s diagnostic model.

The RAD engine not only supports individual learners, it delivers insights gleaned from student populations in aggregate. RAD fine-tunes learning recommendations according to the difficulty of a skill (or skill area) and can detect misconceptions in a subject by evaluating patterns of response selections across student populations. In the future, RAD will be able to adaptively change content through instructional materials.

RAD generates snapshots of learner knowledge that are sent to the recommendations engine.

Our team had a “light bulb moment” after creating ACTNext’s prototype learning mobile app, called the Education Companion, when we realized RAD could serve other learning assessment systems. That was a big moment for the development team!

We found the RAD engine can be extended and configured for any learning assessment system that uses a standard compliance for learning technologies—the IMS Global Caliper standard. In fact, any taxonomy that meets the Caliper standard can connect with RAD.

Adjustments to skill predictions at lower levels are propagated up the hierarchy so that RAD can also yield probabilistic predictions at higher levels.

What is the Dashboard component of RAD?

Our goal is to provide educators, administrators and customer support reps with insight into the learning assessment system, so we developed a dashboard that includes a web-facing interface. You now have a detailed view into what’s happening inside the software service so you can see what RAD is recommending and the diagnostics.

Using the dashboard, educators will be able to check learner progress in any subject. Over time, with longitudinal tracking, a teacher can see the progress of one student, their entire classroom, a whole school, or district.

While the RAD enables students to work at their own pace and in a self-directed way, receiving personalized feedback and help, the dashboard allows teachers and administrators access to the “big picture.”

What else makes RAD different than other learning software?
RAD is a modular, separable component, not hard wired into any particular platform. It can be plugged into different learning assessment systems, providing widespread application possibilities, for various customers.

Download the fact-sheet for more information about RAD capabilities and technical specifications.