Incorporating Adaptive Learning to improve learner retention

by Evnisha Malani
on December 24, 2018

Recent advancements in the field of artificial intelligence has made Adaptive learning a more feasible option than ever. It deals with the shortcomings of online learning such as learner engagement, efficiency, retention, and linear content delivery just to name a few.

Adaptive tech utilizes neuroscience, learning theory, artificial intelligence and learner-centric instructional design to make learning that sticks for your learners.

You can find a decent number of companies that are either making adaptive content or LMSs equipped to deliver them. The underlying principle behind it is same but the level of adaptability varies.

Here’s what goes in an adaptive system:

Low adaptability

 

It’s the level of adaptability you’ll find in most of the MOOCs (Massive open online courses). Low level of adaptability relies on human efforts and needs revisions to make or deliver relevant content to your learners.

  • Learner is assessed through a test and then assigned a learning path according to the analyses of their results.
  • Learner accesses learning content from within the system or other sources and takes a test. Based on initial assessment, relevant learning content and an adapted test is assigned.
  • At this level of adaptability the learner accesses content from within the adaptive system and takes a test. A learning path is defined using analysis and revised content is then assigned to the learner. After taking the revised course, learner is then re-tested on prior weaknesses.
Highly adaptive

 

A highly adaptive system is an ideal one. It exploits AI to identify skill gaps of each learner individually and assigns them specific learning paths with bespoke content.

Adaptability is directly proportional to learner data.

Talking about the data, it is the most prominent challenge in creating and delivering adaptive content.

Data collection tools

The data needed to make an adaptable learning system is collected with the help of adaptive tools. Some of the mentioned tools are exclusively designed to gather and analyze the learner data in order to make the adaptive sequencing better. While others double as an adaptive content authoring platform.

  • Smart sparrow
    Smart Sparrow is an adaptive authoring tool that provides adaptive content services as well. The Platform allows you to create intelligent courses based on your learner needs and Studio service can help with tailored content for your organization.
  • Knewton
    Knewton’s learning analytics deliver student insights so educators can understand which lessons stick, and which don’t. Knewton pulls additional data from course materials to understand which content is the most effective and shares how students interact with the materials so they can be updated and optimized.
  • Scootpad
    ScootPad adapts in real-time to learner needs with the right level of practice, instruction, remediation and assessment to ensure they master every concept.
  • Dreambox
    DreamBox is an adaptive learning software which focuses on improving mathematics for elementary and middle school students using adaptive system. It uses continuous formative assessment in and between lessons, to provide the right next lesson at the right time.

Processing the data for Adaptability
Adaptive tools collect data using real time analytics that help you determine key performance indicators(KPI).

Collecting the data

The data collection is classified in three key areas, Type of data, Difficulty level & item granularity, and Learner’s history.

Types of data

Academic performance, Learning process, Student interests. Other behavioral data can also be collected, such as social behavior (e.g. posting a comment on another student’s feed), ratings (e.g. whether you like an activity or not), or even mood (e.g. identifying how you’re feeling that day).

Difficulty Level and Item Granularity

A problem’s difficulty is determined by using evaluation models such as Kirkpatrick, Webb’s Depth of Knowledge and Bloom’s Taxonomy.

Granularity means the level of detail at which a concept or skill is captured.

The most common categories of data include:
• The general standard or topic.
• The specific concept.
• The discrete knowledge or skill.
• The cognitive difficulty level.

Learner’s History

If the tool does remember how the student has previously interacted with the content, then this information is added to the data pool and considered during the process of changing a student’s path.

Analysing the data


An adaptive system analyzes a learner’s profile for weak areas and offers them remediation by providing adapted content that is relevant to them.

Performance data, learner history, and behavioral data is all too much and too complex for humans to analyze in real time, hence it’s done with the help of algorithms. The most common ways by which algorithmic analysis is done are –

  • Weighting categories of data: weight the number of times a student submitted the correct answer as a higher priority than how much time the student spent on a set of questions.
  • Applying thresholds of mastery: apply a rule such as 80% mastery to determine when a student has met expectations.
  • Comparing groups of students data: compare one student’s learning profile to another similar student’s profile.
  • Calculating probability of mastery: calculates how likely it is that a student has mastered a skill.
  • Applying rules for correct and incorrect responses: send students to the activity that’s already aligned to the correct or the incorrect response.
Skill Selection

Skill selection is done on the basis of learning history. Adaptive tools determine learning paths and skills based on a learner’s response.

If the tool has only one option then the learner is moved to a skill which was aligned to a previous response. While in case of a tool with multiple options and extensive learner history, skill assignment is done on the basis of similar learner profiles.

Adjusting learning via analysis

After the tools have collected data, the system adjusts the content accordingly and provides the students with new required assignments. Content adjustment is done on the basis of three parameters i.e delivery, amount, and design.

As a result of this whole process the learners get tailored content based on their weaknesses and the effective learning time is reduced significantly. This time reduction in learning or training can amplify business benefits for your organization.

If your organization has specific learning needs then adaptive learning is the best bet for 2019. As Bear Grylls puts it “Improvise, Adapt, Survive”. Ask us how.

P.S. Can we send you an email?

Once a week or so we send an email with our best content. We never bug you; we just send you our latest piece of content: