Troubleshooting Thursdays: Adaptive Learning, Part 5: Gamification (Tip 105)

Troubleshooting Thursdays: Adaptive Learning, Part 5: Gamification (Tip 105)

A resource for safe and effective troubleshooting from the leaders in simulation training.

Hello, Troubleshooters! Welcome back once again to TST. 

We’re forging ahead with our series on adaptive learning (AL). If you find yourself with a little extra time on your hands these days, you can go back and check out earlier posts:

Part 1:     Adaptive Learning Overview

Part 2:     Personalized Learning 101

Part 3:     Why Adaptive Learning and Simulation are a Winning Combination

Part 4a:     The Future of Learning

Part 4b:     The Future of Learning, continued

Today we’re looking at adaptive gamification—a growing trend in optimizing adaptive learning that will soon be making its way into training programs everywhere, so that manufacturers can also get more bang for their training buck.

Adaptive Learning, Meet Gamification

In this series, we’ve been talking about computer-based adaptive learning, and how sophisticated algorithms can detect learners’ unique abilities and needs, and then serve them up content that suits them best. Maybe they’ll vary the delivery method (e.g., video, audio) or maybe they’ll change up the lesson plan so that learners can skip ahead if they’ve already mastered the material, or repeat a lesson if they struggled the first time.

In the past, we’ve also delved into the “gamification” of learning. It’s well known that using a game format is valuable for e-learning, because it increases user engagement. It’s fun, so people want to continue playing. This is increasingly critical in the realm of business training, since Millennials now make up over 35% of the US labor force. (A fun fact about Millennials is that they spend more on voluntary donations to game creators than on game subscriptions. That’s how much they love games!) Gen Z are pretty much born gaming. So, education and training are going to be more successful for these groups if games are involved.

There are now lots of “serious games” out there, designed to teach real-world skills. One example is Underground, a game designed to help surgical students practice removing a gallbladder arthroscopically. It’s set on a fictional planet where the object is to rescue cute little aliens from a deep pit. Using game controllers, players end up using the same hand movements required in the actual operation, thereby bonding the techniques into muscle memory.

Much research has been done on adaptive learning in serious games, meaning that the content of the game is being adapted to the learner’s abilities, usually by increasing or decreasing the difficulty level.  

But that is still adaptive learning, not adaptive gamification. If adaptive learning and gamification met, fell in love, and had a baby, that baby would be adaptive gamification.

Adaptive Gamification vs. Adaptive Learning

A little background. Gamification of e-learning programs means taking a learning program and adding the kind of fun features you typically find in video or computer games: leaderboards, stars, badges, chimes, words of encouragement, etc.; features that video games use to keep players engaged. The theory is that these features will help learners stay interested and motivated—Millennials and Gen Z, yes, but other generations, too. Everyone likes a little pat on the back, right?

Well, it turns out that not everyone likes the same kind of pat on the back. Some people thrive on competition and want to see their name at the top of the leaderboard, or at least above the name of the guy in the next cubicle. But others could care less, or are even intimidated by competition. Some people love to connect socially with peers, and some don’t. Some people love to achieve a goal, and others simply remain unmoved by that thought. 

Adaptive gamification refers to adapting these kinds of rewards and challenges, that is, the gaming features rather than the lesson content, to the user’s learning style and emotional type. First, the algorithms must discover which features actually motivate and engage a particular user, and then they must adapt the game elements to appeal to that learner. The theory is that the more engaged learners are, the longer they will spend learning, and the lower the drop-out rate will be.

Does Adaptive Gamification Work?

But does it work? Apparently so! E-learning platforms can experience a high drop-out rate, but one study in this area (Hassan et al. 2019) found that adaptive gamification based on learning styles increased motivation by 25% and reduced the drop-out ratio by 26%. 

Other researchers (Lavoué et al., 2018) conducted an intriguing study in which 266 participants answered an online questionnaire (“the BrainHex questionnaire”) to find out what kind of player personality they were. This simple questionnaire asks participants about their likes and dislikes related to computer games, and then groups participants into one (or a blend) of seven categories: Seeker, Survivor, Daredevil, Mastermind, Conqueror, Socializer, or Achiever. (You can try it yourself and discover your own learning style here.)

The researchers began with an existing, ungamified e-learning platform (Project Voltaire) that teaches French spelling and grammar. Due to the repetitive nature of memorizing involved, the platform designer admitted that many learners drop out early on. To this platform, the researchers added five gamification features: 

  1. Stars (which learners could earn for mastering a grammar rule)
  2. Leaderboards (showing a user’s near neighbors rather than overall best scorer, known as a “within-reach” goal)
  3. Tips (a social connection feature where learners could offer advice to help other users)
  4. Walking landscape (players could see a depiction of themselves hiking in mountains, each correct answer earning them a step on the path to a destination)
  5. Timers (which encourage learners to repeat an exercise in a faster time than a previous attempt)

Researchers then divided the participants into three groups. The first group learned on a platform with features adapted to their learning style. The second group learned with gaming features deliberately mismatched to their style. The third was a control group with no gaming features at all.

The results? Group 1 (matched features) had much more highly engaged users, who spent an average of about 42 minutes longer per week in the learning environment than the mismatched and control groups. They also had higher levels of motivation than the other two groups.

In case you’re interested, features 2 and 5 (leaderboard and timer) were the most effective in increasing user participation in the adapted group, and features 3 and 4 (tips and walking) increased motivation the most.

So there you go, Troubleshooters! Adaptive gamification is one more way to optimize learning and squeeze the most value out of your training budget. Tune in again next week for our last post in this series, on the top adaptive learning tech companies.

Follow Troubleshooting Thursdays

Tune in to Troubleshooting Thursdays for reliable tips, general troubleshooting process, and industry insights. Stay up to date with Simutech Multimedia:

Have a subject you would like Troubleshooting Thursdays to cover? Send Simutech Multimedia an email at [email protected].

Looking to give simulation learning a try? Get started with our award-winning first solution here: Get Demo.

Get started with your demo

Troubleshooting Thursdays:   Adaptive Learning, Part 4b: The Future of Learning, continued (Tip 104)

Troubleshooting Thursdays: Adaptive Learning, Part 4b: The Future of Learning, continued (Tip 104)

A resource for safe and effective troubleshooting from the leaders in simulation training.

Welcome back, Troubleshooters! From all of us at Simutech, we hope you’re all well and weathering this coronavirus storm safely. 

If you’re a regular reader here, you’ll know that the goal of this blog is to help manufacturers stay on top of trends that will affect them, usually with a training and education focus, so they can reduce the downtime that’s such a drag on profitability, and stay ahead of the competition. We’ve recently been doing a series of posts on adaptive learning, and right now we’re concentrating on the future of learning, because it applies directly to employee training, an area where manufacturers can potentially make big gains.

The Future of Learning is the Future of Training

Training obviously has a lot of benefits. Properly trained staff have fewer accidents, which results in fewer OSHA penalties and lawsuits. They stay with the company longer because training makes them more valuable, enables them to take on more responsibility, increases their self-worth and their sense that the company values them, and is a goal along their desired career path. Training can have very tangible financial advantages for an organization; for example, well trained maintenance staff are worth their weight in gold—they can diagnose and repair electrical faults in production line equipment quickly and efficiently, and that’s huge for manufacturers whose downtime costs them in the thousands of dollars per minute. 

While the benefits of training are clear, the cost has always been the obstacle. That’s why new technologies and approaches that are lowering the cost of training are very good news for manufacturers. If you have the vision, you can see how in the not-so-distant future you will be able leverage these advances to train more employees faster, to the same quality standards, and for less money. 

One way to prepare for this is to understand how technology is making mass education possible. Last week we talked about the future of education being transformed by adaptive learning (AL) powered by artificial intelligence (AI). Today we’re looking at the ways that AI is improving, so that it is able to provide an ever more personalized learning experience along with the efficiencies of scale.

AL of the Future: Profiles and Learning Styles

Most of us are familiar with the notion of learning styles, that is, that some people are visual learners and others are aural, verbal, or active, etc. Some like written materials. There’s no shortage of research that confirms there are distinctly different learning styles. There is also a lot of research that confirms that matching (or mismatching) teaching and learning styles can have a significant effect on learning outcomes (e.g., Ford and Chen 2001, Gilakjani 2012).

In a fascinating study out of Pakistan, students from a number of universities studying in four different subject areas—physics, math, botany, and chemistry—were evaluated to determine their individual learning style on four dimensions of learning: sensing/intuition; active/reflective; visual/verbal; and sequential/global. Their instructors’ teaching styles were also evaluated on the same dimensions. 

The researchers found that overall, more students were mismatched with their instructors than were ideally matched. This is because more students prefer visual, sensing and active learning styles, while more teachers teach with a passive lecture style. More importantly, the researchers found that students with matched learning styles significantly outperformed those who were mismatched (mean scores of matched students were 76% while mismatched were 70%). The study also concluded that while visual learning is most dominant, active, global, sequential, reflective, concrete and verbal learning styles all account for an equal proportion of students. 

What all of this shows is that learning characteristics are different for each person and that people respond better to different kinds of learning resources. In addition, not only learning styles, but learners’ goals, abilities, and background knowledge also differ. Obviously, to optimize the AI-based learning of the future, course designers need to take students’ individual profiles into account. And that brings us to…

Profiling (But in a Good Way)

Just like the way your streaming video service creates a profile for you, so it can serve up better viewing suggestions for you, the AL- and AI-based learning systems of the future will create a profiles of users in order to adapt content and content delivery to them.

And they won’t even know it’s happening! Learning programs will collect data in real time based on students’ answers and behaviors, using what they call “stealth assessment” (happening beneath the radar). They will “get to know” each learner and personalize a program of content delivery specific to them. 

In the future, this is likely to happen at a higher level than e-learning programs. A platform-wide AI network will collect data from a much wider set of activities and even from separate e-learning systems, shared to create a deeply individual profile, and giving the AI a much richer data source from which to infer more subtle learning influencers, such as psychological or emotional state. 

Also part of the profile is learners’ ability to adapt to computer-based learning, which may depend on things like age or cultural background. Lee (2001) identified four types of learner adaptation: model learners, disenchanted, maladaptors, and fanatics, all of which have differing levels of ability and satisfaction that affects their learning. Increasingly sophisticated AI will have to take these abilities into consideration as well.

A Golden Opportunity for Continuous Improvement

The key takeaway here for manufacturers is that there are many opportunities for continuous improvement in the area of training, and that computer-based software, including simulation software, that integrates adaptive learning and AI will continue to improve the learning experience as well as student performance. And that’s an efficiency no manufacturer can afford to overlook.

Well, that’s it for today, Troubleshooters! Tune in again next week as we continue exploring AL, AI, and the latest trend that’s certain to have a role in the future of learning—gamification.

Follow Troubleshooting Thursdays

Tune in to Troubleshooting Thursdays for reliable tips, general troubleshooting process, and industry insights. Stay up to date with Simutech Multimedia:

Have a subject you would like Troubleshooting Thursdays to cover? Send Simutech Multimedia an email at [email protected].

Looking to give simulation learning a try? Get started with our award-winning first solution here: Get Demo.

Get started with your demo

Troubleshooting Thursdays: Adaptive Learning, Part 4a: The Future of Learning (Tip 103)

Troubleshooting Thursdays: Adaptive Learning, Part 4a: The Future of Learning (Tip 103)

A resource for safe and effective troubleshooting from the leaders in simulation training.

Hey Troubleshooters! Thanks for tuning in to TST again this week. As you may know, we’ve been writing about adaptive learning (AL) lately. Today, in Part 4 of our series, we’re looking at the future of learning. 

(We should fill you in on the first three posts. To explain—no, there is too much—to sum up, Part 1 was an overview of adaptive learning, Part 2 was about personalized learning, and Part 3 zeroed in on why AL and simulation training dovetail so well.)

AI + AL = The Future of Learning

Provided the world doesn’t end in a zombie apocalypse, it seems clear that adaptive learning is the future of education. But not just adaptive learning: adaptive learning powered by artificial intelligence (AI). 

Adaptive Learning = Efficient Learning

In almost every aspect of modern life, we are constantly striving for increased efficiency. Adaptive learning is more efficient than traditional one-size-fits-all classroom methods because it aims to personalize the learning path for each student, according to their unique strengths and needs. Adaptive learning methods can be used with human instructors, or with automated e-learning programs. And it’s with automated AL that we find the greatest potential for gaining efficiency.

Computer-based AL can, for instance, deliver extra content on topics that are giving a particular student trouble. Or, it can figure out that another student already knows the material, and skip them onto the next lesson. Right out of the gate you gain efficiencies 1) from letting the students who have mastered content proceed “unshackled” and thereby shortening their training time, and 2) by not sending other students on to a new module until they have mastered the content in the preceding one, thereby ensuring they’re not wasting their time attempting to build on concepts they don’t yet understand.

But most importantly, because all of this is done via software and is not limited by the number of human instructors or the budgets to pay them, it can yield tremendous efficiencies through scalability. 

Obviously, at this point in time, a one-on-one human instructor-student relationship is still more intuitive and nuanced than even the best AL programs can claim to be. Human teachers and tutors have a wealth of experience and knowledge. They have deep knowledge of the subject matter and have the advantage over computers by, say, being able to interpret body language that demonstrates boredom or frustration. They can instantly see why a learner is having trouble with a concept, and what the misconception is.

But, while this ideal of individual tutoring and apprenticing is available to some, it is extremely expensive. The future of learning, not just for schoolchildren but in other applications such as employee training, is computer-based adaptive learning driven by artificial intelligence that can more closely approximate the guidance and instinct of a human instructor, but distribute it to vast numbers of learners.

Enter Artificial Intelligence

Earlier we gave the example of AL software sequencing content to suit learner’s abilities. But there’s much more to it than that. For instance, education research has shown that emotional states such as engagement, frustration, impatience, or discouragement can radically impact the quality of learning. So, in order to optimize student success (which is efficient), the AL software has to be attuned as much as possible to the individual learner’s needs so they remain engaged and not frustrated or discouraged. This can mean custom-tailoring not only the sequence of the course content for optimal engagement, but even adapting assessment in real time in order to avoid discouraging learners. There are a lot of moving parts.

And this is where AI and machine learning come in. As the field of computer-based AL evolves, it will depend more and more on complex AI algorithms fed by big data. The more learners that go through the course, the more data that is generated for the algorithms, and the more the system “learns” how to better personalize content for optimal individualized learning, and the more the content maps are refined. It all results in a continuous improvement cycle, whereby the AI-powered AL becomes increasingly more human. (Not as in becoming sentient and taking over the world, just becoming more natural.)

China’s Futuristic Learning System

As it happens, the learning system of the future may already exist. 

In China, where academics are prized, competition is stiff, and parents will make huge sacrifices so their children can get ahead, there is a market for individualized learning tools. One such tool is the Yixue Squirrel AI Learning, an AI-powered adaptive learning system that tutors children in a variety of subjects in the Chinese curriculum. Over two million students are signed on. They learn either online from home, or attend one of over a thousand Squirrel AI learning centers across China, which children in remote and rural areas to access the same quality of education as urban children. 

Squirrel AI uses a three-pronged strategy of diagnosis, content, and assessment. It begins with an assessment to discover a student’s weaknesses and strengths, and then only teaches the material that the student doesn’t know. The diagnosis is extremely granular, resulting in highly individualized courses. It If a student makes mistakes, it tracks them back to where they originally went wrong, works on that material with them, then sends them back to where they left off. Because it is only covering material that the learner doesn’t know, it Yixue claims that it actually reduces student screen time and gives them more leisure time. Students work at their own pace, and Squirrel is constantly collecting data to improve the content, delivery, and assessment.

Learning of the future will be much like this, say experts. As with the Squirrel AI system, human teachers will continue to be required for content prep. Some experts argue that in the future, AL will be “open-ended and assessment-based” because of the difficulty of developing “very deep” AL content for certain subjects.

Well, that’s it for today, Troubleshooters! Tune in again next week as we continue exploring the future of learning. We’re going to be talking about how the learning system of the future will seek to discover and cater to student’s individual learning styles and preferred learning resources, and use those to deliver relevant content. In the next few weeks, we’ll also be looking at another feature that’s certain to have a role in the future of learning—gamification.

Follow Troubleshooting Thursdays

Tune in to Troubleshooting Thursdays for reliable tips, general troubleshooting process, and industry insights. Stay up to date with Simutech Multimedia:

Have a subject you would like Troubleshooting Thursdays to cover? Send Simutech Multimedia an email at [email protected].

Looking to give simulation learning a try? Get started with our award-winning first solution here: Get Demo.

Get started with your demo

Troubleshooting Thursdays: Adaptive Learning, Part 3: Why Adaptive Learning and Simulation are a Winning Combination (Tip 102)

Troubleshooting Thursdays: Adaptive Learning, Part 3: Why Adaptive Learning and Simulation are a Winning Combination (Tip 102)

A resource for safe and effective troubleshooting from the leaders in simulation training.

Greetings, Troubleshooters! Welcome back to TST. We’re currently in the middle of a series of posts on adaptive learning. In case you missed it, Part 1 was an overview of adaptive learning (AL), and Part 2 was on personalized learning 101. Today we’re talking about why adaptive learning and computer simulation training go hand in hand.

Okay, let’s dig in!

Data—AL Needs It, Simulation Feeds It

The major reason adaptive learning and computer-simulation-based learning are such a natural fit is data. One the one hand, you have digital, simulation-based learning software, churning out all manner of data about learners—tracking their progress through the course content, their performance during assessments, the time it takes them to complete a simulated scenario, etc. And on the other hand, there’s AL, just itching for data to feed its algorithms. Talk about a symbiotic relationship!

In order to be truly useful, AL needs data. AL collects data from large numbers of learners to see how  they learn the specific material, and then in turn itself learns from that data (machine learning) so that it can become more intelligent over time with respect to the content sequencing, learning objectives, and learning style, and provide sophisticated, real-time adaptation of course material based on learner performance and behavior.

Admittedly, very basic adaptive learning systems that don’t use machine learning to analyze user data can still guide learners through a custom path based on their performance and behavior. They could in fact do that with just one user and no knowledge of what other users are doing. But in order to create a more personalized and better-informed overall experience for the learner, data from large numbers of users for the machine learning (adjustment of the algorithms) is required, providing higher-level analysis that benefits from seeing the bigger picture.

Just as an example, suppose you subscribe to a movie streaming service and one night you watch Avengers: Infinity War. The next night when you log in, the service suggests you might also like Thor, Ant-Man, Iron Man 2, Justice League, Black Panther, Captain America, and … Gone with the Wind. Now there’s an algorithm that needs a little more data! If that service had more viewers watching those movies, it would know that (almost) no one who’s watching superhero movies is also watching epic historical romances. More data gives AL the ability to more optimally personalize the learning path.

AL Refines Simulation-Based Learning

But it’s not just AL that needs simulation’s data. Simulation learning systems benefit from integrating adaptive learning capabilities. With access to the data generated by computer-based simulation e-learning software, AL can detect, for example, significant spikes in course failings and alert the instructor that something is wrong pedagogically. It can also see if 90% of learners are getting a certain question wrong, and alert content designer that it has to be reworded. It can alert course designers if students are being routed past a particular lesson because they already know the material, but then are failing the assessment. AL can give valuable feedback to the content designers so that they can refine the simulation to address these (and far less obvious) issues, so that the simulation-based education or training system using AL is improving, rapidly and continually.

Why simulation rather than just e-learning?

At this point, you may be asking yourself, why is simulation-based learning in particular a better fit with AL than other forms of e-learning? The answer is simple: simulation learning allows generates more detailed data than just online reading and testing. With simulation-based learning, you can get data about a wider range of activities that give better clues to help the AL diagnose what stage the learner is at and what they might be missing. For example, a strictly online learning program in which students read lessons and watch videos and then answer multiple-choice questions can tell the AL program a certain amount about what the student doesn’t understand. 

But imagine a training simulation for electrical troubleshooting that can tell when a learner virtually “walks” into a simulated greenhouse and stands in a puddle of water while repairing a circuit. Or forgets to turn off the power before working on a fault. That kind of richer, deeper data can be collected from a simulation and used to help an AL system determine that the learner needs a refresher on basic safety before they actually are deployed on the factory floor.  

When training for specific tasks, using a simulation can reveal important details such as how long it actually takes the student to perform that task, and whether they need more practice. It can keep track of how many attempts the learner makes to solve a problem before deciding on an answer. It can make more accurate assumptions about when a learner is guessing, or solving a problem by trial and error rather than by knowledge.

Powered by Sound Pedagogy

Pedagogy is the method and practice of teaching, i.e., how you teach what you teach, so that students can understand it and master it most efficiently. Sound underlying pedagogy is absolutely critical to the success of both simulation e-learning and adaptive learning.

For example, who decides exactly what percentage of correct answers on the assessment of one module means that a student can skip the next one? Who anticipates all of the potential actions a learner might perform in a non-multiple-choice task in a simulation? Who decides what the student’s misconception is based on wrong multiple-choice answers, and what the remedy is? Who writes the hints and other feedback that guide the learners to rethink wrong answers and come up with the right ones? All of these issues and more must be thoroughly anticipated and tested by subject matter experts (SMEs) who have actually taught the material, in order to be sure the algorithms are working as intended. The nice thing is, AL can also keep track of how successful in general learners are with a certain course structure, and monitor how success rates change after tweaks are made to the content. 

All of this is the hard and expensive part. By comparison, the technology is easy!

Purchasing Simulation Software with Integrated AL

Because it’s a big, expensive job, not all manufacturers can afford to create their own simulation-based, AL integrated training programs. Nor is it necessary. If you’re training for generic skills and not on a proprietary system, you can avoid reinventing the wheel and purchase existing training software. Just be sure to research the quality of the program or system first. Not all are created equal, and it really is a question of the vendor’s experience in the subject matter, in sound pedagogical theory, and in the technology.

And that’s it, Troubleshooters! Tune in next week when we pull out our crystal ball and reveal … the future of learning!

Follow Troubleshooting Thursdays

Tune in to Troubleshooting Thursdays for reliable tips, general troubleshooting process, and industry insights. Stay up to date with Simutech Multimedia:

Have a subject you would like Troubleshooting Thursdays to cover? Send Simutech Multimedia an email at [email protected].

Looking to give simulation learning a try? Get started with our award-winning first solution here: Get Demo.

Get started with your demo

Troubleshooting Thursdays: Adaptive Learning, Part 2: Personalized Learning 101 (Tip 101)

Troubleshooting Thursdays: Adaptive Learning, Part 2: Personalized Learning 101 (Tip 101)

A resource for safe and effective troubleshooting from the leaders in simulation training.

Welcome back, Troubleshooters. It’s that day of the week again, and time for TST. Last week we began a series on Adaptive Learning (AL) with a general overview. Today in Part 2, we’re going into a bit more depth—welcome to Personalized Learning 101! 

Let’s Get Personal

On the face of it, probably nothing seems less personal than big data. And yet, big data is a big part of adaptive (or personalized) learning.

As an educational approach, personalized learning is all about catering to the distinct learning needs, skills, strengths, and interests of individual learners. The goal is to facilitate the success of each individual learner by customizing content and content delivery to best suit them. It involves tailoring accommodations, supports, and accessible learning strategies to each student.

In schools, where most work has been done in this field, it can mean restructuring the entire school into smaller learning communities such as teams or programs; replacing academic tracks with differentiated learning (delivering flexibly designed course content to groups of students with similar learning styles and needs); offering a variety of learning pathways, creating personal learning plans for each student, and so on. There is no one method set in stone for personalization of learning.

Technological advances have now made it possible to deliver personalized learning on a mass scale. Instead of one tutor or mentor customizing the curriculum for one student, adaptive learning tools are making it possible to use software powered by complex algorithms to optimize learning paths for individual learners on a mass scale. This ambitious goal requires a lot of preparation up front in terms of detailed curriculum mapping and content development  (e.g., all student responses must be anticipated in advance and clear, unambiguous feedback prepared for each eventuality.) It also requires big data to fuel and refine the algorithms. The more data the algorithms have to work with, the better they approximate the instincts of a human instructor.

Personalized Learning Will Impact Your Training

More and more, organizations are recognizing that the same concepts can be applied to employee training, for the purpose of getting similarly improved training results that are scalable and efficient. For employee training, organizations are increasingly turning to online training, often with a widely diverse employee base, often with different cultural experiences, geographic locations, proficiency levels and multiple generations (ages).

Learners themselves are asking for content they can pull when they need it, rather than having it pushed to them when they don’t

Follow Troubleshooting Thursdays

Tune in to Troubleshooting Thursdays for reliable tips, general troubleshooting process, and industry insights. Stay up to date with Simutech Multimedia:

Have a subject you would like Troubleshooting Thursdays to cover? Send Simutech Multimedia an email at [email protected].

Looking to give simulation learning a try? Get started with our award-winning first solution here: Get Demo.

Get started with your demo

Troubleshooting Thursdays: Adaptive Learning, Part 1: Overview (Tip 100)

Troubleshooting Thursdays: Adaptive Learning, Part 1: Overview (Tip 100)

A resource for safe and effective troubleshooting from the leaders in simulation training.

Greetings, Troubleshooters! Thanks for tuning in to TST today. 

Today is a milestone for us—this is our 100th TST post! That’s 100 articles on topics of interest to manufacturing executives, HR personnel, Directors of Technical Training, and of course, electrical troubleshooters everywhere. You can find our searchable archive of Troubleshooting Thursday posts here, and look for tips on everything from calculating your training ROI to meeting ISO continuous improvement standards, to testing for opens with a voltmeter.  

Our goal has always been to keep manufacturers informed about the many benefits of high-quality electrical troubleshooting training. Today, we’re beginning a new series on adaptive learning, a new feature we’re integrating into our simulation-based electrical troubleshooting training system.

What is Adaptive Learning?

Adaptive Learning (AL) (sometimes called adaptive teaching or personalized learning) is an educational technique that uses computer algorithms to provide customized, optimal learning paths for individual learners. 

Every learner has unique abilities and needs, but all too often in the modern learning environment, education has become a “one-size-fits-all” solution in which the same content is delivered at the same time to a classroom of people of varying abilities, skills, and knowledge levels. Some are way ahead and get bored, while others are bewildered and left behind. AL seeks to turn that reality on its head. 

In an AL scenario, content is delivered to learners via computer software powered by algorithms that use data such as the learner’s responses to questions, tasks and experiences and their response times, etc., to generate real-time feedback such as hints or encouragement, or to decide what content the learner needs to see next. As opposed to one-size-fits-all teaching, AL helps the ones who are behind to catch up, and the ones who are ahead to advance “unshackled,” so that in the end everyone successfully reaches an acceptable proficiency level in the shortest amount of time.

AL technology relies on knowledge and best practices from several different fields: artificial intelligence, education, psychology, brain science, psychometrics (the science of objective psychological measurement), and predictive analytics. The algorithms are often complex, measuring not merely whether a learner’s responses are right or wrong, but how long they take to answer, and, if they are wrong, why they are wrong (are they misunderstanding a key point? are they just guessing?) Based on this data, the software can then determine what the source of the misconception or information gap is, and remedy it through targeted hints or additional lessons. In a nutshell, adaptive algorithms first determine what learners know, and then what they need to see next.

The EdSurge Framework: Three Components of Adaptive Learning

EdSurge is a California-based organization that specializes in education technologies. They have identified three sub-categories of AL: adaptive content, adaptive sequence, and adaptive assessment. 

Adaptive Content 

Adaptive content is real-time feedback that is instantly adapted to the learner’s performance. Usually, this feedback is more helpful than just whether an answer is right or wrong. It may take the form of remediation of an incorrect assumption, encouragement, or additional information or hint questions. Real-time feedback provides additional teaching, and encouragement helps keep users engaged, a critical part of successful learning. 

Adaptive Sequence

Adaptive sequencing refers to AL that continually analyzes student performance data to decide what content the student should see next. This could apply to the type of content, or the sequence in which content modules are presented. In real time, the algorithms create differentiated pathways for individual students by varying the sequence of content—for example, providing additional modules on concepts that a learner is finding difficult, or fast-tracking learners that have grasped the material successfully.

Instead of each learner experiencing every piece of available content, as they do in traditional classrooms, with AL, they only experience what they need to see at that point in time for optimal learning, much the same as happens with individual tutoring or apprenticeship.

Adaptive Assessment

Adaptive assessment refers to adapting the assessment material in real time, that is, altering the difficulty of a test question based on the response to a previous question. The difficulty increases if the learner gets it right, but remains easier if they are struggling. Testing can be a high-stress and discouraging aspect of education. Adaptively assessing learners can result in less discouragement and therefore greater willingness to engage with the material and continue on. Instructors are made aware of the varying assessment paths of each learner.

Adaptive Learning Is Scalable

Of course, the ideal mode of instruction is the one-on-one human tutor or mentor. In one-on-one instruction, a real live teacher brings to bear all the advantages of intuition sharpened by years of experience. The teacher works closely with the learner, observing their verbal and nonverbal cues when they’re having trouble understanding a concept, and adjusting their hints and feedback accordingly. In making those on-the-spot adjustments, that instructor is adapting to the student. It’s a beautiful relationship.

Unfortunately, this kind of attention can only be lavished on one or two students at a time. The real power of AL is that it is scalable. AL technology means that a hundred or even a thousand students at once can benefit from this same kind of individual attention and adjustment. And that’s great news for anyone who’s trying to train large groups in a short amount of time or with a limited budget.

Okay, that’s it for today, Troubleshooters. Please tune in again next week for Part 2, personalized learning 101!

Follow Troubleshooting Thursdays

Tune in to Troubleshooting Thursdays for reliable tips, general troubleshooting process, and industry insights. Stay up to date with Simutech Multimedia:

Have a subject you would like Troubleshooting Thursdays to cover? Send Simutech Multimedia an email at [email protected].

Looking to give simulation learning a try? Get started with our award-winning first solution here: Get Demo.

Get started with your demo