BACK

Adaptive Learning's Next Audience: Struggling K-12 Students

Published March 4, 2015


adaptive-learnings-next-audience-struggling-k-12-students

Report Highlights

  • Struggling K-12 students, particularly struggling elementary-level readers, present arguably the most significant challenge to U.S. schools today.
  • Personalized learning tools driven by adaptive technology, still relatively underutilized at the K-12 level, have in recent years been proven effective at the college level, both for remedial students and the general population.
  • Adaptive learning tools are uniquely well-suited for use by struggling K-12 students and should be used more widely because they directly address some of the most fundamental difficulties faced by teachers of struggling students, including time constraints and an inability to do manually what technology can do at scale: provide real-time, laser-focused diagnostic and predictive learning analysis that's customized for each student and personalizes the learning experience.

Introduction

Every year in the United States, more than 1.2 million students drop out of high school1.

It's the kind of statistic that has been repeated so often that, as tragic as it might be, it has begun to lose its impact. But if the statistic itself doesn't seem shocking, consider this: most of those 1.2 million high school dropouts were, at one point, a struggling elementary school student. And in almost every case, that student's fate was sealed by the time they reached the third grade.



Every day 3,300+ students drop out of high school

The third grade is the year during which certain basic skills, such as literacy, become fundamental to all learning, and any deficiencies quickly begin to snowball. Students who lack such fundamental skills not only miss out on higher level learning, but they're also more likely to act out, which can lead to additional issues that impact students' lives for years to come.

What can we be doing differently? How can we better reach those very young students before they become the high school dropouts of 2022?

The differences in parent-child interactions in low- versus high-income households cause children from high-income families to be exposed to some 30 million more words than children from families on welfare by the time children are three years old.

Struggling Students in the United States

American students are struggling. According to a 2014 research report from the Annie E. Casey Foundation , some 66 percent of U.S. fourth-graders read below their grade level – a number that jumps to 80 percent among fourth-graders from low-income families. The relationship between family income and student literacy is not a simple one, but it has been widely observed, and research has shed some light on several aspects of it. One of the most striking findings, first established in the 1990s and the subject of ongoing research even today, is that the differences in parent-child interactions in low- versus high-income households cause children from high-income families to be exposed to some 30 million more words than children from families on welfare by the time children are three years old.





66% of US 4th graders read below their grade level

Such discrepancies affect not only children's knowledge, but their skills development for years later, and it's not difficult to see how. Vocabulary acquisition and reading ability go hand-in-hand, and both are fundamental to virtually everything that a student will do throughout his or her academic career and beyond. And while early-grade math skills might be a similarly crucial foundation for a student's later STEM studies, a lack of such math skills is rarely as stigmatized as a lack of fundamental reading skills.

As widespread as such problems might be, every student's story is unique. Not every struggling student comes from a low-income background, nor does every struggling student act disruptively in order to mask learning difficulties. For some struggling students literacy might not even be at the core of the problem at all.

In a perfect world, every student, struggling or not, would receive special, personalized instruction based directly on his or her needs. Unfortunately, the 25-student classroom isn't likely to get much smaller anytime soon. So where might teachers be able to turn for support?

Adaptive learning in Higher Education: A snapshot of success - US college professors feel adaptive learning is more likely to have a positive impact on higher ed than MOOCs

While adaptive learning tools might be less familiar to K-12 educators, they've been around on college campuses for some time – and they've made good traction.

  1. The Bill and Malinda Gates Foundation, after an extensive review of eight adaptive learning leaders in 2013 , called adaptive learning "exactly what we need most right now," and announced its intention to award up to $1 million in grants for universities to further adaptive learning
  2. LearnSmart, one of McGraw-Hill Education's adaptive learning technologies, has been demonstrated to improve college students' grades by one full letter, with more B students getting As, and more C students getting Bs. In some courses, LearnSmart was seen to improve pass rates by more than 12 percent, retention rates by more than 10 percent
  3. ALEKS, an artificially intelligent assessment and learning system that uses adaptive questioning to quickly and accurately determine what a student knows and doesn't know in a course, has seen historical pass rates approaching 90 percent. ALEKS largely serves a struggling, or remedial, student population.
  4. Sixty-six percent of college and university presidents feel that adaptive learning is likely to have a positive impact on higher education, while only 42 percent feel similarly about massive open online courses (MOOCs), according to a 2013 survey by Inside Higher Ed and Gallup

Personalized Learning: An Overview

Adaptive learning tools are a relatively new technology based on an age-old pedagogical concept – namely, personalized instruction for each student.

It is well established that the closer teaching comes to a one-on-one interaction, the more effective it is. In the past, such personalized instruction was delivered to a relatively small number of lucky students – those who could consistently get their teacher's attention or whose families could afford a tutor. Personalization required truly excellent educators who had the time and flexibility to work one-on-one with students.

And unfortunately, even the greatest teachers have repeatedly found themselves bound by the constraints of time. A teacher can spend her full seven-hour school day helping students individually and still only spend an average of less than 17 minutes per day with each of her 25 students.

This simple math has caused classroom teachers, from the 19th century on, to "teach to the middle," delivering lessons that are academically appropriate (in terms of difficulty and pace) for most students but that generally leave advanced students bored and struggling students lost. Some struggling students might have been able to ask their teachers for help, and some advanced students might have earned their teachers' praise, but rarely would either of these groups receive any real personalized instruction to meaningfully address their abilities or needs.

Finding the right personalized learning path

Adaptive ed-tech can help students apply a deep understanding of student learning and behavior and use sophisticated data analysis to pinpoint and course-correct learning gaps.

Seeking to overcome such challenges, educators have repeatedly turned to the latest technologies, often with some success. The use of one-to-one computing in particular, while far from a cure-all, has been clearly demonstrated to yield significant improvements in certain struggling student populations.

Ultimately, however, highly effective instructional differentiation needs to be more nuanced and customized than traditional tracking and traditional ed-tech.

There are numerous reasons why a student might be struggling. Still, statistically speaking, there are several all-too-common corollaries for students who perform toward the low end of the academic spectrum. Generally, struggling students:

  1. Come from lower-income families. Lower-income students perform roughly 1.25 standard deviations below high-income students on standardized reading assessments – a gap that has grown more than 30 percent since the mid 1970s.
  2. Live in urban centers and attend resource-constrained schools. While challenges such as income disparity, ethnic and linguistic diversity, and high student mobility rates are hardly unique to urban centers, schools in major cities generally see a higher concentration of these challenges. Likewise, schools in lower-income neighborhoods generally lack access to the most advanced technology as well as a host of other resources that higher-income schools enjoy.
  3. Fall behind early. The third grade has been cited as the most pivotal year in a student's academic career. It's the year during which fluent reading becomes a prerequisite to all learning – and any students without that fluency begin to quickly fall behind.
  4. Are more likely to drop out of high school. A 2012 study found that students who do not read proficiently by third grade are four times more likely than proficient readers to join the 1.2 million students who drop out of high school each year.

Modern Adaptive Learning

Over the past decade, education technologists have applied the latest thinking in cognitive science, instructional design, and data structure and analytics to the education space, leading to the first generation of truly adaptive digital learning tools. By continually measuring students' knowledge (cognition) and confidence in that knowledge (metacognition) and using the most advanced algorithms, these data-driven technologies provide a unique learning pathway for each student. This adaptive technology enables teachers to finally achieve the kind of highly personalized instruction and assessment that they have always sought to deliver, and it is doing so in ways that are highly precise and efficient.

How do they work? At the most basic, adaptive learning tools work as digital learning guides as a student progresses through a lesson. As a student reads a chapter in his or her e-book, watches an educational video, or engages in a multimedia lesson, the adaptive program periodically asks the student to answer questions about the course material, and about the student's confidence in his or her answers. Within seconds, the program can zero in on precisely what a student does and doesn't know (and why he doesn't know it), and can provide additional help where needed: suggesting further reading, additional videos, or other instructional resources. At the core of adaptive technology is learning science, which provides the map of all the ways in which a student might come to know a certain concept or piece of information. Adaptive technology relies on learning science to "understand" a student's knowledge or skill level and determine the most effective, efficient path for that individual student to learn and master the concept in question.

Drilling down a level, adaptive learning algorithms can use a student's answers to home in even more closely on that student's personal learning preferences, state of mind while studying, and a variety of other learning metrics. The program can use this information to provide the teacher with an in-depth profile of each student's learning style, strengths, weaknesses and mastery, facilitating a level of student-teacher familiarity that might take months to achieve in a traditional classroom (often too late) and that would require hours and hours of manual data collection and analysis. Adaptive technology is the engine that powers the promise of data and analytics in education, where we have an opportunity to collect, analyze and act upon literally billions and billions of data points about student learning and behavior.


Benefits of Adaptive

When adaptive-powered personalized classrooms work well, it is hard to argue against their success. Classrooms today and in the future will need the following three characteristics to fully take advantage of the possibilities that personalized learning will bring:

  1. Real-time data analysis. As Karen Cator, president and CEO of Digital Promise and former director of the Office of Educational Technology of the U.S. Department of Education, said at the 2013 ASU GSV Summit, data is the "rocket fuel" powering personalization in education. Technology is giving us more data, more quickly. While student performance data has always been available to instructors through homework assignments and assessments, today’s technology collects data as students are learning – on a constant or near-constant basis – providing instant feedback on individual student performance to educators, enabling them to spot and correct problems sooner.
  2. Personalized instruction. In traditional classrooms, instructors often "teach to the middle," delivering a set curriculum that is generally designed for the average student. In the digital, personalized classroom, instructors have better awareness of what students know and where they are struggling, and can adapt their instruction accordingly. Being able to understand the variety of needs in a classroom and adapt the daily classroom lectures, activities and interactions to meet those needs is critically important, especially as instructors are increasingly working with students who have different learning abilities and backgrounds.
  3. Classroom management. In the personalized classroom, instructors are freed from paperwork and busywork – the more mundane tasks of the classroom that take away from actual teaching time. The time they gain can be spent giving students more individual attention and support.


It's in this way that adaptive learning tools go beyond merely tracking whether or not a student answers questions correctly. At their finest, adaptive tools can apply a deep understanding of student behavior and learning. Through the data mining process, learning patterns inevitably arise, allowing adaptive tools to use insights from the aggregate to become even more powerful and precise; the more students use an adaptive learning engine, the "smarter" it becomes.14

Adaptive learning tools can increase grades by one full letter, achieve course pass rates near 90%, and boost retention by 10%+

So far, adaptive technology has been used primarily in higher education. And in the years that adaptive tools have been available on U.S. college campuses, students and professors have seen results. A 2012 independent study of McGraw-Hill Education's adaptive learning solutions on college campuses found that students increased their grade by one full letter, with more B students getting As, and more C students getting Bs15.

Interestingly, one of the areas in which adaptive learning tools have seen marked success is in remedial college courses. This may be because, as professors have consistently reported, adaptive tools are ideal for guiding students through the foundational elements of a course, helping students to self-identify which concepts they might have failed to grasp initially, where they might look for self-help, and when professor intervention is necessary for more in-depth assistance.

The adaptive assignment achieves true engagement by continually monitoring a student's interest level, and changing course as soon as engagement drops off.

Across the board, adaptive tools help students come to class prepared, allowing professors to focus on higher level lessons rather than reviewing basics and struggling to keep the class on track – and that means results. In one recent case at Lone Star College, professors saw pass rates climb from 50 percent to 75 percent after introducing adaptive math program ALEKS in a summer 2013 mathematics course.16

Though adaptive tools are not yet as widespread in the K-12 space, the schools that have implemented them have seen similar results. At California's Big Bear Middle School, for example, Algebra Readiness eighth graders achieved more than 200 percent of their expected improvement on the California Standards Test (CST) after using ALEKS for one year.17 As adaptive moves further into K-12 schools, new trends may emerge that will allow the designers behind these adaptive tools to make them even-better tailored for specific student populations, but we already have more than enough information to get started.

A Growing Audience For Adaptive: Struggling Students

Let's turn back to our struggling students. Whether we define them as students in the bottom quartile or as students who are at least one grade level behind, let's at least agree that these are the students who are in the most dire need of help.

Perhaps the greatest tragedy in the struggling student landscape today is that, with a high degree of accuracy, we already know all too well who is going to fail. As an industry, we're remarkably good at identifying students who need help, and we have been for quite some time. The problem lies in the fact that the remediation that is generally prescribed to remedy students' learning deficits is inefficient.

Traditional methods require a teacher to spend dozens, sometimes hundreds of hours, helping each individual struggling student to get on track – time and teacher resources that virtually no school has in anything close to adequate supply.

Take, for example, studentswho struggle in reading. We already know that one of the areas in which beginning readers generally need most support is in understanding the fundamental phonology behind our language, involving sound and letter structure – what the general public tends to call "phonics." We know that five- to seven-year-old students with phonologic deficits are students who are at severe risk of reading failure. We can test those students in kindergarten and know that, unless they receive early, specialized intervention instruction, those students will fail their third grade standardized reading exams. We already know that those same students, when they reach the ninth grade, will be at risk of dropping out of school.

So what do we do with these struggling readers to help ensure that they don't become a dropout statistic?

Response strategies

In most cases, we pull them out of class to give them the targeted reading support that they need. This isn't necessarily a bad option – in recent years, intervention practices have become remarkably sophisticated, and even the more basic paper-and-pencil tools designed to support these methods have been seen to be highly effective when used properly.

One of the most widely successful formalized reading-intervention processes in use today is called Response to Intervention (RTI). RTI is an early detection, prevention and support protocol in which students are identified as belonging to one of three intervention-need categories, provided instruction as appropriate to that tier, and monitored continually to ensure ongoing progress. Beyond merely identifying what a student does not know, RTI is designed to help identify what a student does know, and to build a framework of further understanding directly upon that knowledge.

In the fully adaptive classroom, teachers remain the single most important driving force in driving student success.

Unfortunately, virtually all of the traditional intervention solutions designed to support RTI require intensive teacher or specialist guidance, and relatively few schools have the manpower to deliver these solutions en masse.

In a typical elementary school, a reading specialist might attempt to work with several struggling students at a time using book-based instructional tools. Invariably, and through no fault of the educator, the carefully calibrated instructional design begins to lose efficacy, stretched to accommodate a student-to-teacher ratio beyond what the designers intended. This can be particularly frustrating because of another fact that we already know about struggling student populations: that the integrity of instructional design is absolutely crucial to these students.

The bottom line is that teachers are still the most important driver of effective instruction for students. The difference is that adaptive tools allow teachers to fine-tune and further personalize their teaching, directing their attention where it is needed most and to whom it is needed most, instead of simply focusing toward "the middle." Adaptive learning doesn't replace teaching so much as it lessens the need for the teacher to be inherently involved in every aspect of each student's ongoing practice and study – a need that is simultaneously vital and difficult-to-deliver to struggling student populations.

To that point, we know that intensive, long-term one-to-one instruction is not scalable in most schools. That's where technology can lend a hand. In order to be effective in struggling learner populations, educational technology must mimic the experience of one-to-one instruction, and adaptive learning tools are the only educational technology that can deliver such an experience.

Going beyond ‘practice' with adaptive

Let's take a look at an example of how this might work in practice. A group of struggling fourth grade readers are pulled out of their regular classroom for specialized intervention instruction. Under an adaptive learning model, much as in the traditional intervention classroom, these students will, as a group, receive an introductory lesson by a reading specialist before being given an assignment to work on, individually or in groups, as the teacher works her way through the classroom to provide additional one-on-one lessons as needed. The difference lies in the fact that, in the adaptive classroom, the assignment that the students are working on mimics the teacher's one-on-one instruction.

In order to be effective in struggling learner populations, educational technology must mimic the experience of the one-to-one

The adaptive assignment is infinitely more engaging than the traditional one. It's not just superficially "engaging," working under the easy assumption that today's students are "digital natives" and therefore respond better to digital learning tools. The adaptive tool achieves true engagement by continually monitoring a student's interest level, and changing course as soon as engagement drops off.

Beyond maintaining engagement, adaptive tools mean that teaching and learning no longer need to stop once a student enters the "practice" phase of a lesson. In a traditional intervention classroom, once the teacher finishes her introductory lesson and directs students to work on a classroom assignment, students either understand what they're doing or they don't. The students who understand the assignment might complete it without a problem, but they likely won't learn anything new. The students who have a shakier grasp on the lesson might be stuck on the first problem of their assignment and, if the teacher is preoccupied with another student, it might be several long, wasted minutes before they receive the help that they need.

In the adaptive classroom, on the other hand, the practice assignment can immediately identify the student who needs help, and, if the teacher is busy with another student, can provide additional instruction on the fly. If a student still has trouble grasping a concept, the software might try an alternate teaching approach, much as a human teacher would. Time that would otherwise have been spent struggling is instead replaced by useful instruction – just as virtually every educator would want.

Admittedly, this instructional element, geared specifically toward struggling learners, might be the area in which today's adaptive tools still need to be refined, if only because struggling students often respond best to instructional styles that differ somewhat from what works for their peers. But the technology and the research are both there, and we cannot afford to let them go unutilized as students continue to fall through the cracks.

An Adaptive Future

The potential for adaptive technology to dramatically reshape the educational lives of struggling students is difficult to ignore. It's not about introducing additional technology for the sake of technology – it's about facilitating demonstrable results for students, closing performance gaps and increasing engagement.

Every hour – every minute – that an instructor has with a struggling student is crucial. It is entirely unacceptable that schools should be forced to simply give a group of struggling students a traditional practice tool, leave them largely to their own devices as they work through it, and call that "intervention." Until we can provide a human tutor for every struggling student, we should strive to provide each with a digitally adaptive tutor – and we can.

5 benefits of adaptive learning in the classroom: automated intervention, more 1:1 teaching time, continuous learning, increased student engagement, improved outcomes

Adaptive intervention can change lives, and it will. But if it's going to happen on any large scale in the near future, school leaders and digital education companies will need to work together.

It's no secret that an unsettling proportion of struggling students attend schools in underserved communities where any technology, much less the latest adaptive technology, is largely unavailable. And beyond the technology itself, there's the professional development that it requires, without which these valuable tools become little more than a burden to the educators on the front lines.

But the hard fact remains that, even in the most technologically underprivileged schools, adaptive tools are an infinitely more realistic option than a full-time instructor for every student. It's not just that adaptive learning is the best option for many struggling students – the harsh truth is that it might be their only viable option.

Adaptive technology's next big audience must be the audience that needs it the most: struggling students. Every 26 seconds in this country, we're reminded that traditional methods are simply not working. It's time to try something new.


1 Do Something. 11 Facts About High School Dropout Rates.
https://www.dosomething.org/facts/11-facts-about-high-school-dropout-rates

2 Annie E. Casey Foundation. (January 2014). Early Reading Proficiency in the United States.
http://www.aecf.org/~/media/Pubs/Initiatives/KIDS%20COUNT/E/EarlyReadingProficiency2014.pdf

3 Hart, B. & Risley, T.R. (1995). The Early Catastrophe: The 30 Million Word Gap by Age 3.
www.unitedwayracine.org/sites/unitedwayracine.oneeach.org/files/The%20Early%
20Catastrophe%2030%20Million%20Word%20Gap%20by%20Age%203.pdf

4 Newman, A., Bryant, G., Stokes, P., & Squeo, T. (2013). Learning To Adapt: Understanding The Adaptive Learning Supplier Landscape.
http://edgrowthadvisors.com/wp-content/uploads/2013/04/Learning-to-Adapt_Report_Supplier-Landscape_Education-Growth-Advisors_April-2013.pdf

5 Jarrett, J. & Rajan, R. (2013). Jumpstarting Adaptive Learning.
http://www.impatientoptimists.org/Posts/2013/03/Jumpstarting-Adaptive-Learning

6 McGraw-Hill Education. (2011). McGraw-Hill LearnSmart Effectiveness Study.
http://chronicle.com/items/biz/pdf/McGraw-Hill_LearnSmart-Effectiveness-Study.pdf

7 ALEKS Corporation. Math Success with ALEKS
http://media.wix.com/ugd/6ec213_535fc58ce87b4f48913c92172ecfe4fc.pdf

8 Jaschik, S. & Lederman, D. (2013). The 2013 Insider Higher Ed Survey of College and University Presidents.
http://insidehighered.com/news/survey/affirmative-action-innovation-and-financial-future-survey-presidents

9 Rabiner, D. L., Murray, D. W., Skinner, A. T. & Malone, P. S. (2010). A Randomized Trial of Two Promising Computer-Based Interventions for Students with Attention Difficulties.
http://connection.ebscohost.com/c/articles/47656465/randomized-trial-two-promising-computer-based-interventions-students-attention-difficulties

10 Reardon, S.F. (2013). The Widening Income Achievement Gap.
http://www.ascd.org/publications/educational-leadership/may13/vol70/num08/The-Widening-Income-Achievement-Gap.aspx

11 Ahram, R., Stembridge, A., Fergus, E., & Noguera, P. Framing Urban School Challenges: The Problems to Examine When Implementing Response to Intervention
http://www.rtinetwork.org/learn/diversity/urban-school-challenges

12 Hernandez, D.J. (2012). Double Jeopardy: How Third-Grade Reading Skills and Poverty Influence High School Graduation
http://gradelevelreading.net/wp-content/uploads/2012/01/Double-Jeopardy-Report-030812-for-web1.pdf

13 Hernandez, D.J. (2012). Double Jeopardy: How Third-Grade Reading Skills and Poverty Influence High School Graduation
http://gradelevelreading.net/wp-content/uploads/2012/01/Double-Jeopardy-Report-030812-for-web1.pdf

14 Knowledge Space Theory is one example of many adaptive technologies spanning humanities and sciences. For more information on how adaptive tools operate using a framework of related educational concepts, see the ALEKS Corporation’s discussion on Knowledge Space Theory:
http://www.aleks.com/about_aleks/knowledge_space_theory

15 McGraw-Hill Education. (2011). McGraw-Hill LearnSmart Effectiveness Study.
http://chronicle.com/items/biz/pdf/McGraw-Hill_LearnSmart-Effectiveness-Study.pdf

16 McGraw-Hill Education. (2014). ALEKS Case Studies and Implementation Strategies.

17 McGraw-Hill Education. ALEKS; Big Bear Middle School
http://www.aleks.com/k12/BBMS_Conquers_Math_with_ALEKS.pdf

BACK