White paper

Intelligent Adaptive Learning

AN ESSENTIAL ELEMENT FOR PERSONALIZED LEARNING

DREAMBOX.COM
 Intelligent adaptive learning is defined as
digital learning that immerses students
in modular learning environments where
every decision a student makes is captured,
considered in the context of sound learning
theory, and then used to guide the student’s
learning experiences, to adjust the student’s
path and pace within and between lessons,
and to provide formative and summative
data to the student’s teacher.
This type of learning tailors instruction
to each student’s unique needs, current
understandings, and interests, while
ensuring that all responses subscribe to
sound pedagogy.
Intelligent adaptive learning can play a
critical role in raising the achievement of all
students by meeting the individual learning
needs of each student in PreK–12 schools.

2 DREAMBOX.COM
 Executive Summary

Imagine having access to a highly qualified personal tutor for every student in your
classroom—on call, ready to tutor students anytime, anywhere. Intelligent adaptive
learning is that, and much, much more. It is a sophisticated, next generation system
that adapts learning to meet the needs of individual students, using the latest
research from the learning sciences.

3 DREAMBOX.COM
 Intended as a supplement to the classroom,
the intelligent adaptive learning system combines An intelligent adaptive learning system
a modularized curriculum, a continuous stream is designed to:
of data as the student engages in the learning
activities, and a cognitive model with which • Serve as a personal tutor to the student
feedback is provided to the student and next
• Adapt the sequencing of the curriculum
steps are determined. Every decision a student
and associated learning experiences
makes is captured, analyzed in the context of
sound learning theory, and then used to guide • Individualize the pace of learning
the student’s learning experiences to an ideal • Regulate cognitive load for
degree of challenge—ultimately aiming for the student
mastery. It adjusts the student’s pathway through
the modularized curriculum, modifies the pace, • Engage students in learning
and engages the student in next generation through gaming
instructional approaches that focus on deep
Each of these design elements is supported
understanding of the discipline, all while
through research. Together they serve as the
documenting the student’s progress, and
rationale to warrant serious review of this genre
providing formative and summative data to
of next generation learning tool by educators.a
the student’s teacher. (See Figure 1.)

Figure 1: A Model of an Intelligent Adaptive Learning System

Classroom Intelligent Adaptive Learning

Real-time data capture of student actions,

solutions, and explorations online

Modular curriculum*

Learning activities*
Database of
Teacher Student student data
Embedded, adaptive,
and continuous Continuous capture
assessment* and storing of data

Adapt sequencing,
Data on Intelligent feedback navigation, pace, pedagogy,
student progress to student and presentation Continuous data feed
Intelligent feedback to system

Cognitive model/
data analysis

*Designed pedagogically to engage students

4 DREAMBOX.COM
 Intelligent Adaptive Learning is Designed to:
1. SERVE AS A PERSONAL TUTOR TO THE STUDENT
Research indicates that tutoring provided by a highly qualified personal tutor is twice as effective as
classroom instruction.b While this is a well-known fact, for most public schools such tutoring has
been cost prohibitive.
An intelligent adaptive learning system is a next generation intelligent tutoring system that uses cognitive
modeling to adapt what is presented to the learner, when it is presented, and how it is presented in response
to the learner’s needs.
This real-time feedback loop is key to the effectiveness of the system as a tutor. When used strategically,
feedback can increase the average student’s scores by 27 percentile points (effect size of 0.79). The aspects
of feedback that had significant, positive effects on learning include: frequency of feedback, provision of
formative feedback specific to the targeted learning objectives, and questioning/learning prompts1 —
all of which are incorporated into the design of intelligent adaptive learning systems.

2. ADAPT THE SEQUENCING OF THE CURRICULUM AND ASSOCIATED

LEARNING EXPERIENCES
The sequencing of curriculum units and learning activities within units significantly influences the depth
and efficiency of learning for students. In an intelligent adaptive learning system, different students use
different paths. The system determines the options based on what a particular student is ready for and with
awareness of the range and diversity of learning experiences needed to ensure coherent connections and
deep understanding. The student is provided a degree of choice, but within parameters designed to ensure
a consistent progression of learning.2 Student choice is also an important element of these learning systems.
Studies indicate that students who are provided choice have higher levels of engagement, which in turn
correlates to more time on task and higher achievement.3
A key element of the individualized adaptive sequencing of curriculum is the importance of basing student
experiences on their prior knowledge.c This not only enables the system to identify and correct the
misconceptions a student might have, but it enables the system to present learning activities in ways that
connect to and build upon that student’s prior knowledge and interests. In addition, this adaptive sequencing
of the curriculum enables cycling through the entire program of study to ensure eventual mastery.

3. INDIVIDUALIZE THE PACE OF LEARNING

This national trend toward student-centered learning challenges the notion of seat-time versus competency-
based learning. It is clear from studies on tutoring that enabling students to work at their own pace to
achieve mastery—with appropriate feedback through the tutor—is an effective learning strategy. In fact,
researchers have found that mastery learning, where learning is held constant and seat-time varies, when
compared with the opposite (seat-time held constant and learning varies) results in significant increases in
student achievement.

4. REGULATE COGNITIVE LOAD FOR THE STUDENT

It is the responsibility of the teacher to adapt learning activities to ensure that students are making steady
academic progress toward targeted learning standards. The range of prior knowledge and skill levels that
students bring to the classroom is broad, but regardless of the students’ starting points, the learning process
should support the each student’s steady progress toward the learning standards.
The area between these two points is called the student’s “zone of optimized learning.” Because this zone
is different for every student, keeping each student within his zone is a challenge. If the tasks presented to
the student are too complex for their skill level, they may become frustrated. On the other hand, if the tasks
assigned are not sufficiently challenging for their skill level, they will become bored. Thus the teacher needs
to continually monitor the student’s learning experience to balance task complexity and skill level. Meanwhile,
the curriculum should also present learning activities that build on the student’s prior knowledge and interest
areas. When expertly orchestrated, the balance between complexity of task and skill level will be challenging,
yet comfortably paced by the student.

5 DREAMBOX.COM
 5. ENGAGE STUDENTS IN LEARNING THROUGH GAMING
Student engagement is a measure of a student’s investment in learning as defined by their perseverance
and willingness to exert effort necessary to comprehend complex ideas and master difficult skills.4 There
are a number of instructional and learning strategies that increase student engagement, including: a logical
sequencing of curriculum, novelty and variety, student choice, intellectual safety (i.e., system assures the
intellectual risks will not be ridiculed), affirmation of the work and progress, and clarity of goals.5,d
Interestingly, this list of strategies parallels the principles of gaming. It is ironic that students who struggle
to focus on learning often have no problem playing strategy games for hours. The research on learning
through gaming reports mixed results.
However, studies suggest that serious games that are effective in achieving positive results follow the five
basic principles that are also key design features of intelligent adaptive learning systems:
1. sequenced challenges
2. “just in time” and “on demand” information
3. performance before competence
4. motivation and attention
5. timely and specific feedbacke
Thus, some intelligent adaptive learning systems use gaming to engage students in assessments, learning
strategies, and learning activities that promote mastery.f
Intelligent adaptive learning is an extremely important asset in the pedagogical repertoire of the teacher/
student team. It is important that educators vet each intelligent adaptive learning resource to ensure
pedagogical alignment, and to plan carefully how intelligent adaptive learning—including the data on the
student’s progress—will be integrated into the larger learning environment.

Figure 2: Zone of Optimized Learning

Based partially on: Csikszentmihalyi (1991, 2000),
Vygotsky (1978), Murray and Arroyo (2002), and Arroyo (2003).

Learning
Frustration or
standards
confusion

The zone
of optimized
Task complexity

learning

w
Flo

Prior
knowledge
Boredom

Skill level

6 DREAMBOX.COM
 Optimizing Learning

Teaching matters. Different instructional strategies within the classroom get different
results. Today, although lecture is still a prominent instructional strategy in K12 schools
in the United States, many teachers have begun to refocus their classrooms toward
more personalized, active, inquiry-based, collaborative, and project-oriented learning.
Results in such classrooms vary, in part due to the expertise and experience of the teacher with these
newer, more innovative instructional strategies. It is clear from research that the instruction and resultant
learning strategies matter. Emergent research indicates that students who are taught by a teacher who
is performing at the 75th percentile in terms of pedagogical excellence, will significantly outscore a
matched group of students taught by a second teacher who is at the 25th percentile. In fact, an average
student will score 14 percentile points higher in reading and 18 percentile points higher in math as a
result being assigned to the first teacher instead of the second.6 Make no mistake, an effective teacher—
and their pedagogy—does make a significant difference in students’ learning trajectories.7

7 DREAMBOX.COM
 Zone of Optimized Learning
To be effective, teachers must optimize learning for the due in part to the prior knowledge he brings to the
many students in their classrooms. This is a challenge situation. Schools often try to maintain a common
in today’s classroom given the variability among class pace, but all too often, it doesn’t sync with the
students in terms of language, prior knowledge, needs of all students. Therefore, to optimize learning
motivation, literacy, numeracy, social/emotional for each student, a teacher must continuously monitor,
maturity, and family support systems for learning. The guide, and adjust for individual and varying learning
Mathematics Common Core State Standards introduce trajectories. The challenge for the teacher is in
additional depth and complexity to this challenge designing the activities at a level that all students find
in their focus on deep understanding of concepts. challenging, but doable, given their wide range of prior
Hence, to meet these standards, students will not only knowledge and varying skill levels relative to
need to acquire facts, skills, and methods through the subject.
memorization, building automaticity, and following Consider this challenge from the perspective of the
algorithms (i.e., surface learning), they will also need to student. Each student has a zone of optimized learning
make sense of the subject area in context of the world that is grounded in the prior knowledge the student
around them and be able to transfer that knowledge to brings to learning, and culminates in the attainment of
new situations (i.e., deep learning). the targeted learning standard (see Figure 3).
Such deep learning calls for cognitive effort on the The teacher must build on the wide range of prior
part of the student, pedagogical skill on the part of knowledge that students bring to the classroom,
the teacher, and sound instructional design of learning scaffolding each through the curriculum toward the
resources. Teachers do often differentiate assignments achievement of the targeted learning standard. To do
for groups of students and provide choice within so requires that the teacher balance the complexity
assignments, which can lead to some degree of of the tasks assigned to students with the students’
personalization and individualization. Anyone who current skill requirements, keeping them in their zone
has been in a classroom knows that every student’s of optimized learning.g
experience in learning a new subject is unique,

Figure 3: Zone of Optimized Learning

Based partially on: Csikszentmihalyi (1991, 2000),
Vygotsky (1978), Murray and Arroyo (2002), and Arroyo (2003).

Learning
Frustration or
standards
confusion

The zone
of optimized
Task complexity

1 learning

2
w
Flo

Prior
knowledge
Boredom

Skill level

8 DREAMBOX.COM
 One student’s pathway through this zone of there to provide the scaffolding and support the
optimized learning8 is shown below in Figure 3. 9, 10 , 11 , 12 student needs when they need it. Moreover,
The dashed line running through the figure represents especially at younger ages, students are often not
that student’s learning trajectory—note that the sufficiently self-directed in their own learning to make
trajectory is not always within the zone. At Point 1, the necessary adjustments, either cognitively—to
the task complexity is too low and it doesn’t challenge manage their own learning—or emotionally— to stay
the student, resulting in his becoming bored and motivated and engaged. The daily challenge is to
demotivated. At Point 2, the complexity is too high ensure that each student stays within his optimized
for the skill level, resulting in the student becoming zone of learning.
confused or frustrated, and again, demotivated to Sound like an impossible job? Given that each
learn. Learning is optimized when the student stays student’s optimized zone is also constantly shifting,
within the zone of optimized learning by building on it is extremely difficult for a teacher with 20 to 30
the prior knowledge he brings to the situation, thus students to reassess and recalibrate each day,
making incremental progress toward the targeted to provide the learning activities, and associated
learning standard. scaffolds, prompts, feedback, explanations, and
One of the reasons for less-than-stellar academic guidance required for genuine personalization. That
performances by students is the lack of strategically type of affordable, real-time responsiveness requires
designed learning experiences in response to feedback loops and real-time data that
situations where students find themselves at the only technology can provide.
boundaries of their zones. Often, the teacher is not

The zone of optimized learning in a particular area of study is unique for each child. It represents the
range of learning activities that will enable the student to make incremental progress from his starting
point—defined by the student’s prior knowledge of the topic and his current skill levels—to the end point,
which is defined by the targeted learning standards. That endpoint includes both content and practice
standards (i.e., knowledge and skill targets).

9 DREAMBOX.COM
 Intelligent Adaptive Learning

Intelligent adaptive learning optimizes Inherent in the design of intelligent adaptive learning
learning by establishing a digital learning systems are five critical factors:
environment that keeps students in their 1. t he content in the form of lessons or activities
in which the learner engages in a sequence
optimized learning zone. It captures every unique to his needs
decision a student makes and adjusts the
2. the instructional strategies that teach and
student’s learning trajectory both within guide the learner
lessons and across lessons. The key attribute 3. m
 easurements of the affect of the student
of the intelligent adaptive learning system toward the learning
is not the immediate correction of every 4. mechanisms for measuring and understanding
student error, but rather that it attempts to what the student does or does not know
“identify the psychological cause of mistakes,” 5. a feedback mechanism whereby the data acquired
provides intelligent feedback and prompts about the learner informs the next round of content,
for reflection and rethinking by the student, instruction, and motivation the student encounters14
and “thereby lower(s) the probability that The first time the student uses the intelligent adaptive
such mistakes will occur again.”13 learning system, they take an adaptive assessment
that places them within the modularized curriculum.
Imagine a personal tutor who constantly checks From there, the student’s pathway through the
for understanding in real-time by analyzing large curriculum, their pace of learning, and the feedback
datasets of a student’s actions and interactions, they receive are responsive to the individual student’s
often comparing them to a knowledge base of known needs and experience. The student has intelligent
misconceptions or errors commonly committed by personalized support, which adapts the sequencing of
other students studying the same topic. This tutor the curriculum, the pace, pedagogy, and presentation of
provides multiple pathways to learning with real-time lessons to optimize their learning. All these adaptations
intelligent feedback and access to progress reports for are in response to their actions and reactions within
students, teachers, and parents. Intelligent adaptive the system. See Figure 4 for a visual description of the
learning systems often include feature sets that student’s experience, feedback loops, and cognitive
students find engaging. Examples include gaming, or modeling that guide the student’s intelligent adaptive
providing students a modicum of choice as to which learning experience.
activities they pursue—within set parameters of their
current level of expertise and their targeted goals.

10 DREAMBOX.COM
 Many teachers use intelligent adaptive learning to intelligent adaptive learning is based on advances in
provide students with individualized tutoring that artificial intelligence (AI) systems. These AI systems
guides the students in their learning of specific topics. were built to utilize a set of knowledge-based expert
These teachers have access to detailed reports on rules to mimic the behavior of experts. Intelligent
the progress students are making on the learning adaptive learning systems also use such symbolic
standards. representations to continuously monitor and guide the
The level of sophistication of today’s intelligent user experience in ways that optimize the learner’s
adaptive learning systems is far superior to similar experience.16 Inherent in the cognitive model is the
technologies of the past. Computer-assisted determination of the level of mastery required of the
instruction (CAI), adaptive hypermedia, intelligent student in order to move on to new curricular units;
tutors, computer-supported collaborative learning a pedagogical philosophy of how much scaffolding
(CSCL), and individualized learning systems (ILS) of (intelligent feedback) to provide to the student, and
the past typically lacked the speed and sophistication when to provide it; the logic in the sequencing of
of today’s intelligent adaptive learning systems.15 The curriculum units and lessons presented to the student;
processing speeds, advanced learner analytics, and and the process by which the system determines the
sheer magnitude of data collected and analyzed, student’s optimal learning style and subsequently
combined with the sophisticated models of cognitive presents the content to the student.
development, optimize the learner’s experience. The The very name of this learning resource—intelligent
current day intelligent adaptive learning systems adaptive learning system—implies that the system
readily adjust instruction, content, and motivation combines intelligence and adaptivity. In fact, both
based on a student’s current state of prior knowledge, are essential elements, and both are based in sound
expertise and skill levels, and readiness for educational research. The intelligence aspect is
progressively advanced complexities of learning. embodied in the cognitive modeling, which combined
Cognitive modeling is the lynchpin of the system. with the continuous, real-time collection and analysis
The genesis of the cognitive modeling that enables of user data, results in intelligent feedback to the user.

Figure 4: Classroom Environment Leveraging an Intelligent Adaptive Learning System

Classroom Intelligent Adaptive Learning

Real-time data capture of student actions,

solutions, and explorations online

Modular curriculum*

Learning activities*
Database of
Teacher Student student data
Embedded, adaptive,
and continuous Continuous capture
assessment* and storing of data

Adapt sequencing,
Data on Intelligent feedback navigation, pace, pedagogy,
student progress to student and presentation Continuous data feed
Intelligent feedback to system

Cognitive model/
data analysis

*Designed pedagogically to engage students

11 DREAMBOX.COM
 This continuous stream of intelligent feedback made available to the teacher. While teachers
is provided to the user to keep him on a positive will utilize the system in different ways, ultimately,
learning trajectory. The data stream is also used each teacher-student pair is responsible for
by the intelligent adaptive learning systems to maintaining a positive learning trajectory for the
determine whether the user is ready to move on student toward mastery of the targeted standards.
to new curricular units, and based on the user’s Teachers typically consider the intelligent
progress, it determines which curriculum units adaptive learning system as only one of multiple
would optimize his learning trajectory. A third learning strategies and resources that ensure this
way that these data are used is in the reports goal is reached.

The elements of intelligent adaptive learning that contribute to the “intelligence”

attribute include the following:
• Intelligent analysis of a student’s solutions. The system interacts with the student by analyzing the
data from the student’s actions as he solves problems, explores concepts, and makes decisions.
The system not only can tell what is wrong, but is also sufficiently intelligent to pinpoint where the
misconception or misunderstanding occurred that is causing the error.
• Interactive problem solving support. The solution analysis described in the previous paragraph
enables the system to provide extensive, detailed feedback and provides prompts to the
student that cause them to rethink their strategies and solutions, and ultimately correct their own
misunderstandings or mistakes. This “intelligent help” includes hints that prompt reflection on the
problem and its context when a student is stuck. The intelligent adaptive difference lies in viewing
these hints as an opportunity for critical thinking as opposed to simply “telling” a novice student what
the “next step” of an expert’s strategy would be. In this way, the system emulates the actions of an
effective tutor.
• Intelligent adaptive learning systems are built around a modularized curriculum that is individualized
for and by the student.

The elements of intelligent adaptive learning systems that contribute to the adaptive
learning with those curriculum units and lessons include the following:17, 18
• Curriculum sequencing. Based on the intelligent analysis of a student’s gains in knowledge
and understanding, the system sequences the student’s progression through the modularized
curriculum. This is done by providing the optimal planned sequence of curriculum units as the student
demonstrates readiness, and the customization of learning tasks with varied pedagogies within the
module, again based on student data.
• Multiple learning experiences. The intelligent adaptive learning system typically provides multiple
pedagogical approaches to teaching each concept. This includes a variety of learning experiences,
activities, and contexts. The research on deep learning indicates that these multiple experiences
are necessary in order to achieve the student’s deep versus surface understanding of the concept.
Encouraging and enabling deep learning requires that the tasks be meaningful, at an optimal level of
difficulty for the student, and contextualized in ways that enable students to build schemas so they
can make sense of the concept within the world around them.
• Customized presentation and pace. The system accumulates information about the student that is
then used to dynamically generate and present digital content to the student. Diagnostic, adaptive
assessments are embedded within each lesson to assess mastery in a fluid, transparent way that
doesn’t create anxiety for students. The system varies content, sequencing of content, and format as
it optimizes the experience for the learner, frequently offering variations in the way learning activities
are presented—which is often necessary to develop deep understanding. Typically, these learning
activities are filtered to align with learner preferences. However, the longer a student stays in a
module, the more alternates they may experience as they strive for mastery. As a student progresses
through the system their pace is determined by how quickly they demonstrate mastery of a concept,
thus pace varies across learners.

12 DREAMBOX.COM
 The 2 Sigma Challenge

Can we get the learning results attainable through tutoring,

but in a cost effective way?
The research has been clear for over a A study by Bloom and colleagues demonstrated that
quarter of a century, students who engage academic achievement is less about prior achievement
or aptitude, and more about the type of instruction
with a tutor in a one-to one situation the student receives. In Figure 5, the blue, yellow, and
significantly outperform students taught in orange lines (left to right) comprise the student scores
conventional classrooms using lectures, and for the three instructional approaches, respectively,
also in mastery learning.h The differences lecture, mastery learningi and individual tutoring.19
All are bell shaped curves, but the individual tutoring
are truly significant. A student scoring at the results in higher scores relative to the other two
50th percentile in a conventional classroom, instructional approaches.
if tutored full time, would score at the 98th Consider the score of the average student in each
percentile on the same material. This is type of learning; the average student’s score would
two standard deviations (sigmas) above be positioned on the chart at the peak of each curved
line. As the chart indicates, if the student were taught
the lecture method, thus the researchers through lecture, his achievement score would be
labeled it the 2 sigma challenge (i.e., how significantly lower than if he were engaged in mastery
to achieve these same achievement levels learning or individual tutoring.20 Over time, the 2 sigma
through group instruction). challenge has raised a number of important questions.
First and foremost, “What are the critical elements
of tutoring that contribute to the significantly better
results?” And second, “How can we achieve these same
results without the expense of one-to-one tutoring?”
(i.e., within the current time, budget, and teacher-
student ratio constraints of the existing system).

13 DREAMBOX.COM
 Regarding the first question, researchers have Interesting to note, the learner actions that did
identified critical elements of tutoring. They describe not produce increased achievement were those
a tutoring framework consisting of five steps: that the students initiated themselves, such as
1. tutor asks an initiating question self explanations. Overall, the study found that the
significant learning gains accomplished through
2. student responds the tutoring sessions are a result of the student-
3. t utor provides feedback as to the accuracy tutor interactions.
of the response Finally, after decades of research, there are some
4. tutor scaffolds to improve or elaborate the answers to the second question, “How can we
student’s answer in a series of exchanges achieve these same results without the expense
of one-to-one tutoring?” (i.e., within the current
5. tutor gauges student’s understanding21
time, budget, and teacher-student ratio constraints
Analyses of tutorial sessions looked at critical of the existing system). The question has caused
elements of tutoring that contributed to shallow researchers to question whether the 2 sigma effect
learning (acquisition of facts, processes, or methods), is accurate. Some researchers suggest that the
and to deep learning (sense-making and transfer), effects of human tutoring are more typically in the
both of which are important to attainment of learning range of +0.79, perhaps extending to 2 sigmas
standards.j The researchers reported: in special circumstances. These researchers
• It was the active learning by students that influenced explain that Bloom’s 2 sigma data were skewed
the level of learning more than the specific action upward due to the use of a combination of human
taken by the tutor. The more constructivist the student tutoring and mastery learning. Second, the use of
response (e.g., spontaneously self-explaining, asking computer-based tutors and adaptive testing,
questions, drawing, taking notes, summarizing), the and adaptive learning have evolved over the
deeper the learning. last 20 years.
• Explanations by tutors resulted in shallow rather than Given the power of today’s next-generation
deep learning. The researchers speculated that this adaptive learning, it is not surprising that
might be due to a shift of the tutor’s focus from the researchers looking into the 2 sigma challenge
student to the content, thus eliciting less active are finding that the results from computer-based
learning by the student. systems approximate those from human tutoring.
• The most significant increases in deep learning were
a result of the students’ self-reflective comprehension
monitoring (i.e., students continuously gauging their
understanding of the material).

Figure 5: Differences in Academic Achievement for Students Taught

through Lecture, Mastery Learning, and Individual Tutoring
Source: Bloom et al., 1984 as cited by Koller, D., Stanford University, at TedGlobal 2012.

Individual tutoring

Mastery learning

Lecture

Achievement score

14 DREAMBOX.COM
 The Research Behind
Intelligent Adaptive Learning

Educators have been seeking a cost-effective way to provide individualized tutoring for
decades. Technology now makes that feasible through intelligent adaptive learning.
Essentially, the intelligent adaptive learning system is an intelligent program that “listens to and observes” a
student’s interactions with the learning activities from curricular lessons; it then analyzes that data, interpreting
it through the lens of its cognitive model. This enables the system to provide intelligent feedback to the user, to
determine the sequencing of curricular units and lessons, to make decisions about the sequencing of activities
within a lesson, and to inform the educator through data reports on the student’s progress. The adaptivity is
accomplished as the system applies the cognitive modeling to interpret the user data pedagogically, using that
interpretation to adapt instantaneously within the lesson, adapt the sequencing of curriculum, filter and select
learning activities, customize the presentation, adjust the pace of learning, and adapt the navigation system.
The design elements of the intelligent adaptive learning have great potential for increasing the pace and depth
of student learning including:
1. t he effectiveness of support, through the continuous stream of intelligent, formative feedback to the user,
the system, and the teacher
2. the individualized, non-linear sequencing of curriculum and learning experiences
3. t he individualized pace of learning
4. the regulation of cognitive load
5. student engagement through gaming
The research basis for each is discussed on the following page.

15 DREAMBOX.COM
 1. The effectiveness of support through a significant increase in learning, or in some cases with
continuous stream of intelligent feedback actual decreases in learning. The one exception to
this was verbal praise, provided it is specific and
to the user, the system, and the teacher
sincere, with reference to accomplishments through
The reason tutoring works stems back to the strategic student effort.25
use of feedback. When feedback is specific to learning
Researchers have also found that if the feedback is
objectives, it has a powerful impact on learning.
to impact learning positively, it must be focused on
Feedback via computers has been found to be less
the specific task on which the student is working,
threatening than face-to-face situations.
hence the value of the realtime, intelligent feedback
That translates into increased attention paid to to the student. Furthermore, the feedback must be
learning on the part of the student.22 In 2010, Wiliam substantially more than simply an indication of a right or
argued that effective feedback can double the rate wrong answer, and go beyond a quantitative evaluation
of learning. In his discussions on effective forms of to qualitatively address how a student arrived at his
feedback, Wiliam identifies five categories provided answer. A meta-analysis on formative assessments
pedagogically: 1) clarifying learning intentions and found increasingly positive results in student outcomes
sharing criteria for success, 2) engineering effective as the feedback becomes more specific and more
classroom discussions, activities, and tasks that elicit closely associated with the specific activity in which
evidence of learning, 3) providing feedback that moves the student is engaged.26
learners forward, 4) activating students as instructional
In a recent investigation on the effectiveness of
resources for one another, and 5) activating students as
types of feedback, Timmerman and Kruepke found
the owners of their own learning.23 The feedback within
that explanations and remediation were much more
the intelligent adaptive learning system focuses on but
effective than simply providing a correct answer.
one of these five categories; category 4 is excluded
since intelligent adaptive learning systems are built for Nyquist classified formative assessment as strong,
individual users, with no collaboration or cooperation moderate, and weak (see Figure 7), supporting
among or between users. In 2010, meta-analyses by the contention that the type of formative assessment
researchers from McREL24 reported that, on average, provided matters. The strongest type of formative
the provision of such feedback increased student assessment occurred when students were “given
results by 28 percentile points for the average student. information about correct results, some explanation, and
Alternative forms of feedback in the form of praise specific activities to undertake in order to improve.”27
or extrinsic rewards were associated with little to no With the advent of data-driven decision-making tools
and the use of formative assessment increasing
numbers of educators are using data to drive
Figure 6: Design Elements of instruction. Unfortunately, the feedback that schools
Intelligent Adaptive Learning System offer learners fails to significantly impact learning due
to delays in the feedback loops, the lack of specificity,
Intelligent Adaptive Learning or the non-expert source. As an example of the latter,
Real-time data capture of student actions, one researcher found that, in elementary schools,
solutions, and explorations online 80 percent of the feedback to learners was from
other learners, and 80 percent of that feedback was
Modular curriculum* inaccurate.28 Overall, researchers have found that in
Learning activities* a typical classroom, students receive little if any daily
Database of
student data feedback. The use of the intelligent adaptive learning
Embedded, adaptive,
and continuous Continuous capture system can change that.
assessment* and storing of data
A critical factor in the provision of feedback is the
set of conditions under which the feedback is provided
Adapt sequencing,
navigation, pace, pedagogy,
to the learner. Feedback is more effective under low-
and presentation Continuous data feed threat, comfortable conditions, which allow the student
Intelligent feedback to system
to pay attention to the feedback.29 In his meta-analysis
of over 800 research studies, researcher Hattie writes,
“Students learn most easily in an environment in
Cognitive model/
Data on student progress data analysis which they can get and use feedback about what
Intelligent feedback to student
they don’t know without fearing negative reactions
*Designed pedagogically to engage students
from their peers or their teacher.”30 Intelligent adaptive
learning systems are designed to provide this type of
positive feedback.

16 DREAMBOX.COM
 Figure 7: The Percentile Increase in the Academic Performance of the Average
Student when Provided with Weak, Moderate, and Strong Formative Assessments

Strong formative assessments: Students are given

information about correct results, some explanation, 21%
and specific activities to underake in order to improve.

Moderate formative assessments: Students are given

information about correct results with some explanation, 15%
and some specific suggestions for improvement.

Weak formative assessments: Students are

given information about correct results 10%
with some explanation.
0% 10% 20% 30% 40% 50%

2. Individualized sequencing of the Studies indicate that students who are provided
curriculum and associated learning choice have higher levels of engagement, which
in turn correlates to more time on task and higher
experiences for each student
achievement.32 Take for example studies conducted
Intelligent adaptive learning systems are designed in elementary schools in New Zealand. Research cited
around a modularized curriculum. The sequencing in a seminal paper on engagement by Fredricks33
of curriculum units and learning activities within these reported those students who were offered choice as to
units significantly influences the depth and efficiency which tasks to perform and when and where to perform
of learning for students. Research from the literature them, “worked strategically and persisted longer in the
bases of mastery learning, student engagement, and face of difficulty” than did students in a comparison
motivation provide support for such adaptivity of the group.34 Furthermore, Fredricks reports that student
curriculum. engagement begins to decline as they transition to
Mastery learning is an early precursor of the intelligent middle school. That researcher offered an explanation
adaptive learning system. It also used a modularized in noting that as elementary students transitioned
curriculum, but in a rigidly prescribed order, requiring into middle school the classroom structures became
students to work within each curriculum unit or module much more rigid, essentially allowing less choice in
until they retained an established level of mastery the learning environment, which may contribute to or
before moving on to the next, again, in the prescribed accounted for the decline in cognitive engagement
sequence. This new generation of adaptive learning after elementary school.35 Research has also found
not only allows a flexible pace through the curriculum; that the impact of feedback is optimized when there
it customizes each student’s pathway through are appropriately challenging tasks, and students
the lessons. While some sequencing of lessons is are intrinsically motivated to complete such tasks.36
prescribed, in cases where the one is a prerequisite to Student motivation stems from a number of different
others, student pathways are generally determined by perceptions and experiences. Students are more
student readiness and choice. The path is determined motivated to learn when they: 1) feel capable and
in part by the system, based on what a student is ready competent to do what is expected of them, 2) perceive
for, in part to provide a range and diversity of learning stable links between actions and achievement, 3)
experiences to ensure coherent connections and deep have clarity of purpose, and 4) value the subject.
understanding, and, in part by the choices the student Without some degree of motivation, students will not
makes, within parameters established by the system. All pay attention to the task at hand and such inattention
such determinations are designed to ensure consistent decreases the potential for learning.37 Experts in
progressions of learning.31 This adaptive sequencing cognitive science research have shown how intrinsic
of curriculum is supported through research from motivation of students in learning can be enhanced
literature bases on motivation and engagement, prior by building on students’ prior knowledge.38, 39, 40
knowledge, and the theory of flow. Moreover, it enables Making a connection to what interests the student
the presentation of multiple representations of related results in a shift from extrinsic to intrinsic motivation,
ideas, thus enabling contextualization of content and where students focus on their studies because they
processes, ultimately resulting in deeper learning. are motivated by the content itself, with less focus on
earning a grade or pleasing teachers or parents.

17 DREAMBOX.COM
 As described earlier in this paper, keeping a student tapping into the knowledge base from several sources
in his optimized zone of learning requires that the (i.e., schemas) simultaneously. Students involved in
learning activities presented to the student be doable, deep learning need to do the same.
yet challenging. That requires knowledge of the The importance of avoiding cognitive overload has
student’s current skill level, and the complexity of several implications for the design of intelligent
task for which he is ready. Too complex a task for the adaptive learning. First, the development of learning
skill level will frustrate the student, and too simple activities should use screen real estate to display
a task for the skill level will bore the student. The items that the user might need as background in
intelligent adaptive learning system is able to strike completing a task in recognition of limited working
the right balance due to the continuous flow of memory. Second, the learning activities should use
data from the student actions within the system, a combination of visual and textual, or visual and
something that is difficult for a teacher to do without auditory information in order to leverage the full
a technology-based solution such as an intelligent capacity of the working memory. Visuals can support
adaptive learning system. learners with difficult-to-understand concepts.
Visualization and modeling increases conceptual
3. The individualized pace of learning understanding and the likelihood of deep learning and
The research basis for competency-based learning, transfer between situations. For example, students
i.e., holding student learning constant and adjusting who struggle with mathematics often fail
the time, is grounded in studies of mastery learning. As to see relationships between mathematics elements.
noted above, mastery learning was an early precursor Such representations may take the form of number
to intelligent adaptive learning. Typically, both lines, animations demonstrating concepts, drawings,
intelligent adaptive learning activities and mastery fraction/number equivalents, virtual manipulatives
learning are conducted in a low-stress environment, for exploring mathematical properties, etc. According
where the student is allowed sufficient time and to the Institute for Educational Sciences, there is
provided sufficient instruction and scaffolding to moderate evidence for visualization of concepts in
succeed. A researcher who reviewed over 377 studies mathematics.43 Third, sound media design principles
on mastery learning reported an average moderate should be used in screen display to reduce extraneous
effect size of +0.58; that translates into a gain of items, manage cognitive load on working memory,
approximately 22 percentile points for the average and enhance thinking through sound design. A fourth
learner due to their involvement in mastery learning.41 important consideration is the purposeful development
of schemas of understanding by students. This is
4. The regulation of cognitive load accomplished through pedagogical approaches
One of the reasons it is so important to balance that focus on learning with understanding, in
the complexity of the task (that is presented to the addition to ensuring that the student has facilitywith
student) with their current skill level is to avoid learner strategies and algorithms for solving problems
frustration or boredom (i.e., keep them in their zone of efficiently and effectively.
optimal learning).
5. Student engagement through gaming
The culprit that typically causes frustration is cognitive
overload—when the material presented is too complex Student engagement matters. According to Fredricks,
for the student to organize, integrate, synthesize, and increases in cognitive engagement are directly related
understand the concepts contained in the materials. to increases in learning. Cognitive engagement is
There is a neurological reason for cognitive overload a measure of a student’s investment in learning;
and resultant frustration. Neuroscience research his thoughtfulness and willingness to exert effort
points out the limitation of the human brain in its necessary to comprehend complex ideas and master
working memory capacity—the part of the brain used difficult skills.44 There are a number of instructional and
to think. Humans can hold 7 (±2) textual or auditory learning strategies that increase student engagement,
representations, and 4 (±1) visual representations including: a logical sequencing of curriculum, novelty
in working memory.42 The 7 textual or auditory and variety, student choice, intellectual safety (i.e.,
representations can be single words, phrases, or more system assures the intellectual risks will not be
complex schemas. The key to avoiding overload is ridiculed), affirmation of the work and progress, and
to build the capacity of the learner to incrementally clarity of goals.
build schemas of understanding of the topic of study. Gaming is an example of a learning strategy that
A schema is an underlying organizational pattern, embodies several of those listed above. Many
structure, or conceptual framework. Experts studying intelligent tutoring systems use gaming in ways that
complex tasks will hold multiple schemas in their appeal to PreK–12 students.
working memory in order to solve problems by

18 DREAMBOX.COM
 It is ironic that students who find it difficult to • Games and creativity. A study by Michigan State
concentrate on schoolwork for even short periods University researchers studied effects of information
of time, often have no difficulty playing computer technology on the creativity of middle school
games for hours at a time. That irony is not lost on (12-year-old) students. Their results indicated that
educators. According to Gee,45 educators should students who played any type of video game
apply the principles of game design to increase exhibited more creativity than their peers who did
the depth of learning by improving the quality and not play. The researchers compared students’ game
timeliness of feedback (i.e., formative assessment), playing, Internet use, computer use, and cell phone
and by increasing learner motivation, attention, and use, and found that only game playing predicted
engagement in learning. In fact, there is extensive creativity. Results did not vary by gender or race.50
research underway investigating these tools for • Psychological Impact. A report by the Pediatric
educational purposes across a range of grade levels, Clinics of North America reported on positive and
subject areas, student subgroups, and learning negative effects of gaming.51 The positive effects
outcomes.46, 47, 48 included: 1) action games that improved visual-spatial
These four reports provide a range of the many skills, 2) educational games that teach specific
emergent studies in this area: knowledge and skills, 3) exergames that improve
• Simulations and gaming in science. A publication by physical activity levels, and 4) prosocial games
the National Research Council (NRC) summarized that increase empathy,52 helping, and possibly
research on games and simulations in the study of reduce aggression.
science. The NRC found that there was promising Meta-analyses from hundreds of studies on gaming
evidence that simulations could promote conceptual indicate that there is, on average, a significant effect
understanding of science, and moderate evidence on learning, sometimes positive and sometimes
that simulations in science increased students’ negative. The question before researchers is how to
motivation to learn. Though existing studies on leverage the motivational factors into positive results.
gaming in the learning of science seem promising, Studies suggest that the following five principles of
the NRC noted that the body of literature is too thin effective gaming also serve as important elements
to generalize results. of intelligent adaptive learning systems:53, 54
• Online digital content and gaming in preschool 1. sequenced challenges
literacy. A rigorous study found that integrating 2.“just in time” and “on demand” information
digital content from public television video and online
games increased early literacy gains significantly 3.performance before competence,
in preschool children in comparison to groups of 4. motivation and attention
children using more traditional approaches.49
5.timely and specific feedback.

19 DREAMBOX.COM
 In Summary
The intelligent adaptive learning system should be seen as a diagnostic tool, a learning resource, and a source
of valuable data for the teacher, student, and parents. Ultimately, it is the responsibility of the teacher and school
to provide a learning environment that uses the latest research based strategies and technologies to enable
and optimize student learning. An intelligent adaptive learning system is an extremely important asset in the
pedagogical repertoire of the teacher/student team.
As educators make choices about the learning resources they use to advance their students’ learning, they use
criteria established in their instructional materials selection policies. All learning resources used in a school
district, school, or classroom must be vetted. Those criteria should be used when considering intelligent adaptive
learning as well. Three of the key criteria are pedagogy, developmental appropriateness, and alignment to
standards, in addition to the tenets described in this paper.
Given the promising, emerging research on the potential of intelligent adaptive learning systems to individualize
and personalize learning, educators can be optimistic about their potential to improve student achievement.
That said, it will be important that the educators vet the resource to ensure pedagogical alignment, and to plan
carefully how the intelligent adaptive learning—including the data on student progress—will be integrated into the
learning environment.

Sponsored by Intel® Education Initiative

The Intel® Education Initiative: Empowering Tomorrow’s Innovators, recognizes that curiosity, critical thinking and a
strong foundation in math and science are necessary for tomorrow’s workforce to compete for the high-tech jobs of
the 21st century in an increasingly global economy. Intel is involved in education programs, political advocacy and
technology access efforts that enable today’s youth to develop skills they need to be the innovators of tomorrow.
Over the past decade alone, Intel and the Intel Foundation have invested more than $1 billion and Intel employees
have donated close to 3 million volunteer hours toward improving education in more than 60 countries.

20 DREAMBOX.COM
 Endnotes
1. Hattie, J. (2009). Visible Learning: A synthesis of over 800 meta-analyses relating to achievement. New York: Routledge. Location
6091-6096.
2. Sonwalker, N. (2005). Adaptive learning technologies: From one-size-fits-all to individualization. Research Bulletin. Issue 7. Educause.
Center for Applied Research.
3. Fredricks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). School engagement: Potential of the concept, state of the evidence. Review
of Educational Research, 74(1), 59.
4. Ibid.
5. ibid.
6. Marzano, R. J. (2007). The art and science of teaching: A comprehensive framework for effective instruction. Alexandria, VA:
Association for Supervision and Curriculum Development.
7. Nye, B., Konstantopoulos, S., & Hedges, L. V. (2004). How large are teacher effects? Educational Evaluation and Policy Analysis,
26(3), 237–257. Accessed 07/20/12 from http://epa.sagepub.com/content/26/3/237.short.
8. Murray, T., & Arroyo, I. (2002). Toward measuring and maintaining the zone of proximal development in adaptive
instructional systems.
9. Csikszentmihalyi, M. (1991). Flow: The psychology of optimal experience: Steps toward enhancing the quality of life: Harper Collins
Publishers.
10. Csikszentmihalyi, M., & Wolfe, R. (2000). New conceptions and research approaches to creativity: Implications of a systems
perspective for creativity in education. International handbook of giftedness and talents, 81-93.
11. Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Cambridge, MA: Harvard University
Press.
12. Arroyo, I., Beck, J. E., Beal, C. R., & Woolf, B. P. (2003). Learning within the zpd with the animalwatch intelligent tutoring system.
University of Massachusetts, Amherst.
13. Park, O., & Lee, J. (2003). Adaptive instructional systems. Educational Technology Research and Development, 25, 651–684.
14. Ibid.
15. Brusilovsky, P., & Peylo, C. (2003). Adaptive and intelligent web-based educational systems. International Journal of Artificial
Intelligence in Education, 13(2–4), 159–172.
16. Sonwalker, N. (2005). Adaptive learning technologies: From one-size-fits-all to individualization. Research Bulletin, 7.
Educause. Center for Applied Research.
17. Ibid.
18. Biggs, J. B., & Moore, P. J. (1993). The process of learning. New York: Prentice Hall.
19. Bloom, B. S. (1974). An introduction to mastery learning theory. In J. H. Block (Ed.), Schools, society, and mastery learning.
New York: Holt, Rinehart & Winston.
20. Ibid.
21. Chi, M. T. H., Siler, S. A., Jeong, H., Yamauchi, T., & Hausmann, R. G. (2001). Learning from human tutoring. Cognitive Science,
25(4), 471–533.
22. Bloom, B. S. (1984). The 2 sigma problem: The search for methods of group instruction as effective as one-to-one tutoring.
Educational Researcher, 13(6), 4–16.
23. Hattie, J. (2009). Visible Learning: A synthesis of over 800 meta-analyses relating to achievement.
New York: Routledge. Location 6091-6096.
24. Blok, H., Oostdam, R., Otter, M. E., & Overmaat, M. (2002). Computer-assisted instruction in support of beginning reading
instruction: A review. Review of Educational Research, 72(1), 101–130, as cited in Hattie, 2009. Location 8649-8558.
25. Wiliam, D. (2010). The role of formative assessment in effective learning environments.
In The Nature of Learning: Using Research to Inspire Practice, OECD Publishing. doi: 10.1787/9789264086487-8-en
26. Beesley, A. D., & Apthorp, H. S. (2010). Classroom instruction that works second edition research report. Denver, CO: McREL.
27. Dean, C., Hubbell, E. R., Pitler, H., & Stone, B. J. (2012). Classroom instruction that works: Research-based strategies for increasing
student achievement. Alexandria, VA: Association for Supervision & Curriculum Development.
28. Nyquist, J. B. (2003). The benefits of reconstruing feedback as a larger system of formative assessment: a meta-analysis,
unpublished mater of science. Vanderbilt University. As cited in Wiliams, 2010, pp. 125–126.
29. Lou, Y., Abrami, P. C., & d’Apollonia, S. (2001). Small group and individual learning with technology: a meta-analysis.
Review of Educational Research, 71(3), 449–521. as cited in Hattie 2009. Location 8549-8558.

21 DREAMBOX.COM
 30. Nutshall, G. (2007). The hidden lives of learners. NZCER Press, as quoted in Hattie, J. (2009). Visible Learning:
A synthesis of over 800 meta-analyses relating to achievement. New York: Routledge. Location 278-286.
31. Ibid.
32. Hattie, J. (2012). Know thy impact. Feedback for Learning 70(1), 18–23. Accessed 09/15/12 from
http://hoelscherb.pbworks.com/w/file/fetch/58701950/Know_thy_impact.pdf
33. Sonwalker, N. (2005). Adaptive learning technologies: From one-size-fits-all to individualization. Research Bulletin. Issue 7.
Educause. Center for Applied Research.
34. Fredricks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). School engagement: Potential of the concept, state of the evidence.
Review of Educational Research, 74(1), 59.
35. Ibid.
36. Timmerman, C. E., & Kruepke, K. A. (2006). Computer-assisted instruction, media richness, and college student performance.
Communication Education, 55(1), 73–104, as cited in Hattie, 2009.
37. Boekaerts, M., & Universeiteit, K. (2010). The crucial role of motivation and emotion in classroom learning.
In The Nature of Learning: Using Research to Inspire Practice, OECD Publishing. doi: 10.1787/9789264086487-8-en
38. Bransford, J. (2000). How people learn: Brain, mind, experience, and school. National Academies Press.
39. Dweck, C. (2007). Mindset: The new psychology of success. Ballantine Books.
40. Edelson, D. C. (2001). Learning-for-use: A framework for the design of technology-supported inquiry activities. Journal of Research
in Science Teaching, 38(3), 355–385.
41. Hattie, J. (2009). Visible Learning: A synthesis of over 800 meta-analyses relating to achievement. New York. Routledge. Location
6091-6096.
42. Mayer, R. (2010). Learning with technology. In The Nature of Learning: Using Research to Inspire Practice, OECD Publishing.
43. Fredricks, et al., 2004.
44. Gee, J. P. (2007). Learning and games. The John D. and Catherine T. MacArthur Foundation Series on Digital Media
and Learning, 21–40.
45. Shute, V. J., Rieber, L., & Van Eck, R. Games … and … learning. Trends and Issues in Instructional dDsign and Technology,
3. Accessed 6/15/12 from http://myweb.fsu.edu/vshute/pdf/shute%20pres_i.pdf
46. Shute, V. J., Ventura, M., Bauer, M., & Zapata-Rivera, D. (2009). Melding the power of serious games and embedded
assessment to monitor and foster learning. Serious games: Mechanisms and effects, 117–130.
47. National Research Council. (2011). Learning science through computer games and simulations. Committee on Science Learning:
Computer Games, Simulations, and Education, Margaret A. Honey and Margaret L. Hilton, Eds. Board on Science Education, Division
of Behavioral and Social Sciences and Education. Washington, DC: The National Academies Press. Accessed on 9/1/12 from
http://download.nap.edu/cart/download.cgi?&record_id=13078&free=1
48. Penuel, W., Pasnik, S., & Bates, L., Townsend, E., Gallagher, L., Llorente, C., & Hupert, N. (2009). Summative Evaluation of the
Ready to Learn Initiative. Preschool Teachers Can Use a MediaRich Curriculum to Prepare Low Income Children for School Success:
Results of a Randomized Controlled Trial. Accessed 8/12/12 from http://cct.edc.org/rtl/pdf/RTLEvalReport.pdf
49. Jackson, L. A., Witt, E. A., Games, A. I., Fitzgerald, H. E., von Eye, A., & Zhao, Y. (2011). Information technology use and creativity:
Findings from the children and technology project. Computers in Human Behavior. Accessed 8/15/12 from
http://news.msu.edu/media/documents/2011/11/33ba0f16-a2e9-4d36-b063-2f540f115970.pdf
50. Prot, S., McDonald, K., Anderson, C., & Gentile, D. (2012). Video games: Good, bad, or other? Chapter in Children, Adolescents,
and the Media: Pediatric Clinics of North America (pp. 647–658). Elsevier, Inc.
51. Gentile, D. A., Anderson, C. A., Yukawa, S., Ihori, N., Saleem, M., Ming, L. K., et al. (2009). The effects of prosocial video games
on prosocial behaviors: International evidence from correlational, longitudinal, and experimental studies. Personality and Social
Psychology Bulletin, 35(6), 752–763. Accessed 7/12/12 from http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2678173/
52. Gee, 2007.
53. Shute et al., 2009.
54. Gee, 2007.

22 DREAMBOX.COM
 a. Perry, V. (2010). Adaptive and Intelligent System Technologies. Usability.gov. Accessed 11/11/12
from http://www.usability.gov/articles/012010news.html.
b. Bloom, B. S. (1984). The 2 sigma problem: The search for methods of group instruction as effective as one-to-one tutoring.
Educational Researcher, 13(6), 4–16.
c. Bransford, J., Brown, A. L., & Cocking, R. R. (Eds.) (2000). How people learn: Brain, mind, experience, and school.
Washington, DC: National Academies Press.
d. Arroyo, I., Beck, J. E., Beal, C. R., & Woolf, B. P. (2003). Learning within the zpd with the animal watch intelligent tutoring system.
Amherst, MA: University of Massachusetts.
e. Fredricks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). School engagement: Potential of the concept, state of the evidence.
Review of Educational Research, 74(1), 59–109.
f. Gee, J. P. (2007). Learning and games. In K. Salen (Ed.), The ecology of games: Connecting youth, games, and learning.
The John D. and Catherine T. MacArthur Foundation Series on Digital Media and Learning. Cambridge, MA: The MIT Press.
g. Shute, V. J., Ventura, M., Bauer, M., & Zapata-Rivera, D. (2009). Melding the power of serious games and embedded assessment to
monitor and foster learning: Flow and grow. In Ritterfeld, U., Cody, M., Vorderer, P. (Eds.), Serious games: Mechanisms and effects.
New York: Routledge.
h. In the literature, this is referred to as Vygotsky’s zone of proximal development.
i. Labeled the “2 sigma challenge,” Bloom, in the 1980s showed that students who were tutored outperformed students in
conventional classrooms by 2 standard deviations (2 sigmas).
j. Shallow learning is memorization, acquiring facts, skills, or methods, typically resulting in a quantitative increase in knowledge.
Deep learning is when the student makes sense of the concept, representing that deep understanding by interpreting and
applying the concept to a new situation (often referred to as transfer).

23 DREAMBOX.COM
About DreamBox
DreamBox Learning, the leading K - 12 education technology provider, is
committed to radically transforming the way the world learns. As the only dual-
discipline solution rated STRONG by Johns Hopkins EvidenceforESSA.org in both
mathematics and reading, DreamBox provides schools with high-quality adaptive
learning solutions proven to accelerate student growth. Built by teachers,
for teachers, DreamBox empowers educators with robust data analytics and
content-specific professional development to complement instruction. DreamBox
supports 400,000 educators and approximately 6 million students in all 50
states, the District of Columbia, Puerto Rico; and throughout the United Kingdom,
Australia, Canada and Mexico. For more information, visit www.dreambox.com.

© 2022 DreamBox Learning, Inc.

© 2023 DreamBox Learning, Inc. All rights reserved.
All rights reserved.