A Scientific Learning Whitepaper

1956 Webster Street, Suite 200

Oakland, CA 94612
888-816-0100
www.SciLEARN.com

Generalization: Making Learning More than a “Classroom Exercise”

Logan De Ley, M.A., M.S.

Senior Researcher
October 12, 2016

©2016 Scientific Learning Corp. 1

What is Generalization?

As educators, we ask students to do many things, from memorizing multiplication tables, to writing
essays, from reciting poetry to conducting scientific experiments. Yet even the most elaborate
classroom project is but a means to an end; all of these activities are intended to help students become
informed citizens, effective communicators, reflective decision makers, and insightful problem solvers,
whatever their future pursuits. Achieving this goal means teaching for generalization, and this focus
affects what we teach, how we teach, even when and why we assess.
Generalization (or transfer) of learning is the ability to take skills or concepts learned in one context and
apply them to novel problems in different contexts. Many problems are superficially different but
structurally similar; generalization requires looking past the superficial differences to perceive the
deeper relationships. For example, consider two problems that require an understanding of
percentages: answering a percentage problem on a worksheet in a quiet classroom versus mentally
calculating the sales tax on a large purchase in a busy store. While these are very similar problems, they
are presented and organized differently, in the context of very different social and emotional cues.
Much as we might like to assume that “if we teach it, they will apply it,” this view has been debunked by
more than a century of research. Studies show that learners often fail to apply known skills and
strategies to novel problems, whether the learners in question are rats in mazes, adults solving puzzles,
or school children in the classroom (Marini & Genereux, 1995). These findings have spurred further
research to identify conditions that foster generalization for various skills.
Student success hinges on the acquisition and generalization of many skills, including cognitive skills
and academic skills. Cognitive skills, such as memory, attention, processing and sequencing, provide a
foundation that is critical for other kinds of learning, both in and out of school. Educators build on this
foundation through academic instruction – teaching students how to read, perform mathematical
computations, place current events in historical context, etc. Despite the differences in these skill sets,
generalization is crucial for both cognitive and academic skills.

Generalization in Cognitive Skill Development

The question of “does it generalize?” has been particularly controversial in the area of cognitive skill
development. Cognitive skills are conceptualized as general-purpose abilities that serve all kinds of
learning. Indeed, research has shown that strong cognitive skills predict later academic success
(Duncan, et al., 2007; Swanson, Jerman, & Zheng, 2008). Still, some critics have contended that
cognitive skill training is too narrow, and too different from academic learning, to help students who
are struggling in the classroom.

Recent research in the domain of working-memory skills has added fuel to this controversy. For
instance, a 2013 meta-analysis by Melby-Lervag and Hulme found no evidence that working memory
training produces sustained benefit in academic skills, such as arithmetic and decoding. In 2009,
Holmes, Gathercole, and Dunning reported significant gains in working memory following the use of an

©2016 Scientific Learning Corp. 2

adaptive program. However, when the same authors later focused on generalization, they found that
student gains did not transfer to improved performance on academic tasks – even for tasks known to
place high demands on working memory (Dunning, Holmes & Gathercole, 2013). They concluded that
the adaptive program used in their studies was effective at building the targeted working memory
skills, but that it neglected to provide the scaffolding and support needed to generalize these skills into
academic tasks and other real-life situations.

According to Amso and Scerif, in their 2015 literature review, the area of visual attention training has
yielded similar results. Repetitively exercising one cognitive domain may produce near-transfer without
far-transfer. In other words, learners may get better at tasks that are very similar to the training task,
without generalizing their learning more broadly, or showing improved educational achievement. The
authors concluded that the simplest training approaches are not optimal for promoting generalization.
Instead, programs should take a more complex approach, strengthening multiple top-down and
bottom-up processing systems.

Bridging the Gap

An example of this more complex approach can be seen in the Fast ForWord family of products. Based
on decades of research into how students learn, the products build top-down skills, such as working
memory and attention, along with bottom-up skills such as perceptual processing. Furthermore, the
products build these cognitive skills in the context of language and literacy activities, promoting
transfer by bridging the gap between learning isolated skills and applying them in meaningful activities.
Fast ForWord supports generalization through the combined principles of simultaneous development
and cross-training. Each Fast ForWord exercise focuses on a specific language or reading task while
simultaneously developing underlying cognitive skills such as memory, attention, and processing.
Different exercises are combined within a Fast ForWord product to cross-train each skill using different
tasks and stimuli. Simultaneous development and cross-training ensures that each product engages
multiple components of the highly interconnected networks of brain structures involved in language,
reading, and learning.
The design of the Fast ForWord product sequence also supports generalization and the application of
cognitive skills to learning academic subjects. In the most typical sequence, a student begins with a
product that builds core cognitive and language skills, using tasks that bear little resemblance to
familiar academic activities. Subsequent products continue to build on those core cognitive skills, but
they increasingly focus on more advanced language and reading skills, using tasks modeled after best
practices academic instruction. Because they receive extensive practice in applying their new cognitive
skills to authentic reading and language arts activities, Fast ForWord users are well positioned to
generalize these skills to the classroom and beyond.
School-based research indicates that students who use the Fast ForWord products are successful at
generalizing their cognitive skill gains. For instance, a longitudinal study conducted with the Everett
Public Schools (Scientific Learning, 2010) found that students made significant reading achievement
gains during the year that they used the Fast ForWord products, and that they continued to make

©2016 Scientific Learning Corp. 3

reading gains in subsequent years. Another study, conducted with the St. Mary Parish Schools
(Scientific Learning, 2013) evaluated student results across the subject areas of English Language Arts,
Math, Science, and Social Studies. After district-wide Fast ForWord adoption, students improved their
achievement levels in all four domains, as shown in figures 1-4.

4th Grade ELA 4th Grade Math

St. Mary Parish School District State St. Mary Parish School District State
90 90

Basic or Above on Math (All Testers)

Basic or Above on ELA (All Testers)

80

Percent of Fourth Graders

80
Percent of Fourth Graders

70 70

60 60

50 50

40 40
Fast ForWord Use Fast ForWord Use
30 30
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Year Year

4th Grade Social Studies 4th Grade Science

St. Mary Parish School District State St. Mary Parish School District State
Basic or Above on Science (All Testers)

90 90
Basic or Above on Social Studies

w13stm01 w13stm02
4th graders in district who are Basic or above -- compared to state 4th graders in district who are Basic or above -- compared to state
Percent of Fourth Graders

80 80
Percent of Fourth Graders

ALL Testers ALL Testers

Started using FFWD in 2006-7 Started using FFWD in 2006-7

70 70
(All Testers)

60 60

50 50

40 40
Fast ForWord Use Fast ForWord Use
30 30
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Year Year

Figures 1-4. Results on the Louisiana Educational Assessment Program (LEAP) for fourth graders in the Saint Mary
Parish district, compared to students statewide, for test years 2003-2012. The Fast ForWord products were adopted
by district elementary schools between the 2007 and 2009 test years.
w13stm04
4th graders in district who are Basic or above -- compared to state w13stm03
ALL Testers 4th graders in district who are Basic or above -- compared to state
ALL Testers
Results like these indicate that the Fast ForWord approach builds cognitive skills that generalize,
Started using FFWD in 2006-7
Started using FFWD in 2006-7

helping students to attain and sustain improved learning trajectories and to get more out of the
classroom curriculum, across disciplines.

Generalization in Classroom Instruction

Teaching for generalization doesn’t require starting over with entirely new methods of instruction.
Established, time-tested approaches can be very successful, when used at the right time for the right
material. For example, despite their philosophical differences, both discovery learning and direct

©2016 Scientific Learning Corp. 4

instruction can promote generalization when used appropriately. Likewise, well-timed and thoughtfully
constructed assessments can be much more than the final resting place for so much classroom learning.
Educators can promote cross-disciplinary thinking by presenting multiple examples, revisiting ideas in
different contexts, and assigning projects that cross domain boundaries. The use of multiple examples
is especially important for helping students to distinguish the core features of a concept from
superficial features linked to a specific context (Gick and Holyoak, 1983). At the same time, building
deep, domain-specific knowledge is also critical since it helps students recognize structural similarities
between prior learning and new problems encountered in different contexts.
Figuring something out for oneself requires deep and active mental processing, which is one reason
why discovery learning approaches can promote generalization. These approaches work well for
concepts that can be readily induced from the feedback learners generate through exploration, and
they help to prepare learners who can invent novel solutions (McDaniel and Schlager, 1990). Direct
instruction approaches can also promote generalization by efficiently helping learners to build deep
domain-specific knowledge, and to focus on core concepts.
Direct instruction can also be integrated into a blended approach known as assisted- or enhanced-
discovery learning. In one study of this approach, Chen and Klahr (1999) attempted to teach a
challenging principle of experimental design to children in grades 2-4, using varied levels of support.
They found that only the highest level of support led to effective learning. When they combined
discovery learning with leading probe questions and direct instruction, even the youngest students
were able to learn the principle, and the oldest students were able to generalize the principle to new
conceptual domains. Indeed, a large body of research indicates that the blended approach is superior to
either unsupported discovery learning or direct instruction alone (Alfieri, Brooks, Aldrich & Tenenbaum,
2011).
Another approach that successfully blends discovery learning with direct instruction is widely used in
Japanese math classes. Students are presented with solved problems and asked to explain why the
solutions are correct or incorrect. This task combines the active, analytic thinking of discovery learning
with the targeted focus of direct instruction. Research with students in the U.S. found that this method
improved the learning of math concepts and markedly improved generalization (Siegler, 2002).
Assessment can also play an important role in teaching for generalization. We know that testing boosts
learning by helping students strengthen the retrieval component of memory. There is also a growing
body of evidence that testing boosts generalization. Studies show that taking a practice test is more
effective as a study strategy than rereading a text or reviewing notes (Carpenter, 2012). Tested
students retain what they have learned over longer periods of time, they perform better when faced
with new test formats, and they are more likely to make inferences both within and across knowledge
domains. Informal testing with feedback can also improve metacognitive awareness – helping students
to understand their own thinking and learning, and to recognize when they need to change their
learning strategies.

©2016 Scientific Learning Corp. 5

Conclusion

After all the “high stakes” tests have been taken, grades recorded, and diplomas handed out, the real
test begins. Have your students built the knowledge, skills, and dispositions that will serve them in their
lives and help them become productive members of their communities? The answer to this question
hinges on whether your instructional approaches and materials promote generalization.

Bibliography and further reading:

Alfieri, L., Brooks, P. J., Aldrich, N. J. & Tenenbaum, H. R. (2011). Does discovery-based instruction
enhance learning? Journal of Educational Psychology, 103, 1-18.

Amso, D., & Scerif, G. (2015). The attentive brain: insights from developmental cognitive
neuroscience. Nature Reviews Neuroscience, 16, 606-619.

Carpenter, S. K. (2012). Testing enhances the transfer of learning. Current Directions in Psychological
Science, 21, 279-283.

Chen, Z. & Klahr, D. (1999). All other things being equal: Acquisition and transfer of the control of
variables strategy. Child Development, 70, 1098–1120.
Duncan, G. J., Claessens, A., Huston, A. C., Pagani, L. S., Engel, M., Sexton, H., Dowsett, C. J.,
Magnuson, K., Klebanov, P., Feinstein, L., Brooks-Gunn, J., Duckworth, K., & Japel, C. (2007).
School readiness and later achievement. Developmental Psychology, 43, 1428-1446.
Dunning, D. L., Holmes, J., & Gathercole, S. E. (2013). Does working memory training lead to
generalized improvements in children with low working memory? A randomized controlled
trial. Developmental Science, 16, 915-925.

Gick, M. L. & Holyoak, K. J. (1983). Schema induction and analogical transfer. Cognitive Psychology, 15,
1-38.

Holmes, J., Gathercole, S.E., & Dunning, D. L. (2009). Adaptive training leads to sustained
enhancement of poor working memory in children. Developmental Science, 12, F9-F15.

Marini, A. & Genereux, R. (1995). The Challenge of teaching for transfer. From A. McKeough, J.L.
Lupart, A. Marini (Eds.) Teaching for Transfer: Fostering Generalization in Learning (pp.1-20).
Mahwah, N.J.: Lawrence Erlbaum Associates.
McDaniel, M. A. & Schlager, M. S. (1990). Discovery learning and transfer of problem solving skills.
Cognition and Instruction, 7, 129-159, Hillsdale, NJ: Lawrence Erlbaum Associates.
Scientific Learning Corporation. (2010). Improved reading achievement by students in the Everett
Public Schools who used Fast ForWord® products: 2006-2009, Scientific Learning: Research
Reports, 14 (2): 1-6.

©2016 Scientific Learning Corp. 6

Scientific Learning Corporation. (2013). Improved longitudinal achievement in English Language Arts,
Math, Science, and Social Studies by students in St. Mary Parish Schools who used Scientific
Learning Products, Scientific Learning: Research Reports 17 (6): 1-8.
Siegler, R. S. (2002). Microgenetic studies of self-explanations. In N. Granott & J. Parziale (Eds.),
Microdevelopment: Transition Processes in Development and Learning (pp. 31-58). New York:
Cambridge University.

Swanson, H. L., Jerman, O., & Zheng, X. (2008). Growth in working memory and mathematical problem
solving in children at risk and not at risk for serious math difficulties. Journal of Educational
Psychology, 100, 343-379.

Other readings:

Agocs, M. (2011). Fifth graders make significantly greater gains than a comparison group across
multiple subjects after Fast ForWord. The Science of Learning Blog.
http://www.scilearn.com/blog/fifth-graders-make-gains-across-multiple-subjects

Agocs, M. (2011). Longitudinal study shows significant Fast ForWord gains endure over time. The
Science of Learning Blog. http://www.scilearn.com/blog/fast-forword-gains-endure-over-time-
longitudinal-study .

Jenkins, B. (2012). Human intelligence and the brain: Mapping intellectual ability. The Science of
Learning Blog. http://www.scilearn.com/blog/human-intelligence-brain-mapping-intellectual-
ability

Marzano, R. J. (2011) Art & science of teaching: The perils and promises of discovery learning.
Educational Leadership, 69 (1), 86-87. http://www.ascd.org/publications/educational-
leadership/sept11/vol69/num01/The-Perils-and-Promises-of-Discovery-Learning.aspx

©2016 Scientific Learning Corp. 7