Overview

Roy D. Pea

SRI International

The Sciences and Technologies of Learning

National Science Foundation

June 2, 1999

Overview Major advances in the learning sciences over past

several decades

Powerful interactive learning environments are building on these developments

Defining and tackling the challenges of scaleup and sustainability

How advances in computing and communications are creating exciting opportunities to address needs

An emerging nexus of technology advances, learning sciences and educational policy

Revolutionary advances in sciences of learning

National Academy of Sciences “How People Learn” (1999)

The nature of expertise

Development of concepts and reasoning abilities

New pedagogies for deep learning of complex subjects

Roles of teacher learning

New assessment approaches for higher standards

Powerful roles for effective use of technologies

Aspects of the sciences of learning

The knowledge-intensive nature of expertise Expertise is not simply general abilities nor use of general

strategies Experts’ extensive knowledge affects what they notice and

how they organize, represent, and interpret information in their environments

Expert knowledge organized in large coherent frameworks

Experts notice features and meaningful information patterns unnoticed by novices

Expert knowledge reflects contexts of application--it is not reducible to isolated facts

Expert knowledge does not guarantee pedagogical knowledge

The importance of representational competencies for expertise

Expertise often involves the skillful creation, use, and interpretation of symbolic expressions (written language, mathematical equations, graphs, technical diagrams, proofs, computer programs)

Experts have greater meta-representational proficiencies than novices—knowing which representational forms are most suitable for asking and answering specific kinds of questions

Experts have facile understanding of the mappings between different representational forms (e.g., algebraic functions to graphs or numerical tables)

Experts are able to assemble arguments, designs, theories, and other complex artefacts that are subject to challenge and testing in a community of peers

The development of concepts and reasoning abilities

Young children rapidly come to make sense of number, language, and causality

In their efforts to make sense of the world, children form robust conceptions that may conflict with the formal knowledge that is later taught (e.g., intuitive physics)

The development of metacognition is a crucial aspect of acquiring expertise and becomes a strategic competency for learning:

Knowledge about one’s knowledge and its limits

Control knowledge about thinking and learning: planning, monitoring, and revising one’s efforts

Contextual and cultural influences on learning

Participation in social practices is a crucial form of learning outside school and in school

The broad diversity of social practices in different cultural contexts creates special challenges for engaging students’ prior knowledge in school

Learning is promoted by social norms that value a search for understanding

Learning is assisted by the family and social environment in which activities provide opportunity for learning through participation

From learning sciences theory to learning environment design Not a simple translation

Physics constrains but does not dictate bridge design (Herbert Simon)

The field of the learning sciences is raising important questions and inquiries:

Rethinking what is taught

Rethinking how it is taught for understanding

Reframing how learning is appropriately assessed Powerful examples of Interactive Learning Environments (ILEs)

that build on our understandings from the sciences of learning SimCalc’s MathWorlds

The Knowledge Integration Environment

WorldWatcher: Scientific visualizations for global investigations

Cognitive tutoring systems

SimCalc: Democratizing access to the Mathematics of Change

Enable all students to develop full understanding and practical skills with the Mathematics of Change and Variation, including fundamental concepts of calculus

As early as Grades 5-8—against a backdrop of ~10% taking High School Calculus, 1.5% taking AP Calculus

Collaborators: Jim Kaput (U. Mass, Dartmouth)

Jeremy Roschelle (SRI International)

Ricardo Nemirovsky (TERC)

Rutgers-Newark; Syracuse; San Diego USI

How can technologies and engaging learning activities change the experiential nature of the Mathematics of Change and Variation by tapping more deeply into students’ cognitive, linguistic, and kinesthetic resources?

Target:

Age:

Who:

Learningsciences

Questions:

SimCalc: Co-evolution of technology and MCV curriculum

Compute within a symbol system

One-way serial links

(e.g., function plots graph)

1970’s 1980’s 1990’s

Multiple linked representations

(formulas, graphs, numerical tables)

MBL/CBL: Physical data collection and

symbolic representation

Source: Kaput, NCTM 2000

New Big Three

The “New Big 3” for Learning the Mathematics of Change and Variation

words

Coordinate graphs

Symbolic formulas

Numericaltables

RATES

words

Coordinate graphs

Symbolic formulas

Numericaltables

TOTALS

PHENOMENAPhysical

CyberneticAbstract

MBL/LBM MBL/LBM

Source: Kaput, NCTM 2000

SimCalc: Co-evolution of MCV curriculum and technology Curriculum: With technology use in activities of predicting,

comparing, designing, build on student experiences with physical change (motions, seasons, aging, growth,

flows) symbolic change (smaller numbers, steeper curves)

Advanced topics: Connections between variable rates and acculumation Velocity, acceleration, limits

Contextualizes other mathematical topics such as: Slope, rate,ratio, proportion Areas of geometric figures

Example of a SimCalc activity

SimCalc outcomes

Technology linkages between experiential phenomena and mathematical representations become conceptually linked in student’s mathematical competencies.

After a three-month supplementary course in MCV using MathWorlds, students from the most troubled high school in Newark NJ achieved near-ceiling effects on assessment items that challenge university calculus students

Testing low-SES school mainstream Grade 6-10 students indicated higher levels of performance after MCV coursework than high-SES Gr 11-12 students taught traditional calculus. They….

Relate slope of position graph to speed of a motion and to the corresponding velocity graph

Infer total distance covered, given by velocity graph, demonstrating accumulation of area under a curve

Now and Future SimCalc

MathWorlds implementation in Java (Roschelle, SRI International)

Incorporation of Java MathWorlds in ESCOT project testbed of interoperable middle school math components

TERC’s LBM (“Line Becomes Motion”):

To incorporate kinesthetic experience, students use mathematical functions created on a computer to control physical devices (like motorized toycars)

MathWorlds commercially available in Flash ROM on TI-83Plus graphing calculators (Fall ‘99) and PCs (Key Curriculum Press, Fall 2000)

Massive teacher development with NJ and Mass SSIs and San Diego USI; T-Cubed workshops run by TI

KIE: Knowledge Integration Environment

To promote coherent knowledge integration in science learning that is reflectively and critically used (versus unconnected facts and beliefs)

Middle to high school sciences Marcia Linn, Jim Slotta, et al (UC Berkeley) and

diverse scientist partners and organizations Expertise involves connected ideas and models

used for reasoning. Do learners develop more integrated

understanding and models when they engage in meaningful collaborative projects using technologies that support key cognitive and social aspects of scientific inquiry and “make thinking visible”?

Target:

Age:

Who:

Learningsciencesissues:

KIE Technology

KIE is a client-side front-end to the World-Wide Web where student project activities are supported by:

SenseMaker: software that ‘scaffolds’ thinking and the organization of critically-considered evidence in scientific argument

KIE Project units

An associated KIE Evidence Database

Mildred the Cow Guide: a provider of reflect process prompts (what to do next and how)

SpeakEasy: net forum for project participants to share issues

Written reflections and class discussions

Student teams work with and/or create scientific evidence in three kinds of supplementary units (2 days to 2 weeks long) Theory comparison projects (e.g., dinosaur extinction, life on

Mars)

Design projects (e.g., an energy-efficient home in the desert using scientific principles)

Critique project (e.g., science tabloid claims on energy conversion)

Scientist partners (e,g., NASA Ames):

Post web evidence for pre-college science teachers

Suggest debates, critiques, or design projects for learners

Mentor students using personal web pages

KIE Curriculum

Students can be effectively encouraged to integrate their knowledge through simple prompts for reflection on their ideas (Mildred the Cow)

Students can develop well-formulated scientific arguments

Net-based discussions enable more students to voice their ideas about the science, especially girls

Major improvements in integrated understanding of project topics such as light, heat, temperature, and sound

KIE Outcomes

Many of KIE’s nearly 20 projects have been classroom-tested

KIE has become WISE (Web-Based Integrated Science Environment)

…and has spawned Project SCOPE:

Science Controversies On-Line: Partnerships in Education

New NSF-funded effort (UC Berkeley, SCIENCE magazine, U. Washington)

Will develop ‘controversy communities’ of scientists and science learners, focusing on controversies that concern leading research scientists and also connect to citizen interests, e.g.,

World-wide control of malaria

Evidence for life on mars

Deformed frogs (environmental chemical or parasite?)

KIE Now and Future

WorldWatcher: Scientific visualizations for global inquiry

Students at all grade levels and in every domain of science should have the opportunity to use scientific inquiry and develop the ability to think and act in ways associated with inquiry…

(National Science Education Standards, National Research Council, 1996, p. 105)

Using visual reasoning for pattern perception in inquiries involving complex data sets

CoVis and later WorldWatcher global warming curriculum as examples

Who: Daniel Edelson (Northwestern U), Roy Pea, Douglas Gordin (now at SRI International)

The multi-agency GLOBE Project coordinated by NSF provides another example

A visualization of temperature data for the Northern Hemisphere displayed by Transform, a powerful, general-purpose visualization environment widely used by scientific researchers

A visualization window from the WorldWatcher software displaying surface temperature for January 1987.

Summary

statistics for

entire image

Summary

statistics for

current

selection region

Readout

showing

lat/long,

country/state,

and data

value for

current mouse

location

Current

mouse

location

Current

selection

region

The interface to the library of energy balance data in the WorldWatcher global warming curriculum

A tenth grade student’s hand-drawn visualization of global temperature for July (Edelson, Gordin, & Pea, J. Learning Sciences, 1999.

Questions about visualization For what domains are visualizations particularly

crucial for promoting understanding? How does the use of these visualizations

influence mental imagery and reasoning in problem solving both while using and when without access to the computer-generated visualizations?

How do how these representations ease the tasks of understanding and using knowledge about the conceptual systems they depict?

We need an empirical science of representational design for understanding complexity, not only capturing and displaying it.

Intelligent tutoring environments Better and more efficient learning of well-structured domains: algebra

I, geometry, algebra II, college algebra Middle school to remedial college Pittsburgh Advanced Cognitive Tutor Center (Koedinger, Anderson,

Corbett); new NSF research center (CIRCLE) Cognitive Tutors conjoin a research base from cognitive psychology

(ACT-R) and artificial intelligence with curriculum content in mathematics from math educators

Key tenets of theory:

Learning by doing, not listening or watching

Production rules represent performance knowledge Units are modular, so isolate skills, concepts, strategies Units are context-specific, so address when as well as how

In search of “2-sigma effect” where human tutors excel over classroom instruction by two standard deviations (Bloom, 1984)

Target:

Age:

Who:

Learningsciencesquestions

What cognitive tutors do Provide a cognitive model that incorporates different strategies and

typical student misconceptions Provide model tracing that follows a student through their individual

approach to a problem (context-sensitivity) Uses knowledge tracing to assess student knowledge growth through

graded levels of competence, and adaptively select activities for learning (“just-in-time” assistance in reasoning)

PUMP algebra tutor provides 1 standard deviation improvement:

Results after 3 years of replicated studies of urban school use in Pittsburgh and Milwaukee indicate increases of 15-25% on standardized tests (SAT subtest, Iowa) and 50%-100% better on problem solving and representation use measures.

Students highly motivated, reduce embarrassment, and succeed Teachers are able to shift their attention and support to struggling

students

The view from research to practice

Too much like Saul Steinberg’s famous New Yorker poster of Manhattan…...“Everyone knows about the advances in the learning sciences”

Really? These advances are too rarely reflected in

educational practices.

New volume on Learning Research and Educational Practice (Bransford and Pellegrino, Co-Chairs)

* Source: 1999 NAS report on “Bridging Learning Research and Educational Practices”

Linear flow modelThe usual means of knowledge transfer through “dissemination” has rarely worked for bringing research to bear broadly on practice*

Reciprocity-of-influence model

* Source: 1999 NAS report on “Bridging Learning Research and Educational Practices”

Defining the challenges of scaleup and sustainability Most studies with designs of interactive learning

environments informed by the sciences of learning are:

Small-scale efforts

Not sustained Common problem of “lethal mutation” of innovations Cultural and linguistic diversity of school environments Importance of attention to standards, accountability,

assessment at a local level Teacher professional development Marketplace issues: from prototypes to sustainable

products and services with needed support

Scaling of innovations

The successes of learning technology innovations are typically accompanied by “researcher hothouse effects”

Common problem of “lethal mutation” of innovations (Ann Brown)

Why? Teachers are designers!

Teachers continue to design curricula in their classroom uses and local adaptations (four phases of curricula)

Need to localize for success rarely supported by teachers’ understanding of the design rationale for why the innovation has its features and practices

Cultural and linguistic diversity of school environments

Standards, accountability, assessment

Curriculum practices are strongly driven by systems of accountability and assessment

Standards provide an important common language for expected outcomes

Educators need usable and compelling forms of assessment in tandem with innovative curricula and technologies for learning

Performance and portfolio assessments are making headway as more meaningful guides to progress

Teacher professional development

Teacher Professional Development (TPD) is a critical component of all education reform efforts

Formal TPD approaches (e.g., summer institutes, collaboratives) can offer motivating, collaborative learning experiences but find it hard to: scale to large numbers sustain collaboration back at teachers’ home sites provide cost- and time-effective support through the

change process tailor content to local school, district initiatives build infrastructure for sustainable TPD (and reform)

systems Difficulties confirmed in evaluation of NSF’s SSI TPD work

Evaluation of NSF’s SSI TPD efforts* Most states provided limited TPD time and the SSIs typically supplemented

formal TPD activities with less than 1 week a year

No SSI had resources to reach all teachers needing TPD—only a minority

Follow-up procedures require many opportunities for assistance, feedback, and reflection in coaching, meeting with others involved in the process or other connections with colleagues

Intros of new practices require time for discussion, questioning, risk-free practice, sharing and reflection, revision

Interaction with colleagues very important, since teachers often work in isolation and lack opportunities to observe others, share their expertise and experience, or practice new techniques.

Good TPD helps build learning communities within, among, and beyond schools

(Source: Corcoran, Shields and Zucker, SRI International, March 1998)

The research-commerce culture divide

Marketplace issues: from prototypes to products and services with necessary support

Two cultures: different audiences, purposes, pressures The divide may narrow as….

Research greets complexities of practice

Grant agencies seek scale and sustainability

Companies seek innovations and to leverage external research

New models for public-private partnerships will need to evolve (beyond “technology transfer”)

Tackling the challenges of scaleup and sustainability

A design research orientation With partnership models that can work in

bringing together necessary expertise and realism to scaleable learning improvements

With networked improvement communities that seek to augment collective intelligence for some purpose and develop sustainable solutions

A Design Research Focus

Design research Challenges the traditional basic-applied science

distinction

Embraces situational complexity and works to manage it through to solutions, and reflect them as cases

“Learning engineering”: Iterative design over multiple generations of a research-guided intervention to improve learning

The need for partnership models

Brief examples:

SCOPE: Science Controversies On-Line: Partnerships in Education (UC Berkeley and SCIENCE magazine)

LeTUS “design circles” of middle school science teachers, curriculum and assessment experts, learning researchers, technology developers (Northwestern and Chicago Schools; U. Michigan and Detroit Schools)

Tackling design research toward scaleable models

Networked improvement communities

Communities that seek to augment collective intelligence for some purpose using the net:

ESCOT integration teams

TAPPED IN and ongoing teacher professional development

CILT and industry alliance program Infrastructure is coming together for schools, homes

ESCOT is a digital library of linkable component tools and a community of teachers, researchers & developers creating, improving, and testing these technologies in real classrooms with real curricula

Principal Investigators: Jeremy Roschelle, Roy Pea, Chris Digiano, Jim Kaput

Curriculum Databaseorganizing information that links concepts, activities & technologies

Software Innovationdesigning pedagogically-soundre-usable, linkable components

Integration Teamscomposing or structuring lessons that tie components to curriculum

The ESCOT Testbed

Towards a digital library of re-usable components for middle-school mathematics

• Key ESCOT Partners:

• SRI International• Key Curriculum Press (Geometer’s Sketchpad)• The Show Me Center, University of Missouri• Swarthmore (MathForum), • University of Colorado, Boulder (AgentSheets)• University of Massachusetts, Dartmouth (SimCalc)

‘Best of class’ graphs, tables, calculators, dynamic geometry, simulations, … 100 or so core elements

Enable plug and play, mix and match Linked multiple representations and

other core educational features

ESCOT Integration Teams put components together

Teacher: Pedagogical Design Developer: Component Design Web facilitator: Web Design (& teamwork)

It’s the right time

Java: a common platform XML: integration glue Web: coordinate distributed work Standards (e.g IMS)

Labelling for search (metadata)

Plug & play, mix match

Linked representations

Web-based teacher professional development (TPD) environment designed with easy-to-understand ‘virtual conference center’ metaphor (‘social computing’ research)

Multi-user, chat, and shared Web browsing

Supports use of assessment and curriculum development tools Significant growth and demand

Over 3,400 registered users, 14 partnership organizations (as unmarketed R&D)

Technical plans for enabling large-scale implementations Strong brand identity and evangelists

NY Times “How to get the most from computers in the classroom”

Highlighted in US Dept of Education’s ”What works”

Working with LA Unified and state of Kentucky in major reform plans Funding by National Science Foundation private foundations, “tenants,” and

corporate sponsorship (Sun Microsystems)

Room description goes here...

Introducing CILT

Roy Pea (SRI), Marcia Linn (UC Berkeley), John Bransford (Vanderbilt), Barbara Means (SRI), Bob Tinker (Concord Consortium)

Concord Consortium

Center for Innovative Learning Technologies

A distributed center for advancing LT R&D

MISSION:

To serve as a national resource for stimulating research on innovative, technology-enabled solutions to critical problems in K-14 learning in science, mathematics, engineering and technology.

Open structure with annual workshops for harvesting knowledge and leveraging diverse efforts

Working on “theme teams” of high-priority led by 2-3 senior researchers and a post-doctoral scholar

Visualization and Modeling

Ubiquitous Computing

Community Tools

Assessments for Learning

C

L

I

T

CILT’s Industry Alliance Program

How we are working on the research-commerce divide through industry alliances

Intel’s senior partnership with CILT community Texas Instruments and hand-held learning

environments Palm Computer sponsorship of CILT educational

software design competition

Some closing observations...

Underlying dynamics of forces in the technological landscape

Moore’s Law

Microprocessing capability doubles every year or at least every 18 months

Metcalfe’s Law

The value of a network is the square of the number of nodes connected to that network

Supply chain efficiencies and virtual companies are revamping global business and will affect every life sector

(PITAC 98 Report)

Fast PCs and information appliances, “fat pipes,” digital content

President’s Information Technology Advisory Committee Report to the

President (August 98)

Vision of Transforming the Way We Learn

“Any individual can participate in on-line education programs regardless of geographic location, age, physical limitation, or personal schedule. Everyone can access repositories of educational materials, easily recalling past lessons, updating skills, or selecting from among different teaching methods in order to discover the most effective style for that individual. Educational programs can be customized to each individual’s needs, so that our information revolution reaches everyone and no one gets left behind.”

It’s not enough of a “Grand Challenge” Enabling this vision requires re-inventing learning substantively, not

only the HOW and WHEN of learning We will do better at re-inventing learning if we heed the PITAC

visions of transforming the ways we:

Communicate

Deal with information (I/O)

Work

Design and build things

Conduct research

Deal with the environment

Do commerce Each of these areas in society is spawning new literacies and

required skills for an informed and proficient citizen. Keeping education apace of the needed learning curve is the Grand

Challenge

Looking forward, computational media will...

Because of their use in research and society, continue to create new content, in mathematics and science such as complexity theory, neural nets, emergence

Allow broad accessibility of powerful ideas, and alter the age level and sequencing of curriculum we will need to invent to meet the demands of a new knowledge age

Thus require ‘partnership model’ research and development at the edges of content-coming-to-be — collaborative innovations and empirical investigations of co-evolving subject matter, technology and appropriate curriculum

Such research by definition will be at the interstices of the disciplinary areas by which the National Science Foundation is organized

Let’s work together to rise to this Grand Challenge for learning and its affiliated sciences…….

THANKS FOR YOUR TIME

Please visit us at

CILT.org

and

SRI.com/Policy/CTL!!

Large research challenges

Infomating the physical environment for learning with ubiquitous computing and transmitting (Spohrer’s WorldBoard concept extending web URLs to geo-located things-in-the-world)

Developing Bayesian and other machine learning approaches to user-profiling of sufficient power that they may infer a learner’s interests and abilities from their net-based interactons, and offer up relevant resources to learn

Pervasive knowledge integration environments with rich and age-appropriate metadata cataloging of web resources for inquiries to develop high-standards learning

Lifelong digital portfolios of learning

Facing the challenge

“We predict that ‘educational portals’ providing a gateway to the Internet, the world’s greatest library, will emerge in K-12, postsecondary and corporate training markets.”

The Book of Knowledge: Investing in the growing education and training industry. M. Moe, K. Bailey, & R. Lau. Merrill-Lynch In-Depth Report, April 9, 1999.

Research in context of learning portals is needed

To grow connected learning communities based on...

Quality (from research and experience)

Cooperation (we share information to help one another learn)

Collaboration (learning together)

Communication

To accelerate distributed learning….

Effective use of better standards-based learning resources and assessments

Teacher professional development

Effective student use of the Internet for learning

School-home connections

To bring customers-providers together more effectively

Emerging Learning “Solutions”

Concept: Service provision via web-based network from any device*

* As in: Sun Microsystem’s “WebTone” Microsoft’s “Digital Nervous System”

ASPs

CSPsISPs WAN Connectivity

Zero-Admin

ServersLAN

PCs

Macs

Thin clients

HHC

Web-phones

Schools Homes

K-12 Portals as leverage point for investment