Mapping Learning


Jim Spohrer
Education Research & Technology
August, 1992

Contents
1. Introduction: Motivation and Goals
2. Cognitive Theories: What is learning?
3. Skills and Competencies: What to learn?
4. Tasks and Activities: What to do to learn?
5. Pedagogy: What to do to foster learning?
6. Tools and Materials: What artifacts foster learning?
7. Automation: What to use computers for?
8. Diffusion: How to influence practice?
9. Concluding Remarks
References

1. Introduction: Motivation and Goals

A map is a tool of great value to explorers. A map not only acts as a repository for what is known about the landmarks in an area, but is also useful for planning where to set out to make new discoveries. Michael Carter, the Director of ATG's Education Research and Technology group, has commissioned a number of us in that group to begin "mapping learning." A map of learning-related issues may help us work more effectively as a team in a time of limited resources, or at least help us gain a better appreciation for alternative perspectives on learning.

The purpose of this document is to provide a brief overview of the main landmarks in seven areas of active learning research. Seven areas were selected because of their fundamental importance to many of our group's endeavors. As we gradually come to agree on what the main landmarks are in each of these areas, we may well be on the road towards a shared map. The seven areas are:

1. Cognitive Theories: What is learning?
2. Skills & Competencies: What to learn?
3. Tasks & Activities: What to do to learn?
4. Pedagogy: How to foster learning?
5. Tools & Materials: What artifacts foster learning?
6. Automation & Amplification: What to use computers for?
7. Diffusion: How to have an impact on practice?

We all have our implicit answers to these questions, and our answers heavily bias our attitudes about which partnerships and R&D efforts should be pursued. This document is an attempt to give short simple answers to these questions of fundamental importance. If this document helps stimulate discussion, then it will be a partial success. Ultimately, the success of this document and its sister documents depends on the degree to which they help us to articulate our implicit assumptions and biases about key landmarks on a shared map of learning technologies. Towards this end, please point out key learning technology landmarks or ideas that have been overlooked or misrepresented in this first draft.

2. Cognitive Theories: What is learning?

Theories of learning fall into three broad camps:

1. Behavioral and associationist theories of learning focus on drives, responses, stimuli, and rewards. Drives and stimuli impel animals with central nervous systems to generate responses, and rewards cause certain stimulus-response connections to be strengthened. Despite the seemingly low-level nature of behavioral theories of learning, some quite sophisticated high-level theories of learning are based on behavioral foundations (cf. Miller and Dollard). Landmark works include those by Thorndike, Tolman, Hull, Skinner, Gagne, Anderson (see reference section for some more details).

2. Cognitive theories of learning focus on knowledge representation and epistemology as well as human information processing and problem solving strategies. Of primary importance is understanding the knowledge structures and processes underlying competent performance on particular tasks in some domain, the initial knowledge state of a learner, and the processes by which a learner is transformed from novice to expert. Predicted misconceptions and errors are used to evaluate the plausibility of competing theories. Landmark works include those by Newell and Simon, Anderson, Glaser.

3. Meta-cognitive theories of learning focus on the cognitive processes that allow learners to self-monitor and self-regulate their thinking. More than others, this group emphasizes the important role that social interactions (social cognition) and environmental interactions (situated cognition) play in fostering the development of thinking and learning skills. Constructivism, active learning, apprenticeship, communities of practice, and portfolio-based assessment are key concepts in the full range of meta-cognitive theories of learning. Landmark works include Palincsar&Brown, Scardemalia&Bereiter, Schoenfeld, Resnick, Collins, Brown&Newman, Brown, Collins,&Duguid.

More on the landmark works can be found in the annotated reference section. The annotations include some quotes and notes that are intended to provide readers with a quick sampling of certain related works.

Dewey, Piaget, Werthheimer, Vygotsky, and Bruner each reacted to the inadequacies of behavioral theories of learning and the traditional model of education. Within the works of these scholars lie many of the seeds of the cognitive and meta-cognitive camps. Some of key insights they offered are summarized below:

Dewey is often referred to as the father of the progressive movement that occurred in American education around the turn of the century. Dewey thought of learners as inheritors of civilization and reconstructors of accumulated experience. Though the terminology of the time was different, Dewey's notions of active, situated (education as a process of living, not of preparing to live), social learning would fit well with the concerns of researchers in the meta-cognitive camp.

Piaget is best known for his work on elucidating the early stages of childhood development. Discrete stages of competencies fits well with the cognitive learning theories notions of novice/expert differences in representation and reasoning capabilities. His view of child as scientist, trying to make sense of the world and actively constructing knowledge, as opposed to simply memorizing it, is an extreme departure from typical behavioral theories.

Wertheimer is best known for his work in Gestalt psychology. Among other things, his work demonstrated that learners who simply memorize procedures are less able to modify those procedures in new situations. On the other hand, learners who actively construct the procedures from more basic principles while performing real world tasks learn when and when not to apply the procedures as well as skills for customizing and extending the procedures when necessary.

Vygotsky introduced the notion of a zone of proximal development in which a learner could interact effectively with a facilitator to acquire new competencies. The zone of proximal development refers to the zone between the things one can already do and the things it would be foolhearty to attempt. The middle range are those things which one can reasonably hope to master either on our own, with the right tools or cognitive artifacts, or with the right facilitators. Much of the work in situated cognition has its roots in this work.

Bruner helped popularize the notion of alternative modes of representation (actions, pictures, and symbols) tied to developmental stages. Representing knowledge in all three modes supports deeper understanding of material, such that problem solving could be done efficiently in the most appropriate mode and difficulties occurring in one mode could be overcome by falling back to a more fundamental mode.

Despite the large number of innovative ideas expressed in the classic works of these scholars, behavioral theories of learning to date have had the largest impact on the practice of education and instruction. This is in part because advocates of behavioral techniques are more often ready, willing, and able to articulate operational and assessable methodologies for the development of instructional materials. Nevertheless, the way learning is conceived and the way it is observed changes over time, and the next paradigm shift will almost surely be in the direction of the cognitive and meta-cognitive camps.

3. Skills and Competencies: What to learn?

The question of what to learn is a negotiation between the needs of society and the wants of individuals. One, admittedly crude, measure of society's needs is what jobs are available -- jobs for pay, jobs pro bono, and jobs for fun. So part of the answer to what to learn is what people need to know to do the range of jobs that society will offer them over the course of their lives. The SCANS report takes this approach and identifies five competencies that lifelong learners and workers in an information-rich, services-oriented economy should possess:

Resources: identifies, organizes, plans, and allocates resources
A. Time - Selects goal-relevant activities, ranks them, allocates time, and prepares and follows schedules
B. Money - Uses or prepares budget, makes forecasts, keeps records, and makes adjustments to meet objectives
C. Material and Facilities - Acquires, stores, allocates, and uses materials or space efficiently
D. Human Resources - Assesses skills and distributes work accordingly, evaluates performance and provides feedback

Interpersonal: Works with others
A. Participates as Member of a Team - contributes to group effort
B. Teaches Others New Skills
C. Serves Client/Customers - works to satisfy customers' expectations
D. Exercises Leadership - communicates ideas to justify position, persuades and convinces others, responsibly challenges existing procedures and policies
E. Negotiates - works toward agreements involving exchange of resources, resolves divergent interests
F. Works with Diversity - works well with men and women from diverse backgrounds

Information: Acquires and uses information
A. Acquires and Evaluates Information
B. Organizes and Maintains Information
C. Interprets and Communicates Information
D. Uses Computers to Process Information

Systems: Understands complex inter-relationships
A. Understands Systems - knows how social, organizational, and technological systems work and operates effectively with them
B. Monitors and Corrects Performance - distinguishes trends, predicts impacts on system operations, diagnoses deviations in systems' performance and corrects malfunctions
C. Improves or Designs Systems - suggests modifications to existing systems and develops new alternative systems to improve performance

Technology: Works with a variety of technologies
A. Selects Technology - chooses procedures, tools or equipment including computers and related technologies
B. Applies Technology to Task - understands overall intent and proper procedures for setup and operation of equipment
C. Maintains and Troubleshoots Equipment - prevents, identifies, or solves problems with equipment, including computers and other technologies

A Three-Part Foundation

Basic Skills: Reads, writes, performs arithmetic and mathematical operations, listens, and speaks
A. Reading - locates, understands, and interprets written information in prose and in documents such as manuals, graphs, and schedules
B. Writing - communicates thoughts, ideas, information, and messages in writing; and creates documents such as letters, directions, manuals, reports, graphs, and flow charts
C. Arithmetic/Mathematics - performs basic computations and approaches practical problems by choosing appropriately from a variety of mathematical techniques
D. Listens - receives, attends to, interprets, and responds to verbal messages and other cues
E. Speaks - organizes ideas and communicates orally

Thinking Skills: Thinks creatively,makes decisions, solves problems, visualizes, knows how to learn, and reasons
A. Creative Thinking - generates new ideas
B. Decision Making - specifies goals and constraints, generates alternatives, considers risks, and evaluates and chooses best alternative
C. Problem Solving - recognizes problems and devises and implements a plan of action
D. Seeing Things in the Mind's Eye - organizes, and processes symbols, pictures, graphs, objects, and other information
E. Knowing How to Learn - uses efficient learning techniques to acquire and apply new knowledge and skills
F. Reasoning - discovers a rule or principle underlying the relationship between two or more objects and applies it when solving a problem

Personal Qualities: Displays responsibility, self-esteem, sociability, self-management, and integrity and honesty

A. Responsibility - exerts a high level of effort and perseveres towards goal attainment
B. Self-Esteem - believes in own self-worth and maintains a positive view of self
C. Sociability - demonstrates understanding, friendliness, adaptability, empathy, and politeness in group settings
D. Self-Management - assesses self accurately, sets personal goals, monitors progress, and exhibits self-control
E. Integrity/Honesty - chooses ethical course of action

The competencies and skills needed for employment were the result of in-depth interviews with workers and their managers in fifteen kinds of jobs. In future releases of this study many more jobs will be analyzed. Some of the jobs that are currently not included in the sample set are scientist, architect, engineer, doctor, lawyer, professor, diplomat, anthropologist, sociologist, psychologist, artist, musician, historian, journalist, politician, athlete, actor. It would be interesting to read a critique of the SCANS report from Brady, author ofWhat's Worth Teaching , and other advocates of a cultural enrichment curriculum.

Bruner and many others have argued that as our society becomes more complex and keeps more and better records, curricula are becoming overloaded with too many details that all students are being encouraged to learn. One way to deal with the problem of too much to learn is to rely on specialization. So rather than thinking about all the skills, concepts, and competencies every individual should aspire to, we can think about the mix of skills, concepts, and competencies that society currently needs and that individuals may be predisposed to attain. There are a lot of people who are against specialization because it narrows people's horizons and puts them into a niche. However, if trends continue, the average person starting work today will have four separate careers by the time they retire -- not different companies and doing the same thing, but different careers. As learning technologies become more effective the barriers to entry on new careers can be reduced. A learner can decide to specialize several times in different areas over the course of one life span. Schools could be set up to allow students to experience several different highly specialized careers in the course of their twelve years of mandated schooling. Each career would add to their portfolio, as well as prepare them to be a flexible and adaptive part of the highly dynamic learning organizations that make up society. Even more ambitious would be for social policy makers to address the four career per life span statistic and design retooling "time-outs" into the very fabric of society. The notion of life-time learning would then be reified in a tangible and significant way.

Most learning goals that emerge naturally, either at work or play, are idiosyncratic. The line between learning and information access blurs when we look at naturally occurring learning goals such as gathering facts or finding out about new procedures. Performance support is the phrase used to describe the learning, doing, referencing activities viewed together instead of separately. What to learn on a day to day basis is hard to predict and plan for, but when the need arises there is a great deal of motivation to satisfy the learning goal. Collins&Brown, as well as others, are compiling evidence that situated learning (cf. Dewey's notion of learning as living, not planning to live) is far more effective than abstract learning out of context. They frequently cite the following result from a study by Miller and Gildea: listening, talking, and reading the average 17-year-old has learned vocabulary at a rate of 5,000 word per year (13 per day) for over 16 years. By contrast, learning words from definitions and sentences abstracted from the context of normal communication, the way vocabulary has often been taught, is about 100 to 200 words per year. Moreover, what is taught turns out to be almost useless in practice. This is not to argue against teaching vocabulary, only to bring under scrutiny how it is currently taught. Ironically perhaps, the performance support and situated cognition movements taken together imply that when learning is not a separate focus of attention but part of a natural doing and referencing context, learning rates sky rocket.

Another answer to this question of what to learn is based on the interests and predispositions of individuals. There is some research on how learners become interested in learning about certain topics. A classic in this area is the work of Miller and Dollard on the role of imitation in social learning. For children and many adults, role models are an important factor in shaping what to focus on and to learn. The excerpt from Bruner in the annotated reference section also touches on the topic of predispositions.

So in summary, the answer to the question of what to learn can be viewed from several perspectives: (1) what should everyone learn, and what should only some individuals learn, (2) what does an individual need to learn in order to be general purpose and versatile, and what does an individual need to learn in order to be special purpose and efficient, (3) what does society need people to learn, and what does the individual want to learn, (4) what learning needs emerge from immediate tasks demands and can be resolved quickly, and what learning needs are recurring and profound requiring major commitments and sustained effort.

4. Tasks and Activities: What to do to learn?

Gagne describes six traditional modes of instruction: tutoring session, lecture, recitation class, discussion class, laboratory, and homework. Students who participate in these modes of instruction typically have the objective of completing requirements, passing tests, or performing better on the job. Each of these traditional modes of instruction has associated tasks and activities. Laboratory (and sometimes tutoring sessions and homework) may require students to design and construct some artifact as part of a project. Lectures are the least active, and students typically listen, take notes, and perhaps ask a question or two. Homework may be largely passive, such as reading, or active, such as problem solving or writing. Laboratories are typically active, but may be little more than unguided procedure following exercises that involve manipulation of tools and materials. Tutoring sessions are typically the most active and guided, involving a great deal of social interaction and reasoning. Recitation and discussion classes may be either largely active or passive, depending on how much preparation the student puts in and how engaged the student becomes in the activity. Discussion classes offer the possibility of a lot of social interaction, as well as a chance to be involved in brainstorming, critiquing, and discussion management activities. Most activities can be made active (active listening, active reading) as well as reflective (writing is rewriting, reading is rereading). However, the traditional modes of instruction coupled with traditional modes of assessment can leave all but the best students in a passive, unreflective, unmotivated state.

Collins and Brown describe both traditional apprenticeship mode of instruction and a newer cognitive apprenticeship mode of instruction. These modes of instruction give rise to activities such as modeling, coaching, scaffolding, and others described more fully in the next section. Playing instructional games, another traditional mode of instruction, has a particular refinement known as role playing that has been rapidly evolving in recent years <look-up reference>.

When learning activities are being compared, some of the important factors that can have a profound impact on the quality of the learning experience are: (1) who sets the learning goal (Piaget, Vygotsky, Scardemalia&Bereiter), (2) what is the situation in which the learning goal emerged and must be satisfied (Dewey, Collins&Brown), and (3) what is the complexity of the learning goal (Gagne, Glaser). For example, if a person is on the job and needs to learn how to make a table in a word processing application this is quite different from a teacher telling a classroom full of students that they will next be learning how to solve Schroedinger's equation in quantum physics. In one case, what to learn and what to do to learn is self-regulated, situated, and relatively simple and in the other it is neither self-regulated, situated, nor simple. Brown, Collins, and Duguid provide a rationale for why situated and self-regulated learning goals speed the rate of learning as well as improve retention. As to the complexity dimension, how to learn can be as simple as looking up the answer in a book or doing a little trial and error experiment (i.e., formatting a table in word) or as complex as organizing a multi-year, multi-million dollar research effort involving thousands of people (e.g., the human genome project).

Another dimension to consider is that there are many ways to learn the same thing - so learners have a range of activities from which to choose. (Or do they? What if instructional designers make these decisions for learners.) Learners can attempt to derive answers on their own, find the answer in some artifact in their environment, or seek out others who might provide what they are seeking. Going it alone may mean trial and error experiments or reasoning from first principles. Finding the answer in some artifact may mean consulting a book. Seeking out help from others may mean going to an expert, teacher, or colleague or enrolling in a class. If what is to be learned is not a simple concept, but a more complex skill or competency, then practice with appropriate opportunities for evaluation and feedback may be required. To gain practice or access to others may mean that the learner has to become a member of some community with perhaps the ritual complexities of initiations and rights of passage.

In sum, in most corporate training centers and public schools, instructors set learning goals for groups of roughly thirty students sitting in classrooms. In these settings, teachers introduce topics, lecture, lead discussions, take field trips, lead laboratory or other hands on activities, answer questions, give assignments, provide feedback, and review topics. Gagne and Mager are noted authorities in the area of traditional instructional design -- given a learning objective each has specified detailed methodologies that can be used to determine effective sequences of tasks and activities for learners. The effectiveness of the activities recommended by these traditional instructional design methodologies is being challenged by cognitive and meta-cognitive camps. The challenge is typically on the grounds that social, situational factors, cognitive, and meta-cognitive factors that can enhance the educational effectiveness of learning activities have not been properly taken into account (e.g., should the instruction take place in a typical classroom setting, should the students work alone or in teams, how can passive, non-constructive, non-reflective activities be recast in active, constructive, reflective ways, who should be in charge of setting the specific learning goals).

5. Pedagogy: What to do to foster learning?

A landmark study in the area of fostering learning is Bloom's "2 Sigma" paper. In this work, he compares traditional classroom instruction to mastery learning and one-on-one tutoring. Traditional classroom instruction is largely a knowledge communication model in which an instructor gives lectures, assigns exercises, and gives tests for assessment purposes. Gagne and Mager describe methodologies for getting the most out of traditional classroom instruction. Mastery learning involves reteaching material based on an analysis of student problems until performance criteria are met. One-on-one tutoring involves having a tutor present whenever the learner needs help. Not surprisingly mastery learning was much better than traditional "teach once, test once" classroom instruction (C students attain B students' level), and one-on-one tutoring was much better still (C students attain A students' levels).

These findings essentially reflect applying additional resources to provide more and more appropriate feedback during the learning process. When students are working on an assigned task, and can get help as they encounter problems, they will be able to solve problems faster with less frustration and spend less time thrashing. A tutor or facilitator supplies the gentle corrective actions that Bruner and Vygotsky discuss at great length as necessary for maintaining sustained productive learning in the zone of proximal development. In Section 7, intelligent tutoring systems will be examined as a potential use of technology to supply the advantage of one-on-one tutoring to all students.

An alternative approach to fostering learning is termed cognitive apprenticeship. Collins, Brown, and Newman describe three success models for cognitive apprenticeship:

1. Palincsar and Brown's reciprocal teaching of reading has proven effective in raising students' scores on reading comprehension tests, especially those of poor readers. Their approach concentrates on four basic skills: asking questions, summarizing paragraphs, making predictions, and clarifying difficulties. After reading a paragraph, whoever is playing the role of teacher exercises each of the four basic skills. Initially, the teacher (or student playing the teacher role) acts as a role model performing the four tasks, and later the teacher acts as a coach: guiding, prompting, critiquing. The teacher provided scaffolding is eventually faded as the students learn to perform effectively on their own. Evaluation studies on students with pretest comprehension accuracies of about 15% were able to perform at the 85% accuracy level after 20 training sessions. This method is effective in part because it helps students see reading as a cognitively intensive constructive activity focussed on producing specific intermediate results as opposed to simply sounding out words and trying to remember details in a one pass operation.

2. Scardemalia and Bereiter's procedural facilitation of writing has proven effective in raising the amount of self-monitoring and self-regulating activities evident in students' thinking aloud protocol data while they are engaged in writing activities. In addition, the judged thought content of student writing increased relative to control groups. Their approach concentrates on five basic subskills: generating ideas, improving ideas, elaborating ideas, identifying goals or the purposes of ideas, and composing ideas. A teacher (or student playing the teacher role) acts as a role model for other students and uses prompt cards associated with each subskill to illustrate how to plan an essay. Later, other students try to plan their own essays and use the prompt cards to help them make effective use of their time. During this so called soloing process, others evaluate and comment on the students progress. Evaluation studies based on very detailed coding of student thinking aloud protocol data show a ten-fold increase in the planning that students exhibit. This method is effective in part because it helps students grow out of a linear knowledge telling approach to writing and grow into an iterative idea planning approach to writing. A key issue in this work is the notion of agency in the zone of proximal develop -- who is in control of the learning process the learner or the facilitator or both?

3. Schoenfeld's method for teaching mathematical problem solving has proven effective in increasing the use of heuristic problem solving techniques by students. His approach does not simply focus on teaching specific heuristics, but also the control strategies and belief systems that allow effective use of the heuristics. The teacher demonstrates the model of problem solving illustrating the purpose of heuristics, control strategies, and beliefs. Students are encouraged to challenge the teacher with difficult problems and evaluate the teachers performance, reversing traditional roles. Schoenfeld advocates small group problem solving as a way to encourage articulation, and the teacher moves between the groups stimulating reflection by asking questions: what are you doing, why are you doing it, and how will succeeding help find a solution. Small groups allow different students to model the different roles required in problem solving: moderator, generator, evaluator. Schoenfeld encourages so called postmortem analysis in which solutions are reviewed and students are encouraged to tell their own story or abstracted replay of the path from problem to solution.

A process model of expert performance of a task is a prerequisite to applying any of these techniques. Glaser takes this one step further and argues for three prerequisites to effective instruction: model of where the student is starting from, model of target expert performance, methods to facilitate the transition from novice to expert. Collins, Brown, and Newman describe the characteristics of the ideal learning environment along four dimensions:

1. Content
Domain knowledge
Heuristic strategies
Control strategies
Learning strategies

2. Methods
Modelling
Coaching Intrinsic motivation
Scaffolding and fading
Articulation
Reflection
Exploration

3. Sequence
Increasing complexity
Increasing diversity
Global before local skills

4. Sociology
Situated learning
Culture of expert practice
Exploiting cooperation
Exploiting competition

In Section 7, some of these ideas will be revisited.

6. Tools and Materials: What artifacts foster learning?

The places where learning is practiced are littered with artifacts. For example, classrooms have chalk, blackboards, pencils, paper, books, rulers, desks; offices have computers, telephones, overhead projectors, transparencies, clocks, teleconference centers, copy machines, printers, manuals; homes have television sets, magazines, newspapers, product information booklets, cameras, tape recorders, radios, toys, and games. What are the main categories of artifacts? How are the artifacts used to foster learning?

There are just a few main categories of artifacts. Gagne refers to six types of media for instruction: objects, oral, text, graphics, animation, video, and teaching machines. Objects are typically thought of in the Montessori sense, as things that can be manipulated to foster learning as in Bruner's action mode of representation or Piaget's concrete operation mode. Alternatively, objects may be thought of in the Don Norman sense (cf. Design of Everyday Things, Things That Make Us Smart), as things that are cognitive artifacts, models, simulations, or representations that foster particular types of reasoning or specific inferences.

Departing from Gagne's categories, artifacts can be clustered by functionality as follows:
(1) Information communication (acquire, store, distribute, display): camera,book
(2) Information processing & decision support artifacts: computers, slide rules
(3) Measurement & observation artifacts: rulers,clocks,scales, microscopes
(4) Manipulatives, stuff & specimens: blocks, chemicals, specimen slides
(5) Furniture: tables, chairs
Communication artifacts are part of a framework in which media, messages, senders, and receivers all play a role. The traditional role of learner has been as the passive receiver of information, but the coming cognitive and meta-cognitive paradigm shift is moving the learner's role more into the constructor, composer, critiquer, and sender of information.

7. Automation & Amplification: What to use computers for?

There are three primary aspects of the learner's experience that can be automated or amplified: (1) cognitive, perceptual, or motor functions, (2) artifacts in the physical environment, and (3) people in the social environment. For example, calculators automate some of the mental procedures that learners execute, online books automate a part of the physical environment, and intelligent tutoring systems automate aspects of learner-tutor interactions. Obviously, automation and amplification change the tasks and activities that can be performed. Typically, one is gaining an economy of effort, but in exchange for something else. When deciding what to put into silicon three issues should be considered: (1) gains (new affordances), (2) loses (decreased mobility, reduced incidental learning), (3) feasibility (can we build it, how long and how much, can it be maintained). It is impossible to fully evaluate these three issues with respect to a new technology before deploying the technology. Nevertheless, with experience one should be able to develop increasingly sophisticated hypotheses.

Given Bloom's finding on the advantages of one-on-one tutoring (C students attain A student levels as reported in the previous section), it is not surprising that intelligent tutoring systems are one of the most promising areas for automation with respect to educational effectiveness measured by pre-test/post-test gains against control groups. Valerie Shute provides one of the most recent surveys of the educational effectiveness of intelligent tutoring systems. Four systems were highlighted in her survey:

1. Anderson, Farrell&Sauers' Lisp Tutor was used to provide students with a series of Lisp programming exercises and tutorial assistance as needed during the solution process. Evaluation studies have compared the system to human tutors and traditional instruction. The system works almost as well as human tutors in promoting faster learning with no degradation in outcome performance. In fact, for students will lower incoming pre-test scores the system seems to improve outcome scores over traditional instruction. Without a tutor, students tend to explore many more unproductive paths and this often leads to thrashing and frustration.

2. Shute & Glaser's Smithtown system provides a microworld environment for learning scientific inquiry skills in the context of basic microeconomics theory. Evaluaion studies compared the system to traditional classroom instruction and a control group. The system works as well as traditional classroom instruction to improve outcome performance, but required less than half the time. Without the constraints of the microworld, students tend to explore many more unproductive paths when working exercises.

3. Lesgold, Lajoie, Bunzo, and Eggan's Sherlock system provides a coached practice environment for learning electronics troubleshooting. Evaluation studies compared students using the system for 20 hours versus students receiving 20 hours of on the job training. The group using the system performed significantly better than the control group. The system allowed students to process far more experiences than they could normally get in any less than a few years of on the job experience.

4. Bonar, Cunningham, Beatty, & Weil's Pascal Tutor assists novice programmers in designing, implementing, and testing Pascal code. Evaluation studies compared students using the system to students receiving traditional classroom instruction. The group using the tutor complete the course in one third the time of traditional classroom instruction with comparable outcome performances. Students who did not use the tutor spent more time exploring unproductive paths.

Each of these systems required several years to develop and deploy. In essence, they all work by keeping the learner on task and preventing useless and frustrating floundering. Because of their effectiveness, especially in terms of decreasing training time, they are all in use today. With better tools, the development times of systems like these could be dramatically reduced.

Progress in the area of intelligent tutoring systems is advancing rapidly, especially in the area that Moyse and Elsom-Cook refer to as knowledge negotiation systems. In many tutoring domains there is no single correct representation of knowledge and the interpretation of knowledge must be jointly constructed between tutor (facilitator) and learner. Another area of increasing activity is in the area of expert critiquing system, such as those developed by Fischer and those developed by Silverman. Finally, coached role play simulators are reaching a level of maturity that indicates evaluation studies may not be forthcoming in the not too distant future (cf. systems developed by Stevens at SEI/CMU and the Dustin system developed by Ohmaye while at ILS/NWU).

In addition to intelligent tutoring systems and simulation-based learning environments that attempt to situate learners in realistic physical and social contexts, computers can be used to support the learners in other ways as well. In particular, performance support systems combine learning, doing, and referencing to provide knowledge workers with computer-based support for communications and knowledge building activities. Electronic mail, Scardemalia and Bereiter's CSILE, Lotus Notes (trademark Lotus Development Corporation) are examples of these sorts of systems. Because these tools are not solely focussed on learning and because content is typically supplied by the users of the systems, electronic performance support tools are part of organizational infrastructure for the information age.

Finally, one other role for computers as a support system to learners is in the area of educational games that support incidental learning (cf. Carmen San Diego, trademark ?). Television and Nintendo account for a lot of kids time. The opportunity to support incidental learning and perhaps enhancing certain types of social interactions is an opportunity to good to pass up. Broderbund's Kid Pix and Maxis Software's SimCity, SimEarth, SimAnt series are clearly pointing the way in this area. Training tapes with high entertainment value (cf. Cleese series) and educational game software (cf. Senge's examples, and Knoware, Inc.) is catching on in businesses as well.

8. Diffusion: How to influence practice?

Building a better mouse trap does not guarantee that people will beat a path to your door, and a lot of good ideas in education have yet to have a major impact on practice. Everitt Roger's classic work on the diffusion of innovation is still one of the key landmarks in this area. Diffusion of innovation typically follows an S-curve: initially slow, followed by an acceleration period, and then a gradual tailing off. Because diffusion is such a problem, waiting to have an innovation before planning for diffusion may mean that diffusion of that innovation will never get a chance to occur. Alternatively, setting up a diffusion network without an appropriately compelling innovation may invalidate those channels for later use. For these reasons, innovation development and innovation diffusion are something of a chicken and egg problem. However, a number of organizations, such as the National Diffusion Network, are being set up by organizations to foster the diffusion of ideas that are proven effective according to some mutually agreed upon criteria. By having some organizations specialize on diffusion of technologies and others specialize on the development of technologies, each organization can make more efficient use of resources.

Typical phases in the diffusion of technology from a user's perspective are awareness, interest, trial, evaluation, and adoption. Innovations are rarely upward compatible or discrete replacements of existing tools, and therefore significant learning curves are associated with accommodating the innovations. Bessant and Simon discuss the need for "technological midwifes" to overcome one of the biggest problems in diffusion -- a lack of human compatibility in the changes that information technologies bring to organizations.

Exactly what human compatibility means with respect to technology adoption remains vague. However, a sense of what is meant by human compatibility can be obtained from reports such as the 1974 National Academy of Engineering's report on "Issues in Educational Technology:"

"Teacher/Learner Acceptance of Technology: Quality and relevance of course materials bear heavily upon whether teachers and learners will accept technology. But these are not necessarily the most important conditions. What appears to be more critical are attitudes towards the environment where the teaching and learning process take place, and what is brought into it through technology. Teachers and learners are confronted by a disturbing problem of exercising responsibility of choice at a quality level. Making the proper choice is part of the philosophical debate over accepting the options offered by technology.

"Why doesn't the school learning environment admit more technology and use it? A teacher decides to use or not to use technology according to his or her perceptions of how technology affects the concept of self as teacher. What questions go through the minds of teachers faced with the choices respecting the use of technology? They often seem to be: What will happen to my role as teacher if technology brings in an environment not fully under my control? What will happened to my students' role and development when more options and choices are made accessible to them? What will happen to our relationship? Will I like or dislike the new activities that will replace some of the things I do now? Will I have to be trained or retrained to use the new technology? Do I want to be? Do I believe that the extra effort and trauma of change will make me a better teacher and help my students learn more? Will it make my work easier or harder? What credit will I get for using technology? If the technology really works, will I be working myself out of a job? I know where my career line is now;where is the career line that strengthens pay and security for innovation and adoption of technology? If I am going to use technology, how do I bring it under my control? How do I have something to say about the content and quality of what comes in? How do I know what technology actually does to people under my charge? Who will pay for the technology? How can I get what I want without a hassle from administration, school board, and taxpayer?"

Bessant and Simon describe similar questions asked by workers in business organizations when new tools and technologies are introduced.

If human compatibility is the limiting factor in the rate of technology diffusion, then perhaps economic imperative is the main driving force. Senge's learning organizations are chiefly organizations that can obtain and maintain a technological edge over the competition -- using communication and simulation tools to support improved responsiveness to customer needs. Chubb and Moe argue that as long as schools are not driven by the same economic imperatives as businesses, the diffusion rate of technological innovation in the schools will remain extremely slow compared to businesses. Perhaps it is more likely that as businesses prove the effectiveness of particular technologies in supporting sustainable, profitable learning organizations, government policy makers will provide opportunities and incentives for those technologies to diffuse into schools. NASDC (New American Schools Development Commission) and America 2000 initiatives seem to indicate that this model of diffusion is at work, from businesses into schools. It is also interesting to note that service-oriented businesses are eager to court very young workers. Often employees can provide students training in using technology in the context of business experiences (as well as pay) that many students find a "better" investment of their time than assigned homework activities. A recent San Jose Mercury News article (9/5/92, p. 5) reported:

Business-school cooperation urged


High school principals and employers together should decide the appropriate training for students who do not want to attend college, the head of the National Alliance of Business said Friday. "We want... employers to get... involved in systematic change" in school curricula for students not bound for college, said NAB President William Kolberg.

Fewer than half the educators and business officials surveyed by the group said high school graduates are prepared in math, writing, and reading, Kolberg said. They also said those students are ineffective communicators and lack the ability to listen, work independently and work on teams.

In sum, the role of business in the diffusion of learning technologies to the schools will probably dramatically increase over the next decade.

9. Concluding Remarks

This paper has surveyed some of the key landmarks that should be included on a map of learning. Seven areas have been touched upon:

1. Cognitive Theories: What is learning?
2. Skills & Competencies: What to learn?
3. Tasks & Activities: What to do to learn?
4. Pedagogy: How to foster learning?
5. Tools & Materials: What artifacts foster learning?
6. Automation & Amplification: What to use computers for?
7. Diffusion: How to have an impact on practice?

In sketching short answers to these questions, this paper has begun the process of making explicit some of the tacit knowledge our group is using to make decisions about which partners and R&D activities are worthwhile to pursue. There are several possible next steps that might be taken to continue the process of explicating our tacit knowledge. For instance, a new set of areas/questions might be identified that provide a different way to carve up a map of learning. Alternatively, within the context of these seven areas, key landmarks that have been overlooked might be added and landmarks that have been misrepresented might be reportrayed more accurately. Another possibility might be to begin an invited speakers series or other activities for the group that would give everyone a deeper, shared understanding of the landmark works. It is my hope that formally or informally we find some way to continue the process of explicitly mapping learning.

References

Advisory Committee on Issues in Educational Technology, Commission on Education, National Academy of Engineering (1974) Issues and Public Policies in Educational Technology. Lexington Books, D.C. Heath and Company. Lexington, MA. page 64-65.

Anderson, J.R., Farrell, R., Sauers, R. (1984) Learning to program in Lisp. Cognitive Science, Norwood, NJ, Vol 8, pp. 87-129.

Becker, Henry Jay (1992) Computer-based integrated learning sysem in the elementary and middle grades: A critical review and synthesis of evaluation reports. Journal of Educational Computer Research, Vol. 8, No. 1, pp.1-41.

Bessant, John and Simon, Bennet (1985) The diffusion of information technology: experience in industry and commerce. In Information Technology and Education. Smith, David J. (Editor). Commissioned by the Econonmic and Social Research Council (ESRC) Education and Human Development Committee, School Goverment Publishing Company, Surrey, UK.

Bloom, B.S. (1984) The 2 sigma problem: The search for methods of group instruction as effective as one-on-one tutoring. Educational Researcher, 13(6), 4-16.

Bonar, J., Cunningham, R., Beatty, P., and Weil, W. (1988) Bridge: Intelligent tutoring system with intermediate representations. Technical report, Learning Research and Development Center, University of Pittsburg, PA.

Brady, Marion (1989) What's worth teaching? Selecting, organizing, and integrating knowledge. New York Press, Albany, NY.

Brown, Ann L. (1992) Design experiments: Theoretical and methodological challenges in creating complex interventions in classroom settings. The Journal of the Learning Sciences, 2(2), 141-178. Lawrence Erlbaum Associates, Inc., Hillsdale, NJ.

Brown, John Seely , Collins, Allan, and Duguid, Paul (1988) Situated cognition and the culture of learning. Institute for research on learning, report no. IRL 88-0008; Bolt, Beranek and Newman, Inc. Research Report 6886. Shorter version in Education Researcher, vol. 18, no. 1, February, 1989.

Bruner, Jerome S. (1966) Toward a theory of instruction. Harvard University Press. Cambridge, MA.

Chubb, John E. and Moe, Terry M. (1990) Politics, markets, and America's schools. The Brookings Institution, Washington, D.C.

Collins, Allan, Brown, John Seeley, Newman, Susan E. (1989) Cognitive apprenticeship: Teaching the crafts of reading, writing, and mathematics. In Resnick, L.B. (ed.), Knowing, learning, and instruction: Essays in honor of Robert Glaser, Lawrence Erlbaum Associates, Hillsdale, NJ.

Finn, Chester E. (1991) We must take charge: Our schools and our future. Macmillian, Inc., NY, NY.

Gagne, Robert M. (1965) The conditions of learning. Holt, Rinehart, and Winston, Inc., NY,NY.

Galleger, Krout, and Egido (1990) Intellectual team work. Lawrence Erlbaum, Hillsdale, NJ.

Gardner et. al. (1990) Towards a scientific practice of science education. Lawrence Erlbaum Publishers, Hillsdale, NJ.

Glaser, R. (1976) Components of a psychology of instruction: Toward a science of design. Review of Educational Research. 46, 1: 1-24.

Guba, Egon G. and Lincoln, Yvonna S. (1989) Four generation evaluation. Sage Publications. Newbury Park, CA.

Kohl, Herbert (1991) I won't learn from you! the role of assent in learning. Milkweed Editions, Minneapolis, MN.

Lesgold, A., Layoie, S.P., Bunzo, M., and Eggan, G. (1990) A coached practice environment for an electronics troubleshooting job. In Larkin, Chabay, and Sheftic (Eds) Computer assisted instruction and intelligent tutoring systems: establishing communication and collaboration, Lawrence Erlbaum Associates, Hillsdale, NJ.

Lumsdaine, A.A. and Glaser, R. (Eds) (1960) Teaching machines and programmed learning: A source book. National Education Association. Washington, D.C.

Lyman, Peter (1984) Reading, writing and word processing: Toward a phenomenology of the computer age. Qualitative Sociology. Spring, Summer.

Mill, Walter (1990) The integrated instructional system report. Educational Information Products Exchange (EPIE), NY, NY.

Miller, David (1985) Popper selections. Princeton University Press. Princeton, NJ.

Miller, G.A., and Gildea, P.M. (1987) How children learn words. Scientific American, 257, 3, 94-99.

Miller, Neal A. and Dollard, John (1941) Social learning and imitation. Yale University Press, New Haven, CT.

Moyse, R. and Elsom-Cook, M.T. (Editors) (1992) Knowledge negotiation. Academic Press, San Diego, CA.

Mulkeen, Thomas A. and Cooper, Bruce S. (1992) Implications of preparing administrators for knowledge work organizations: A case study. Journal of Educational Administration, Vol. 30, No. 1, pp17-28.

Norman, Don. (1989) Design of everyday things. Things that make us smart.

Palincsar, A.S. and Brown, A.L. (1984) Reciprocal teaching of comprehension-fostering and monitoring activities. Cognition and Instruction, [1],117-175.

Reiser, Robert A. (1987) Instructional technology: A history. In Instructional Technologies: Foundations, R. Gagne (editor), Lawrence Erlbaum Associates, Hillsdale, NJ.

Resnick, Lauren B. and Klopfer, Leopold E. (1989) Toward the thinking curriculum: Current cognitive research. Association for Supervision and Curriculum Development (ASCD).

Rogers, E. and Shoemaker, F. (1971) Communication of innovations. Free Press/Macmillan, London.

Rogers, Everitt. The diffusion of innovation.

Saltz, Eli (1971) The cognitive bases of human learning. The Dorsey Press, Homewood, Illinois.

SCANS report (1991) What work requires of schools. The Secretary's Commission on Achieving Necessary Skills (SCANS), U.S. Department of Labor.

Scardemalia, Marlene and Bereiter, Carl (1991) Higher levels of agency for children in knowledge building: A challenge for the design of new knowledge media. The Journal of the Learning Sciences, 1(1), 37-68.

Scardamalia, M., Bereiter, C., and Steinbach, R. (!984) Teachability of the reflective processes in written composition. Cognitive Science, [8], 173-190.

Schlosstein, S. (1989) The end of the American era. Cogden and Weed, NY, NY.

Schoenfeld, A.H. (1984) Mathematical problem solving. Orlando, FL: Academic Press.

Sergiovanni, T.J. and Moore, J.H., editors (1989) Schooling for tomorrow: Directing reforms to issues that count. Allyn and Bacon, Boston, MA.

Shute, Valorie J. (1990) Rose garden promises of intelligent tutoring systems: Blossom or thorn? Paper presented at the Space Operations, Applications and Research (SOAR) Symposium, June 1990, Albuquerque, NM.

Shute, V.J., and Glaser, R. (1992) An intelligent tutoring system for exploring principles in economics. In Snow and Wiley (Eds.) Improving inquiry in social science: A volume in honor of Lee J. Cronbach, Lawrence Erlbaum Associates, Hillsdale, NJ.

Sizer, Theodore R. (1985) Horace's compromise: The dilemma of the America high school. Houghton Mifflin, Boston, MA.

Su, Justine Z.X. (1992) What schools are for: An analysis of findings from a US national study. International Review of Education, Vol. 32, No. 2, March, pp. 133-153.

Wright, Ian (1992) Critical thinking: Curriculum and instructional policy implications. Journal of Educational Policy, Vol 7, No. 2, pp. 37-43.

Annotated References

Advisory Committee on Issues in Educational Technology, Commission on Education, National Academy of Engineering (1974) Issues and Public Policies in Educational Technology. Lexington Books, D.C. Heath and Company. Lexington, MA. page 64-65.

"Teacher/Learner Acceptance of Technology: Quality and relevance of course materials bear heavily upon whether teachers and learners will accept technology. But these are not necessarily the most important conditions. What appears to be more critical are attitudes towards the environment where the teaching and learning process take place, and what is brought into it through technology. Teachers and learners are confronted by a disturbing problem of exercising responsibility of choice at a quality level. Making the proper choice is part of the philosophical debate over accepting the options offered by technology.

"Why doesn't the school learning environment admit more technology and use it? A teacher decides to use or not to use technology according to his or her perceptions of how technology affects the concept of self as teacher. What questions go through the minds of teachers faced with the choices respecting the use of technology? They often seem to be: What will happen to my role as teacher if technology brings in an environment not fully under mycontrol? What will happend to my students' role and development when more options and choices are made accessible to them? What will happend to our relationship? Will I like or dislike the new activities that will replace some of the things I do now? Will I have to be trained or retrained to use the new technology? Do I want to be? Do I belive that the extra effort and trauma of change will make me a better teacher and help my students learn more? Will it make my work easier or harder? What credit will I get for using technology? If the technology really works, will I be working myself out of a job? I know where my career line is now;where is the career line that strengthens pay and security for innovtion and adoption of technology? If I am going to use technology, how do I bring it under my control? How do I have something to say about the content and quality of what comes in? How do I know what technology actually does to people under my charge? Who will pay for the technology? How can I get what I want without a hassle from administration, school board, and taxpayer?"

Anderson, J.R., Farrell, R., Sauers, R. (1984) Learning to program in Lisp. Cognitive Science, Norwood, NJ, Vol 8, pp. 87-129.

Becker, Henry Jay (1992) Computer-based integrated learning system in the elementary and middle grades: A critical review and synthesis of evaluation reports. Journal of Educational Computer Research, Vol. 8, No. 1, pp.1-41.

Bessant, John and Simon, Bennet (1985) The diffusion of information technology: experience in industry and commerce. In Information Technology and Education. Smith, David J. (Editor). Commissioned by the Econonmic and Social Research Council (ESRC) Education and Human Development Committee, School Goverment Publishing Company, Surrey, UK.

"Research on the innovation process over many years has shown that the diffusion of any new idea is never instantaneuos; it follows some kind of S-curve with a slow start, and acceleration as the majority of users take it up and a gradually tailing off pattern as usage approaches saturation. Innovation research is particularly concerned with trying to identify and understand the determinants of this pattern - with a view to extracting guidelines as to how to the shape of the curve might be controlled - e.g. to promote rapid adoption of desirable change and to retard diffusion of undesireable.

"Much of the evidence wich we have from research, however, suggests that one of the biggest problems in diffusion is a lack of human compatability: there is still a need for considerable adjustment to the character and challenges of information technology in organisations. This is to be expected: nnovation theory lumps these factors together under the general heading of "learning effects", and it is on these that I wish to concentrate in this short review. In general there are two main areas of the problem - awareness and implementation; it is useful to discuss these separately.

"Awareness: This stage might almost be regarded as a pre-learning stage, since it involves the activities which need to take place before the firm decides t invest and implement IT. Innovation theory usually breaks this down into five stages: initial awareness, interest, trial, evaluation, and adoption...

...these have tended to be discrete replacements for existing technology - eg a word processor for a typewriter. There is, as yet, little indication of moves towards radical concept of an "integrated office of the future": most commentators see what is currently happening as "office mechanisation" rather than automation.... What appears to be needed is some form of "technological midwife" to ease this process within firms.

"Once the first stage has been reached and the organisation has decided to adopt the technology, the process along the learning curve begins in earnest. Here industrial evidence is interesting; even with relatively low complexity systems there is a significant delay before even average levels of successful operation are achieved - far less best practive. This is important becuase it is often the case that firms do not reckon with the time to learn and expect best practice from day one - and may in extreme cases, reject the technology at an early stage because of this initial disappointment."

Bloom, B.S. (1984) The 2 sigma problem: The search for methods of group instruction as effective as one-on-one tutoring. Educational Researcher, 13(6), 4-16.

Bonar, J., Cunningham, R., Beatty, P., and Weil, W. (1988) Bridge: Intelligent tutoring system with intermediate representations. Technical report, Learning Research and Development Center, University of Pittsburg, PA.

Brady, Marion (1989) What's worth teaching? Selecting, organizing, and integrating knowledge. New York Press, Albany, NY.

traditional view is that the curriculum derives from the needs
of learners, the problems of society, the content of academic
disciplines.
a formal model for the study of the concept of culture...
encompass all knowledge,
suggest new disciplines,
identify areas of study now neglected,
indicate the content most appropriate for general education,
suggest relative significance,
organize content logically,
provide students with conceptual structure,
integrate all parts,
provide mechanisms to expand.

Brown, Ann L. (1992) Design experiments: Theoretical and methodological challenges in creating complex interventions in classroom settings. The Journal of the Learning Sciences, 2(2), 141-178. Lawrence Erlbaum Associates, Inc., Hillsdale, NJ.

"In this section, I trace my own shift toward studying learning in classrooms, although the steps I took reflect changines in the field in general. Beginning in the 1970's, and hence avoiding psychology's major shift away from behaviorist learning theories, I concentrate on the more subtle changes that took place in learning theory after the so-called cognitive revolution. Although the theoretical perturbations of this period were less traumatic than in the previous era, there was nothing short of a sea change in theories of learning, with a corresponding revolution in how learning was conceived, observed, and fostered. There was also a dramatic change in what 'subjects' were required to learn, even in laboratory settings, and an awakening to the fact that real-life learning inevitably takes place in a social context, one such setting being the classroom..."

Brown, John Seely , Collins, Allan, and Duguid, Paul (1988) Situated cognition and the culture of learning. Institute for research on learning, report no. IRL 88-0008; Bolt, Beranek and Newman, Inc. Research Report 6886. Shorter version in Education Researcher, vol. 18, no. 1, February, 1989.

Bruner, Jerome S. (1966) Toward a theory of instruction. Harvard University Press. Cambridge, MA.

"A theory of instruction has four major features.

"First, a theory of instruction should specify the experiences which most effectively implant in the individual a predisposition towards learning - learning in general or a particular type of learning. For example, what sorts of relationships with people and things in the preschool environment will tend to make the child willing and able to learn when he enters school?

"Second, a theory of instruction must specify the ways in which a body of knowledge should be structured so that it can be readily grasped by the learner. "Optimal sctructure" refers to a set of propositions from which a larger body of knowledge can be generated, and it is characteristic that the formulation of the structure depends upon the state of advance of a particular field of knowledge. The nature of different optimal structures will be considered in more detail shortly. Here it suffices to say that since the merit of a structure depends upon its power for simplifying information, for generating new propositions, and for increasing the manipulability of a body of knowledge, structure must always be related to the status and gifts of the learner. Viewed in this way, the optimal structure of a body of knowledge is not absolute but relative.

"Third, a theory of instruction should specify the most effective sequence in which to present materials to be learned. Given, for example, that one wishes to teach the structure of modern physical theory, how does one proceed? Does one present concrete materials first in such a way as to elicit questions about recurrent regularities? Or does one begin with a formalized mathematical notation that makes it simpler to represent the regularities later encountered? What results are in fact produced by each method? And how describe the ideal mix? The question of sequence will be treated in more detail later.

"Finally, a theory of instruction should specify the nature and paciing of rewards and punishments in the process of learning and teaching. Intuitively it seems clear that as learning progresses there is a point at which it is better to shift away from extrinsic rewards, such as teacher's praise, and toward the intrinsic rewards inherent in solving a complex problem for oneself. So, too, there is a point at which immediate reward for performance should be replaced by deferred reward. The timing of the shift from extrinsic to intrinsic and from immediate to deferred reward is poorly understood and obviously important. Is it the case, for example, that whenever learning involves integration of a long sequence of acts, the shift should be made as early as possible from immediate to deferred reward and from extrinsic to intrinsic?"

"...It is customary, in discussing predispositions to learn, to focus upon cultural, motivational, and personal factors affecting the desire to learn and to undertake problem solving... There is, for example, the relation of instructor to student - whatever the formal status of the instructor may be, whether the teacher or parent. Since there is a relationship between one who possesses something and one who does not, there is always a special problem of authority involved in the instructional situation..."

"Since learning and problemsolving depend upon the exploration of alternatives, instruction must facilitate and regulate the exploration of alternatives on the part of the learner.

"There are three aspects to the exploration of alternatives, each of them related to the regulation of search behavior. They can be described in shorthand terms as activation, maintenance, and direction. To put it another way, exploration needs something to get it started, something to keep it going, and something to keep it from going random." (pp.40-43)

Byrne, John A. (1992) Management's new gurus: Business is hungry for fresh approaches to the global marketplace. Business Week, August 31, 1992. pp. 44- 52.

Chubb, John E. and Moe, Terry M. (1990) Politics, markets, and America's schools. The Brookings Institution, Washington, D.C.

"It is one thing to know what kind of organization promotes effective education. It is quite another to know how to use public policy to engineer that kind of organization. Social science seemed to know a lot about the organizational end to be achieved and almost nothing about how to get there."

Collins, Allan, Brown, John Seeley, Newman, Susan E. (1989) Cognitive apprenticeship: Teaching the crafts of reading, writing, and mathematics. In Resnick, L.B. (ed.), Knowing, learning, and instruction: Essays in honor of Robert Glaser, Lawrence Erlbaum Associates, Hillsdale, NJ.

"Three success models for cognitive apprenticeship...

"Palincsar and Brown's Reciprocal teaching of reading...
The procedure is as follows: Both the teacher and the students read a paragraph silently. Whoever is playing the role of the teacher formulates a question based on the paragraph, constructs a summary, and make a prediction or clarification, if any come to mind. Initially, the teacher models the process, eventually turning it over to the students. When students first undertake the process, the teacher coaches them extensively on how to construct good questions and summaries, offering prompts and critiquing their efforts. In this way, the teacher provides scaffolding for the students, enabling them to take on whatever portion of the task they can. As the students become more proficient, the teacher fades, assuming the role of monitor and providing occasional hints or feedback... Reciprocal teaching is extremely effective. In a pilot study with individual students who were poor readers, the method raised subjects' reading and comprehension test scores from 15% to 85% accuracy after 20 training sessions. Six months later students were still at 60% accuracy recovering to 85% after only one session.

"Scardemalia and Bereiter's Procedural Facilitation of Writing...
To encourage students to adopt a more sophisticated writing strategy, Scardemalia, Bereiter, and colleagues have developed a detailed cognitive analysis of the activities of expert writers. This analysis provides the basis for a set of prompts, or Procedural Facilitations, that are designed to reduce student information processing burden by allowing them to select from a limited numebr of diagnostic statements...
Scardemalia and Bereiter's teaching method, like Reciprocal Teaching, proceeds through a combination of modeling, coaching, scaffolding, and fading."

"Schoenfeld's Method for Teaching Mathematical Problem Solving...
Like the other two, this method is based on a new analysis of the knowledge and processes required for expertise, where expertise is understood is understood as the ability to carry out complex problem solving tasks in a domain."

Finn, Chester E. (1991) We must take charge: Our schools and our future. Macmillian, Inc., NY, NY.

Gagne, Robert M. (1965) The conditions of learning. Holt, Rinehart, and Winston, Inc., NY,NY.

"Eight different classes of situations in which human beings learn have been distinguished - eight sets of conditions under which changes in capabilities of human learner are brought about. ... In brief, the varieties of learning that can currently be distinguished are:
Type 1: Signal Learning. Classical conditioned response of Pavlov, 1927.
Type 2: Stimulus-Response Learning: Precise response to discriminated stimulus (Thorndike, 1898;Skinner, 1938).
Type 3: Chaining: Chain of two or more stimulus-response conditions.
Type 4: Verbal Association: Verbal chain specialization of motor chains.
Type 5: Multiple Discrimination: Overcoming interfering packages of stimulus-response chains.
Type 6: Concept Learning: A response that identies an entire class of objects.
Type 7: Principle Learning: If Concept A, Then Concept B.
Type 8: Problem Solving: Problem solving is a kind of learning that requires the internal events usually called thinking. Two or more previously acquired principles somehow combined." (pp. 57-59) Media for instrution: Objects, oral, text, graphics, animation, video, teaching machines.

ObjectsOralText GraphicsAnimationVideoTeaching Machine
Present Stimuliyes limitedlimited yesyes yesyes
Direct Attention no yesyes no no yes yes
Provide Model limited yes yes limited limitedyesyes
Furnish Promptslimitedyesyes limitedlimitedyesyes
Guide Thinking no yes yes no noyes yes
Induce Transferlimitedyeslimited limited limitedlimitedlimited
Assess Attainmentsnoyesyes nonoyesyes
Provide Feedbacklimitedyesyes nolimitedyesyes



"Resources of learning: media for instruction:modes of instruction:
The phrase instructional media is used here to refer to the various kinds of components of the learning environment that generate stimulation to the learner.
Modes of instruction are organizations of media to accomplish certain instructional purposes.
tutoring session
lecture
recitation class
discussion class
laboratory
homework"

Galleger, Krout, and Egido (1990) Intellectual team work. Lawrence Erlbaum, Hillsdale, NJ.

Gardner et. al. (1990) Towards a scientific practice of science education. Lawrence Erlbaum Publishers, Hillsdale, NJ.

processing capacity, existing beliefs, problemsolving ability, meta-cognition.

Glaser, R. (1976) Components of a psychology of instruction: Toward a science of design. Review of Educational Research. 46, 1: 1-24.

Kohl, Herbert (1991) I won't learn from you! the role of assent in learning. Milkweed Editions, Minneapolis, MN.

Lesgold, A., Layoie, S.P., Bunzo, M., and Eggan, G. (1990) A coached practice environment for an electronics troubleshooting job. In Larkin, Chabay, and Sheftic (Eds) Computer assisted instruction and intelligent tutoring systems: establishing communication and collaboration, Lawrence Erlbaum Associates, Hillsdale, NJ.

Lumsdaine, A.A. and Glaser, R. (Eds) (1960) Teaching machines and programmed learning: A source book. National Education Association. Washington, D.C.

[cost high and quality immature slow diffusion]

Lyman, Peter (1984) Reading, writing and word processing: Toward a phenomenology of the computer age. Qualitative Sociology. Spring, Summer.

Mill, Walter (1990) The integrated instructional system report. Educational Information Products Exchange (EPIE), NY, NY.

computers motivate students
networked eliminates some of teacher's logistics problems
flexible targeting - individualized
finding problems - diagnosis, tutoring
single vendor
consistent interface vs teachers choice

Miller, David (1985) Popper selections. Princeton University Press. Princeton, NJ.

Miller, G.A., and Gildea, P.M. (1987) How children learn words. Scientific American, 257, 3, 94-99.

Miller, Neal A. and Dollard, John (1941) Social learning and imitation. Yale University Press, New Haven, CT.

"Human behavior is learned; precisely that behavior which is widely felt to characterize man as a rational being, or as a member of a particular nation or social class, is acquired rather than innate. To understand thoroughly any item of human behavior - either in the social group or in the individual life - one must know the psychological principles involved in its learning and the social conditions under which this learning took place. It is not enough to know either principles or conditions of learning; in order to predict behavior both must be known. The field of psychology describes learning principles, while the various social science disciplines describe the conditions." (p. 1)

"...When are social conditions appropriate for imitative learning? How must the social maze be arranged so that imitative behavior is rewarded? ... There seem to be at least four classes of persons who are imitated by others. They are: (1) superiors in an age-grade hierarchy, (2) superiors in a hierarchy of social status, (3) superiors in an intelligence ranking system, and (4) superior technicians in any field. ... All societies seem to rank people, formally or informally, according at least to some of these criteria. The reasons why imitation occurs in these situations is clear. Superordinated persons recognize the cue stimuli which designate the nearness or presence of important goals. The subordinated, seeking these goals, often find it easier to depend on cues given off by the activities of the leaders. The superiors can act as models and critics to aid their inferiors in perfecting the desired habits." (pp. 183-184).

Mulkeen, Thomas A. and Cooper, Bruce S. (1992) Implications of preparing administrators for knowledge work organizations: A case study. Journal of Educational Administration, Vol. 30, No. 1, pp 17-28.

pre-industrial, rural economy (one room school)
industrial, hierarchy, top-down, bureaurcratic
post-industrial, information intensive, global,
schools as models of professional practice

Norman, Don. (1989) Design of everyday things. Things that make us smart.

Duckworth (Harvard)
Open School.

Palincsar, A.S. and Brown, A.L. (1984) Reciprocal teaching of comprehension-fostering and monitoring activities. Cognition and Instruction, [1],117-175.

Reiser, Robert A. (1987) Instructional technology: A history. In Instructional Technologies: Foundations, R. Gagne (editor), Lawrence Erlbaum Associates, Hillsdale, NJ.

Resnick, Lauren B. and Klopfer, Leopold E. (1989) Toward the thinking curriculum: Current cognitive research. Association for Supervision and Curriculum Development (ASCD).

"For many years, mainstream educational practice was informed by a psychology of learning that lived only uncomfortably with the mind or thinking. Derived from associationist and behaviorist principles, it took learning to be an accumulation of pieces of knowledge and bits of skills. This knowledge could be analyzed into hundreds of components ("association," in the technical language), to be placed in learners' heads through practice and appropriate rewards. Theories of classroom management, textbook design, and organization of practice flowed from this basic assumption. Successive refinements of associationist and behaviorist educational psychology, recognizing layers of complexity and difficulty in knowledge and skills to be learned, proposed hierarchies of objectives or forms of learning - as in the widely-known hierarchies of Benjamin Bloom and Robert Gagne (Bloom 1954,1964, Gagne 1974). Problem solving and other activities recognizable as thinking took their place at the top of these hierarchies, which helped to keep alive the idea that there was more to education than acquiring bodies of facts and associations. But these theories isolated thinking and problem solving from the main, the "basic" or "fundamental," activities of learning. Thinking and reasoning became not the heart of education but the hoped-for capstones that many students never reached.

"There were, of course, challengers to the dominant view. For example, Jean Piaget and his followers have argued for over 50 years that knowledge acquired by memorizing is not real knowledge that can be used (e.g., Piaget 1948/1974). Piaget gave us a picture of the "natural" child as a scientist trying to make sense of the world, and of true learning as constructing ideas, not memorizing information in forms given by teachers or texts. A similar critique was offered by Gestalt psychologists such as Max Werthheimer, who showed that practiced performance in school often masked failure to understand why procedures worked and inability to adapt to modifications in how problems were presented (Werthheimer 1945/1959).

However, these critiques were difficult to adopt into mainstream program organized for practiced performances and demonstrations of mastery on school tests. One reason was Piagetian and Gestalt theories of thinking seemed to deny the importance of specific knowledge and to set curriculum of thinking skills apart from learning subject matter. This is illustrated by the difficulties educators had adopting Piagetian theory even in areas where it had great influence: early childhood and science education. In the former, Piagetiam theory stressed the need for children to develop understanding through their own constructive activities and at their own natural rates of development. It was no simple matter - and one not fully resolved - to figure out how to incorporate teaching important knowledge (the alphabet principle applied to reading, for example, or fundamental arithmetic concepts) within the constraints of a program that fundamentally mistrusted all attempts to instruct directly. In science education, the Piagetian interpretation led to a major emphasis on hands-on laboratories and process skills, but these proved difficult to ntegrate - under Piagetian guidance alone - with the scientific knowledge about which reasoning was to occur.

"Modern cognitive theory resolves this conflict. It offers a perspective on learning that is thinking- and meaning-centered, yet insists on a central place for knowledge and instruction. Cognitive scientist today share with Piagetians a constructivist view of learning, asserting that people are not recorders of information but builders of knowledge structures. To know something is not just to have received information but also to have interpreted it and related it to other knowledge. To be skilled is not just to know how to perform some action but also to know when to perform it and to adpat the performance to varied circumstances.

"Today's cognitive science does not sugget that educators get out of the way so that children can do their natural work, as Piagetian theory often seemed to imply. Instead, cognitive instructional researchers are developing a new body of instructional theory based on constructivist, self-regulating assumptions about the nature of learning. This instructional theory is concerned with all the traditional questions of teaching: how to present ans sequence information, how to organize practice and feedback, how to motivate students, how to integrate laboratory activities with other forms of learning, how to assess learning. But each of these questions is addressed differently than traditional instructional theory, for it is assumed that the goal of all of these instructional activities is to stimulate and nourish students' own mental elaborations of knowledge and to help them grow in their capacity to monitor and guide their own learning and thinking. Thinking and learning merge in today's cognitive perspective, so that cognitive instructional theory is, at its very heart, concerned with the Thinking Curriculum."

The role of social communities provides occassions for modeling effective thinking strategies.

Cognitive apprenticeship - learning takes place in the context of actual work, no problem of how to apply abstract abilities. real tasks, contextualized practice (not just component reductions), and plenty of opportunites to observe and model others.

"Our discussion is organized around three areas of research that Glaser (1976) argues are necessary components of any adequate theory of instruction.
1. Research on the processes that underlie competent performance in any particular area.
2. Research on the initial state of learners before instruction.
3. Research on the process of transition from the learner's initial state to the final goal state."

Dewey: reconstruction of accumulated experience
School and Society (1899)
The Child and the Curriculum (1902)
Democracy and Education (1916)
My Pedagogical Creed
Many experiments that work... why not sustained? Needs an organization and a way to make money
Progressive education (schools to serve changing needs of society and individuals in society - dynamic schools)
Child centered education (schools to serve needs of child)
Rousseau
Montessori, child escapes adult domination (handicapped)
Decroly, centers of interest and educative games
Scientific-realist education
Claparede (child's feet, different shoes; child's mind, one curriculum)
Piaget stages in mental development
- physical manipulatoin, symbolic manipulation
- concrete, mental
social reconstruction: needs of society
- catholic, communist, kibbutz
training apprenticeship, trial & error,
Schon (?) The reflective practitioner & Educating the reflective practitioner.

Rogers, E. and Shoemaker, F. (1971) Communication of innovations. Free Press/Macmillan, London.

Rogers, Everitt. The diffusion of innovation.

Bardini, Terry (USC)
Latour, Bruno give me a laboratory and I will move the world.
Melton, C. (?) Categories of human learning.
Richards, Tyde (accelerate tech adoption, event driven, limited capacity to learn and change habits causes running against brick wall, even for early adopters, system driven, strategist expand system, who has an event?)

Saltz, Eli (1971) The cognitive bases of human learning. The Dorsey Press, Homewood, Illinois.

retentional vs inferential

SCANS report (1991) What work requires of schools. The Secretary's Commission on Achieving Necessary Skills (SCANS), U.S. Department of Labor.

Scardemalia, Marlene and Bereiter, Carl (1991) Higher levels of agency for children in knowledge building: A challenge for the design of new knowledge media. The Journal of the Learning Sciences, 1(1), 37-68.

"The relationship between the ideas of constructivism and of agency in the learning process can vary greatly. At one extreme, in behaviorally oriented social psychology, there is the idea of student as agent without any accompanying notion of student as constructor of knowledge (e.g., Mahoney & Thoresen, 1074). Here the idea of agency is primarily that of students' responsibility for their own success or failure in school. At the other extreme is the idea, often linked with Piaget, of the child as natural scientist, building knowledge of the world through acting on it and trying to make sense of results (e.g., Isaacs, 1930). In between is the Vygotskian notion of child and adult engaged in joint activities in a zone of proximal development, with the child functioning as agent insofar as the activities are concerned but with knowledge being an emergent of the social interaction between the child and the more knowledgeable other (e.g., Newman, Griffin, & Cole, 1989). These conceptions relate to quite different teaching models..."

"Elsewhere we (Bereiter & Scardemalia, 1978a) described three idealized models of teaching. The Teacher A model is a task model. The emphasis is one doing work, with learning assumed to be the by product. Teacher's A role is that of supervisor who oversees the quantity and quality of the work done by the students. Although this model most obviously fits the traditional seat-work oriented classroom, it also fits more open classrooms, so long as the emphasis is on activities rather than on knowledge. Studies of teachers who espouse a process approach to sience learning, for instance, indicate that they tend to believe that activities such as collecting seeds, gorwing plants, or playing with laboratory equipment automatically result in learning, without concern for the cognitive processes involved in the activities. Much educational software lends itself to this activity approach - not only conventional courseware and educational games, but also text processing and graphics software, when it used as a medium for activities.

"The Teacher B model is a knowledge-based model. The focus tends to be on understanding, and the teacher's role includes setting cognitive goals, activating prior knowledge, asking stimulating and leading questions, directing inquiry, and monitoring comprehension. We take this to be the prevailing model of how teaching should be carried out. It appears in most of the curriculum guidelines and teachers' manuals we have seen, and it appears in analyses of exemplary teaching (e.g., Collins & Stevens, 1982; Newman et al., 1989). Knowledge-based intelligent tutoring systems also reflect this model (cf. Polson & Richardson, 1988). Although the active cognitive engagement of students figures prominently into this model, most of the high level control of the learning process remains with the teacher or the program. Teacher B tries to be responsive to the knowledge, the interests, and the knowledge needs of students but by this very effort retains control of the educational process.

"The Teacher C model incorporated eveything that the Teacher B model includes but is distinguished by an effort to turn over to students the high level processes that remain under Teacher B's control. Thus, there is a concern with helping students to formulate their own goals, do do their own activating of prior knowledge, ask their own questions, direct their own inquiry, and to do their own monitoring of comprehension. This model is seen explicitly in reciprocal teaching (Brown & Palincsar, 1989; Palincsar and Brown, 1984) and in procedural facilitation, as applied to writing (Scardemalia, Bereiter, & Steinbach, 1984)."

"...The Teacher B model emphases 'growing into' more mature competence; the Teacher C model adds a deliberate effort to promote' growing out of' dependency."

"What growing out of means may be seen by comparing the prototypic example of mother-child interaction with learning as it occurs in adults. Each of us undoubtedly has a zone of proximal development. Between the zone of what we can already do and the range of thins so far beyond our present competence that it wo ld be foolish to attempt them lies the zone of things that we could profitably work to master. When working in this zone, we -- like the child - can often profit from the involvement of a more knowledgeable other person. But there is an important question which can be brought into focus by this question, Who is in charge of the zone of proximal development?

"Much of what we learn as adults is through interaction with peers, where the zone of proximal development is defined by activities the participants can handle cooperatively better than they can alone. In such cases, control of the zone is mutual. When we seek help from someone more knowledgeable, it is usually on our own terms. Although dependent on the other's competence, we retain the regulating role. We know what we want. We know what deficiencies in our own competence the other is supposed to fill and what we hope to gain from their involvement. We resent it if the more knowledgeable person tries to assume a larger role than we intended."

Scardamalia, M., Bereiter, C., and Steinbach, R. (!984) Teachability of the reflective processes in written composition. Cognitive Science, [8], 173-190.

Schlosstein, S. (1989) The end of the American era. Cogden and Weed, NY, NY.

Schoenfeld, A.H. (1984) Mathematical problem solving. Orlando, FL: Academic Press.

Senge, The fifth discipline.

How ever we perceive the problem of what to learn, we do not want to get caught in the all too familiar traps of simply teaching what has always been taught or teaching what we have materials and know-how to teach. These latter two views are not only traps for our public education system, but are also major obstacles to be overcome in business organizations. Senge, and other "gurus" of the business and management consulting world, advocate learning organizations that value dynamic, fast acting, constantly learning project teams over rigid bureaucratic departmental hierarchies in organizations. Chubb and Moe provide a similar indictment of the current organization of our public school systems.

Sergiovanni, T.J. and Moore, J.H., editors (1989) Schooling for tomorrow: Directing reforms to issues that count. Allyn and Bacon, Boston, MA.

Shute, Valorie J. (1990) Rose garden promises of intelligent tutoring systems: Blossom or thorn? Paper presented at the Space Operations, Applications and Research (SOAR) Symposium, June 1990, Albuquerque, NM.

Shute, V.J., and Glaser, R. (1992) An intelligent tutoring system for exploring principles in economics. In Snow and Wiley (Eds.) Improving inquiry in social science: A volume in honor of Lee J. Cronbach, Lawrence Erlbaum Associates, Hillsdale, NJ.

Sizer, Theodore R. (1985) Horace's compromise: The dilemma of the America high school. Houghton Mifflin, Boston, MA.

Su, Justine Z.X. (1992) What schools are for: An analysis of findings from a US national study. International Review of Education, Vol. 32, No. 2, March, pp. 133-153.

Finally, to understand how to foster learning also requires some analysis of why people want to learn in the first place. People find intrinsic and extrinsic motivations for wanting to learn. A recent study (Su, 1992) investigated four categories of learning goals that most people feel pressured to achieve:
(1) Academic goals to succeed in school.
(2) Vocational goals to succeed in a career.
(3) Social, civic, and cultural goals to succeed in a community.
(4) Personal goals to succeed in life (self,family,friends,beliefs).
Failure to achieve learning goals can be rooted in mental or motivational deficits, or active refusal to learn (cf. Kohl's I Won't Learn from You! ).

Wright, Ian (1992) Critical thinking: Curriculum and instructional policy implications. Journal of Educational Policy, Vol 7, No. 2, pp. 37-43.

Anderson, J.R., Boyle, C.F., Corbett, A.T., and Lewis, M.W. (1990) Cognitive modeling and intelligent tutoring. Artificial Intelligence, 42-7-49.

Fischer, G., Lemke, A.C., McCall, R. (1990) Towards a system architecture supporting contextualized learning. Proceedings of AAAI '90.

Hon, D. (1989) Video ultrasimulation: Creating the experience of skilled performance. CHI'90 tutorial. ACM publishing.

Laffey, L., Machiraju, N.R., Chandrok, R. (1991) Organizational memory as a support for learning and performance: Prototype and issues. In Proceedings of the International Conference on the Learning Sciences. Published by the Association for the Advancement of Science., Charlottesville, VA. p. 289-297.

Lesgold, A., Katz, S., Greenberg, L., Huges, E., and Eggan, G. (1991)
Intelligent Coached Apprenticeship System: Experience with the Sherlock
project. In Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics. Charlottesville, VA.

Nichols, P., Pokorny, R., Jones, G., Gott, S.p., and Alley, W.E. (1992)
Evaluation of an avionic troubleshooting tutoring system. Special Report.
Brooks AFB, TX. Air Force Human Resources Laboratory.

Psotka, J., Massey, L.D., and Mutter, S.A. (Eds.) (1988) Intelligent Tutoring
Systems: Lessons Learned. Lawrence Erlbaum Publishers, Hillsdale, New
Jersey,1988

Silverman, Barry G. (1992) Survey of expert critiquing systems: practical and
theoretical frontiers. Communications of the ACM. 33 (4). April.

Spohrer, J.C., James, A., Abbott, K.A., Czora, G.J., Laffey, J. Miller, M.L.
(1991) A role playing simulator for needs analysis consultations. Proceeding of
the World Congress on Expert Systems, Pergamon Press, pp. 2829-2839.

Spohrer, J.C. (1992) Why computer-based instruction is effective. Unpublished
working paper. Apple Computer, Inc.

Stevens, S.M. (1989). Intelligent interactive video simulation of code
inspection. Communications of the ACM. 32:7, 832-843.

See Also