ESTEPWeb.

org: A Case of Theory-based Web Course Design

Sharon Derry and the STEP Team1

To appear in A. O’Donnell & C. Hmelo (Eds) Collaboration, Reasoning and Technology.
Mahwah, NJ: Erlbaum.

Final Draft: December 21, 2002

The web-based course may be the new college textbook. Only it’s potentially better, better than a

book or even a book with a CD and a supporting web site. Web courses are more interactive than

books. Not all are good of course, but many are lovingly and creatively, if not scientifically,

designed. They are perpetually updateable and expandable. They also can be very adaptable.

Our evolving web course (eSTEPweb.org) is associated with a theory-based research and

development effort called the Secondary Teacher Education Project (STEP). One goal of our

work is to achieve a design good enough to virtually insure that prospective teachers acquire

current scientific knowledge about human learning and development in educational settings -- a

field of study designated as the “learning sciences” -- in a form that will be useful to them in

professional practice. Insuring that professional students acquire usable, transferable knowledge

from classroom instruction is a difficult order; research has shown repeatedly that non-trivial

forms of transfer are difficult to achieve. This chapter will tell a story about our efforts to develop

such a design. We will discuss the theory behind our work, key design decisions, and some early

lessons learned from the experience so far. We begin with accounts of the course’s three main

technical components. Two of them are contained in the STEP Knowledge Web (“KWeb” for

short): An online hypertextbook of learning science theories and research (“Theories”) that is

1
Many have been involved in conceptualization and development of the STEP system and its research
program. The ideas in this paper build especially on work and discussions with Rand Spiro, Cindy Hmelo,
David Woods, Matt Delmarcelle, Constance Steinkuehler, Chris Fassnacht, Kate Hewson, John Stampen,
and others.

1
interlinked with a library of video cases depicting stories of teaching and learning in actual

classroom settings (“Cases”). A third component is a site to support collaborative learning

through lesson-design and other problem-based learning activities that make use of the KWeb in

various ways. The discussion of these components will be followed by data from formative

evaluations of our course. We end with a brief discussion of the research issues raised by our

work and a brief description of our research program.

The STEP Knowledge Web

The STEP KWeb is an evolving hypermedia resource library on line, intended to support

professional development that aims to help teachers acquire knowledge about learning sciences

and the ability to use it to think flexibly and analytically about instruction, learning, and the

design of learning environments. The design of KWeb was based in the belief that the conceptual

systems of advanced professional expertise are semiotic and “ill-structured” (e.g., Spiro et al.,

1992), largely because they are repeatedly pulled apart and flexibly reassembled in practice. The

conceptual knowledge of the expert practitioner differs from that of the novice practitioner in that

it is more intertwined with and shaped by many varieties of practice, and is less tightly embedded

in simplifying structural devices, such as hierarchies, through which the subject-matter was

probably originally taught and acquired. The problem of developing teachers as expert users of

the learning sciences is one of designing a system that can accelerate the process of acquiring

conceptual knowledge in a form that interpenetrates activities of professional practice (e.g., the

intertwining of learning sciences with the design and mentoring of classroom instruction).

This way of thinking about web-based instructional design was partly inspired by Cognitive

Flexibility Theory (CFT; Spiro et al., 1991a; 1991b; 1992), a theory of advanced knowledge

acquisition in ill-structured domains. CFT developers advocate advanced professional instruction

that emphasizes, not hides, the complex interactions among conceptual knowledge and the

professional activities that utilize that knowledge. The CFT design hypothesis is as follows: After

2
acquiring a basic body of principled foundational knowledge (e.g., the physical science of

medical practice; the learning science of teaching practice), students’ development toward

flexible “expert” habits of mind, supported by complex memory representations in which

foundational concepts interpenetrate practitioner knowledge, can be accelerated through

classroom instruction in which foundational knowledge is elaborated and restructured through

particular, intensive forms of case-based instruction. The instructional goal is a sophisticated form

of transfer called cognitive flexibility, and it is a kind of learning that can only partially be

accomplished by traditional forms of professional training, which include traditional classroom

instruction combined with supervised practicum or internship placements. CFT is concerned with

creating instructional systems that maximize the effectiveness of limited classroom time by

making it possible for students to systematically encounter and study a larger, more selective

body of real-world cases that are tied to specific instructional goals, than would ordinarily be

encountered in the field in the first few years of practice. Similarly, Derry and Lesgold (1996)

argued that the full range of real-world opportunities to employ important foundational principles

that students learn in class would not be encountered repeatedly during limited field placement

time, or even during the first few years of actual work experience. Hence, there is need to design

and develop instructional systems that accelerate experience by providing students with guided

instruction in the context of a broader range of problems and cases than would otherwise be

possible. Derry and Lesgold cited evidence that systems-based classroom training with Sherlock,

an intelligent tutor for avionics test station troubleshooting that provided students with tutoring on

a library of representative cases, reduced learning time by several years relative to actual job

experience. The cases in Sherlock included representative troubleshooting problems, intelligent

guidance in helping students solve the problems, examples of expert problem solutions, and

opportunities and prompts that helped students use the expert solutions to reflect on their own

problem solving.

3
In CFT instructional systems students: 1. Examine many instantiations of foundational concepts

over multiple cases, usually video cases, representing real-world practice (Figure 1); and 2.

Practice interpreting complex (video) cases in alternative ways that require assembly and

reassembly of multiple domain concepts (Figure 2). While design strategies may vary, CFT

instructional systems use various methods for helping students see connections and interactions

among previously learned foundational concepts (in STEP, ideas from a learning sciences course)

and the real-world cases embodied in the instructional system (such as stories or parts of stories

of lessons in classrooms). Theoretically, such methods help speed students’ evolution of memory

potentials, enabling flexible and creative use of domain knowledge in new settings (Figure 3).

Briefly, in our approach to system design, we developed a web-based hypermedia network

(KWeb) composed of web pages of instruction in learning science concepts that are interlinked

with real-world cases of teaching and learning in K-12 classrooms that instantiate these concepts

in various contexts. Using this network in a CFT way involves devising instructional activities

that encourage students to reflectively explore the conceptual landscape represented by the

network (Spiro et al., 1991a/b), studying multiple cases that illustrate a range of instantiations of

key concepts and themes, and viewing individual cases and case segments (minicases) through

multiple, interacting conceptual lenses. For example, one type of activity we use to encourage

students to interact with the KWeb in this manner is a form of collaborative learning that

represents our own adaptation of problem-based learning (pbl, using lower-case letters to

distinguish our approach from PBL, a widely know instructional method developed by Barrows

(e.g., 1988)). “pbl” is a small group, student-centered from of instruction that is supported by a

suite of specially developed online tools that are part of the STEP site. Students receive a real-

world case to study and a related problem to solve (such as the case of an instructional unit that

must be redesigned to incorporate standards), and they investigate and learn subject-matter

knowledge in the context of discussing the case and design solutions. Students employ and

4
explore the STEP KWeb in their pbl activity and thus are encouraged to “criss-cross” the

theoretical landscape of the learning sciences. Students in our program also study CFT and thus

understand why they are exploring the landscape. Details follow.

ESTEPWeb.org in Three Parts

The main parts of eSTEPWeb.org are:

• Case Library: Video stories of real lessons in real classrooms representing the landscape

of practice

• The Theories Net: A “hypertextbook” containing conceptual knowledge from the

learning sciences and representing the conceptual landscape

• pbl online: Step-by-step scaffolding to help groups carry out authentic design challenges

and other problem tasks that a course manage might design that engage learners in case

analysis and discussion

The following will discuss design of these components and their theoretical psychological

grounding.

The Case Library: A Landscape of Practice

Learning through case study has recently surged in popularity within teacher education programs

(Merseth, 1996; Putman & Borko, 2000 -- add references from Lampert). Yet there are many

open research issues regarding case-based learning, particularly case-based learning with video,

including the question of what students learn from case study, how to facilitate such study, how to

design case-study exercises, and how to design both cases and case libraries to facilitate case-

based teaching and learning. CFT views the case library as representing the landscape of practice

5
and suggests testable hypotheses relating learning to the design and use of case libraries. As

previously discussed, one hypothesis is that systems-based classroom instruction can be designed

to accelerate development beyond what can be accomplished through traditional instruction

combined with field-based professional placements. The goal for such system design should be

cognitive flexibility -- the flexible, adaptive combining and use of domain concepts as

frameworks for interpretation and action in real professional settings – which can be attained by

combining two types of instructional activity: domain criss-crossings and small multiples.

Domain Criss-Crossings as an Instructional Strategy. This instructional strategy draws a

distinction between two kinds of domains: the domains of professional practice (such as

architectural engineering, medical practice, classroom instruction, etc.) versus subject-matter

domains that shape domains of practice (such as physics and art for architecture; biological

sciences for medical practice; learning psychology for teaching, etc.). It assumes that students

come to the instruction with prior knowledge of important, previously-taught (perhaps at a basic

level) subject-matter concepts or themes for which more enduring and sophisticated

understandings are required. This instructional approach guides them in studying a large number

of varied real-word cases that represent the landscape of contexts in which those themes or ideas

might be encountered in relevant domains of practice. In agreement with Wiggins and McTighe

(1998), we propose that the subject-matter themes for intensive study be a few carefully selected

concepts that are of central importance to the domain of practice. Also, the cases selected from

the domains of practice should represent the landscape of important types of contexts in which

those themes or ideas are likely to be encountered. The strategy is intended to help students

develop patterns of understanding representing varied ways in which practice is intertwined with

key subject-matter concepts. This basic idea of domain criss-crossings is illustrated in Figure 1.

6
Enabling this type of activity requires a flexible case library. Our developing online STEP library,

currently containing eight cases for five sub-matter disciplines at the time of this printing, is a

collection of stories about learning and instruction in actual classrooms. A case’s components

include approximately 15-20 minutes of edited video plus supplementary materials, such as

teacher commentary, examples of student work, class handouts, test scores, information about the

school, etc. Our cases range from examples of problematic classroom instruction that could be

improved (redesign cases) to examples of exemplary instruction we would like for students to

emulate (adaptation cases). Most cases lie somewhere in-between.

To enable thematic criss-crossings, our development process involves editing video footage in a

way that captures major to-be-taught themes in small segments of video called minicases. In

STEP, cases are stories representing a collection of such minicases; however, minicases can also

be extracted from their case contexts and reassembled in different groupings for instructional

purposes. For example, a lecture or other presentation for the STEP course might use an assembly

of multiple minicases, drawn from various case contexts, that illustrates a course topic, such as

social knowledge construction. Or, students redesigning an instructional unit on static electricity

can access the library and look at multiple minicases illustrating, for example, instances of

teaching for transfer.

Learning with the Cross-Crossing Strategy. What type of learning does the criss-crossing

activity afford? From the perspective of case-based reasoning (e.g., Kolodner & Guzdial, ), cases

(in this immediate context “cases” is broadly defined to include STEP cases and minicases)

represent possible templates for action and thought, old friends that, if learned, might later be

recalled and adapted and combined to guide action in a new situation. The belief is that new

events in the real world will trigger a “reaching backward” (Salomon & Perkins, 1989) to find

7
relevant cases in memory that can be recalled, combined, and applied to the new situation. From

this perspective, transfer occurs through a process in which helpful previously-learned cases are

recalled, combined and adapted to fit a new situation, a memory process that may resemble an

iterative analogical case-mapping process (case memories are mapped to the current situation,

case representations are adjusted through recombinings to improve the mapping, etc.).

The approach we are implementing for STEP is intended to create theme-based assemblages, or

idea “families” (e.g., Derry, 1996), of (mini)case memories, where the themes represent the most

important enduring understandings that instructors want a STEP course to promote. For example,

a fall course for secondary teacher education at UW-Madison emphasized five themes

recommended by the National Academy of Education as important areas of study for prospective

teachers: 1. Knowledge transfer; 2. The constructive nature of knowing; 3. The social nature of

learning; 4. Cognitive and sociocultural views of motivation; and 5. The importance of providing

students with the opportunity to learn and perform in environments where they receive feedback

and revise their work. The steps in the learning process can be outlined as follows: First, an

instructional episode (for example, a reading assignment and a lecture-discussion class on

“transfer”) creates, within students, some kind of embodied memory representation for one or

more themes. Next, through more advanced instruction involving domain criss-crossings using a

case library, the students’ thematic memory representations are enriched and updated with

exposure to cases that exemplify the theme; hence, for each student individually and for the group

of students in the class as a whole, the assemblage of ideas and related cases associated with a

theme increases in complexity and range, creating a family of memories that is more likely to be

activated in a greater range of future situations.

However, unlike Kolodner and Guzdial (2000), for example, we do not assume that over time and

given many experiences with a single theme in many case contexts, that memories of cases

8
encountered remain intact and unchanged. Students are not expected to recall all specifics of how

a teacher promotes (or fails to promote) transfer as illustrated in the many different cases

encountered. Our assumption is that both constructive and reconstructive memory processes

promote active memory change over time, leading to integration, abstraction, and loss of case

details in the memory representations of the case families that result from study within a CFT-

type system (not only during case-based learning, but also during later recall and use of case

families). We believe this process tends to strengthen memory and later perception for situational

aspects that tend to occur repeatedly across contexts and cases. In the Bartlerian (1932) tradition

we may use the term “schema” to describe these evolving, dynamic, distributed, situationally-

activated, thematically organized case-family memories, although we contrast this use of the term

schema with computer-science views of schemas as relatively static structures with parameters

and slots (e.g., Schank, 1982)

In summary, the kind of individual (and, where applicable, community) learning that the “criss-

crossing strategy” aims for is the creation of schematic attunements to the environments. These

are conceptualized as thematic families of situationally-activated perceptual attunements to

patterns, representing concepts from a subject domain, as they are experienced through study of

instructional cases representing a range of ways those patterns will be encountered within a

domain of practice. Here it is important to re-emphasize that the studied cases must be video

cases, since we make several important bridging assumptions related to the transferability of

learned attunements to real-world professional experience. One testable assumption, based on

psychological analyses showing significant overlap between the perceptual processes stimulated

while watching an activity on film compared to those stimulated during physical participation in

an activity (e.g., Gibson, 1979; Zacks & Tversky, 2001), as well as Barsalou’s (e.g., Goldstone &

Barsalou, 1998) research-supported theory concerning the role of perceptual simulators that

create conceptual understanding through mental enactment, is that embodied memories created

9
during video study using the criss-crossing strategy are likely to be activated later during real-

world professional experience. Such memory activations associated with both conscious and

unconscious, intuitive response to environmental patterns might include attitudes and other

emotional responses in addition to heightened probabilities of particular sorts of understandings

and propensities toward particular sorts of actions that have been observed in video. Hence, the

success of a criss-crossing treatment can be assessed in terms of two hypotheses: range of

transfer and acceleration. Both are tested when one demonstrates that, compared to mentored

field-based experiences of equal time and controlling for differences in students’ entering

thematic/conceptual subject knowledge, the criss-crossing approach increases the range of

transfer situations in which a thematic schema, and consequently the particular interpretations and

responses associated with a thematic schema, are appropriately activated.

The Small Multiples Strategy. The goal of STEP instruction is not merely increasing the

availability of thematic schemas that will be activated to serve as interpretive contexts for guiding

certain types or modes of experience and action, but also on creating new habits of mind

associated with enhancement of tendencies and abilities to view and consider the world

simultaneously from a broad range of multiple perspectives. That is, a goal of STEP instruction is

to enhance the schema assembly process that underlies interpretation of complex instructional

cases and ultimately real-world experience, making this process more fluid, flexible, creative,

complex, automatic and effortless. In Derry’s (1996) terms, this amounts to organizing

instructional experiences toward the goal of enhancing the array of “situational model schemas,”

as well as the speed of activation of such schemas, that a person or group is able to construct and

consider as a basis for understanding and interpreting cases within a domain such as classroom

teaching, whether the case is an instructional video story or a situation encountered as real-world

experience. The small multiples approach as deployed in STEP aims to enhance students’ ability

10
to create, discuss, judge, and recreate multiple situational models that represent valid, intelligent

case interpretations and reinterpretations based on learning science analyses.

To promote this ability, instruction should afford and scaffold prospective teachers in discussing

and otherwise studying cases (and minicases) such that the teachers are encouraged to examine

each case from multiple thematic perspectives. Depending on the sophistication of the students

and the difficulty of the subject matter, the multiples approach may begin with activities in which

students examine complex minicases, considering multiple themes sequentially. For example

students may view a case and then be asked to consider a short segment of that case (a minicase),

first as an example of knowledge construction, then as an example of classroom management, and

finally as an example of formative evaluation. From this strategy of sequential viewing in which

themes are considered one at a time, the instruction would proceed to a more advanced phase in

which students are required to construct more complex analyses in which alternative case

interpretations are constructed by combining and recombining multiple themes, and the resulting

analyses are discussed, compared and evaluated. The basic idea is represented in Figure 2.

To facilitate this process, Spiro (personal communication; Spiro et al., 2001) has proposed the

idea of using metaphorical experiential symbol systems (MESSes), the procedure of tying to-be-

learned themes to symbols or other perceptual enhancements, both auditory and visual, that are

then overlaid on instructional videocases for the purpose of directing students’ attention to the

conceptual themes and thematic interactions as they occur in action. In early learning stages, such

enhancements might help students more easily visualize or notice complex “messes” of ideas in

action. These enhancements could be faded as students gain more skill at identifying targeted

instructional themes within complex cases and interpretive exercises. Spiro has suggested that

such enhancements might be combined with special editing and system design effects to create

11
MTV-type case presentations that help compact and accelerate experience in advanced perceptual

learning, maximizing the benefits of limited instructional time.

Combining Criss-crossing and Small Multiple Treatments. Criss-crossing and multiples

approaches should be combined. A suitable instructional sequence might proceed as follows: 1.

Acquire some basic knowledge of foundational concepts through readings, lectures and other

“efficient” classroom activities. If thematic overlay treatments are to be used in later stages, this

initial instruction should also introduce students to the symbols associated with each theme. 2.

For each theme taught, criss-cross the domain landscape using the CFT system (in STEP, the

KWeb) to support instruction, supporting students as they examine multiple situational contexts

(cases) representing a range of ways in which each foundational theme is instantiated within the

professional domain of practice; 3. Again criss-cross the domain, choosing cases that combine

foundational concepts in various ways, and study these as instantiations of multiple theme sets.

Depending on the sophistication of students, this third step might be divided into two phases: In

the first phase (3a), students might practice interpreting cases, considering foundational themes

one at the time over several re-viewings of each single case. In the second phase (3b), students

are supported in interpreting and reinterpreting cases, integrating multiple themes in each

interpretation. In both phases 3a and 3b of the small multiples instruction, thematic overlays

might be employed initially to assist students in identifying presence of themes and thematic

interactions.

There are a variety of instructional practices that might afford this form of sequencing and type of

instruction. For example, students in a class might receive reading assignments supplemented by

lectures and discussions on the topics of transfer and knowledge construction, to help them

acquire foundational knowledge at a beginning level. Next, the class might together view an

instructional case (a fifteen-minute story of a classroom lesson), then focus on discussing selected

segments of the case (minicases), one at the time, considering for each how the topics of transfer

12
and knowledge construction are exemplified. Third, this activity might be followed by an

assignment requiring students to use the case library for a domain criss-crossing homework

assignment in which they are required to find and analyze 10 minicases illustrating the themes of

transfer and knowledge construction combined. Other approaches, including those that employ

problem-based-learning and contrasting cases will be discussed later in this chapter.

The Theories Web as The Conceptual Landscape

The theories section of KWeb, which is still growing and under constant construction, currently

consists of about 100 densely conceptually interlinked pages that contain explanations and other

instruction for selected learning sciences concepts related to teaching. These pages are

intertwined with about 100 video segments (minicases) in the case library (also growing) that

illustrate instances, uses, or illustrations of learning science ideas at work in the classroom. The

tool chosen (with the input of Chris Fassnacht) for developing and maintaining the system was Z-

Object Programming Environment (ZOPE), a powerful open-source system for supporting

complex web sites. This tool requires a pragmatic structural metaphor for setting up the site and

specifying relationships among theory pages and case nodes in the network. Such structures

impose constraints upon domain representation and are necessary for updating and maintaining a

website that must, as we describe below, constantly evolve in support of both instructional and

research goals. Again with Fassnacht’s input, the STEP team chose a loose family metaphor, such

that every page, or node, in the Theories Web is linked with other nodes in the web that are

designated as its ancestors, parents, children or “other relatives.” When a new node is added, only

its parents and relatives need be specified, then its position in the network with respect to all other

nodes is automatically generated, and any content that is input to new pages is automatically

formatted.

13
The Theories Web is organized into three interlinked idea families: Cognitive Theory;

Sociocultural Theory; and Cross-Theory Ideas. The Cognitive Theory family has two main

branches: Information Processing and Sociocognitive (Developmental) Theory. Each idea family

and its main branches also have several major family branches. This structure has a hierarchical

feel at its highest levels, seemingly in conflict with the basic tenets of CFT, which opposes

reliance on overly simplified structures during advanced instruction. Yet making such structural

context visible to users is a navigational aid that facilitates introductory learning and helps

prevent disorientation in hyperspace (Dias, Gomes & Correia, 1999). Thus, from any entry point

within the Theories section of STEP KWeb, the user can identify the major theory and idea

families to which that page belongs.

While the theory family relationships metaphor is an apparent hierarchy at superordinate levels, it

is a loose family metaphor that affords creation of partially non-directional graph structures

specifying complex interrelationships among pages (there can be three or more parents, for

example, and relational links can be either directional or non-directional). This structure is

capable of representing complex, conceptually valid and instructionally useful relationships that

permit meaningful navigation regardless of where in the network -- from which conceptual

vantage point -- the user explores the system. From any page in the network, any point of entry, a

context is created whereby the user can move easily to pages discussing higher-order (parent)

ideas, another conceptually related page of equal status in the same or any different family

(relatives), or a page on a more specific derivative idea of the current concept (child nodes). In

addition, links within the page encourage the user to access video minicases that can be partially

interpreted in terms of concepts discussed on the user’s current page. An example Theories page

is shown as Figure 4.

14
Designing the conceptual structure for the theories network originally proceeded from a rational

historical and conceptual analysis of the learning sciences domain, without, at first, giving deep

consideration to the structure of the domains of professional practice. However, because web sites

are highly adaptable and potentially widely used, various use contexts have provided natural

mechanisms for ongoing domain analyses. Such analyses are shaped in interesting ways by the

constraints and affordances of the organizing metaphor, the needs of courses and instructors that

the web site supports, and the structure and capabilities of the web-based technology itself.

For example, the content of the Theories section has been influenced recently by the teacher

education courses it supports at Rutgers University and the University of Wisconsin-Madison,

which have chosen to emphasize five key learning science concepts promoted by the National

Academy of Education. Previously mentioned, these are: Transfer, Knowledge Construction,

Social Learning, Student Motivation, and Formative Assessment. Although most of these

concepts were already embedded as pages at appropriate points within the existing network, they

are gradually being expanded through the addition of child and relative nodes, and new cases are

being selected and edited and added to the case library to support these conceptual emphases.

Another example of how use context shapes web content is illustrated by the recent adaptation of

web content to serve learning science courses for kinesiology and music. These majors have

spawned development of pages for concepts such as expertise, giftedness and talent.

How the technology itself shapes conceptual analysis is illustrated by our rule of thumb for

minicase size: A video less than two minutes long can be downloaded quickly with a fast

connection and, if downgraded in quality, with a 56K modem. Hence, the two-minute (or less)

minicase was born. Yet our experience indicates that this is a sufficiently large grain size for

illustrating meaningful and complex instructional events that require complex interpretations

using multiple concepts. Video segments of two minutes or less nicely support conceptual

15
presentations of learning science concepts in various instructional formats, including lecture,

face-to-face, online case discussions in small groups, and individual study using KWeb.

Clearly, then, it is not necessarily the top-level hierarchical structure of the network that makes

some ideas structurally (Meyer, 1975) and instructionally more important than others. Rather,

instructional importance of ideas is largely determined by frequency of links to pages that are

created as the system naturally and dynamically evolves through development and use. It is also

created by frequency of visits students make in specific instructional contexts. For example, both

directed (by assignments and activities) and undirected (student explorations) searches through

the net frequently encounter the NAE nodes. They are, as a result, highly visited nodes and in

many ways structurally more important than the so-called superordinate concepts that provide

overall organization at higher levels of the network.

In sum, domain analysis has proceeded simultaneously and been shaped by system development

and use and by development of cases. This process has re-taught us first hand that learning

sciences in instructional contexts is in fact an ill-structured domain. One source of messiness is

theoretical intermingling: similar but subtly different ideas with similar names are claimed by

different theories. A good example is the term “constructivism,” which is uncomfortably shared

by the sociocultural, sociocognitive, and information processing theories and is used in reference

to instructional methods and theories of learning. Also, there is considerable historical

confounding of theoretical knowledge. For example, the body of knowledge that is often taught

and represented as information processing theory presents concepts derived from early versions

of the theory side-by-side with those derived from later versions. Hence, from the student’s

perspective, one theory may seem to be explaining phenomena in very different ways, a situation

that frequently leads to serious misconceptions that must be challenged, such as the idea that

information processing psychology is synonymous with a behavioral or passive “knowledge-

16
transmission” perspective. Another source of messiness was the absence of good analyses and

basic scientific knowledge about the relationships between domains of real-world professional

practice and the theories that are traditionally taught in professional schools with the intent that

they be “transferred” as interpretive frameworks for practice. Such observations simply confirm

one premise of CFT, that many professional domains are messy and ill-structured and that the

organizational schemes imposed upon them for instructional convenience can often be

characterized as efforts to tame through over-simplified structural metaphors that do not map onto

real-world practice.

Instruction with the Theories Web

The Theories part of the STEP KWeb affords many instructional formats, including those

consistent with advanced CFT instruction for ill-structured domains. For example, when students

are first introduced to a learning science concept, such as knowledge transfer, the theories page

that introduces and discusses that concept can be assigned as reading. When students access the

assigned page, the format of that page invites and encourages them to explore other parts of the

STEP KWeb in a CFT way. As shown in Figure 4, from the assigned page they are able to access

multiple video minicases that offer classroom examples of that concept (criss-crossing the

landscape of the instructional domain for examples of transfer). Or, they are able to move from

the assigned concept page (transfer) to the parent idea(s) for that concept (e.g., prior knowledge

use), to the child ideas of that concept (e.g., analogy and cognitive flexibility), to related ideas

(e.g., knowledge construction, adaptive expertise, etc.), or to various minicases representing the

concept in action. Each page invites and enables further movement through the STEP knowledge

web and beyond, inviting criss-crossing of the conceptual landscape of the “learning sciences.”

Taken together, the Case Library and Theories sections of STEP KWeb offer two different types

of entry points for supporting two different forms of instruction: Theory-based and case-based.

Because minicases can be accessed, viewed and studied in (or out of) sequence within the context

17
of the larger narrative cases of which they are a part, the instructor can select and assign cases or

minicases and organize instruction from a case-based perspective. In STEP KWeb, links between

concept pages and cases are reciprocal, so that every case also points to the concept pages within

the knowledge web that can be used to help students analyze and think about cases. However, this

is a pointing toward a possible subset of useful ideas for thinking about cases, not an explanation

of the case. Students who use links to guide their study and thinking about cases are still required

to construct self-explanations for why a particular set of concepts applies to a particular case.

Thus the system scaffolds but does not supply case analyses for its users, and is designed to

prompt questioning and discussion about why certain links apply to certain cases, and about other

links that might also apply and provide additional valid viewpoints. In this sense the system

supports small multiples instruction. As discussed in the next section, an additional way in which

small multiples instruction is supported in the site is through guided case discussion: An online

environment for guiding case analyses in pre-service courses has been incorporated into the

system and is described in a later section of this chapter. First we turn to several interesting

theoretical issues regarding the bringing together of the domain of practice, represented by the

case library, and the conceptual theoretical world of the learning sciences, represented by the

Theories section of the knowledge web.

Interaction of Case Library and Theories Web

A common goal of professional education is to help learners acquire interpretive conceptual

systems (e.g., the learning sciences) that can flexibly be assembled and imposed as interpretive

lenses upon practice (e.g., classroom teaching) in order guide and inform it. And issue that has

not been addressed in previous literature is how the events of professional practice are naturally

segmented and structured by the physical world, including the physiological systems of human

event perception. This issue is important, for event perception in professional environments is a

cognitive process influenced by both the “top-down” conceptual systems acquired through formal

18
and informal educational experience, and by the “bottom-up” mechanisms of human perception

that have evolved to reflect how humans interact with events in situ. Because top-down and

bottom-up processes constrain and support one another during event perception and

comprehension, instructional designers concerned with transfer of conceptual knowledge to

domains of practice should consider how humans connect with their environment through the

meshing of perceptual patterns triggered by the environment and those triggered by conceptual

memories that are learned through courses and other experiences and presumably recalled and

used during practice.

The classroom landscape teams with hundreds of complexly interwoven events occurring

simultaneously in three-dimensional space and continuously over time. The teacher at the center

must refine, select, edit and otherwise impose order on what would otherwise be an

overwhelmingly complex experience. Fortunately there are features of the physical world that

form patterns and gestalts, and human perceptual systems have evolved means of taking

advantage of them. Zacks and Tversky (2001) synthesized a large body of literature concerned

with perception of events in a dynamic environment viewed as patterns of cognitive activation

that are partially generated in response to correlated physical input received from the

environment. Events as perceived tend to be bounded by certain objective physical properties,

such as change in an actor’s behavior or direction of movements, and thus have beginnings and

endings that are reliably detected across observers. These perceived patterns very often

correspond to chunks of activity representing systems of physical causality: For example, an

agent (e.g., the teacher) acts on objects (e.g., writes on a transparency) with particular, usually

multiple, physical consequences (e.g., a display appears and students stop talking and look toward

the display). There is evidence that such episodes are perceived as bounded objects that have

parts (agent, action, ground, direction) and are located within hierarchical event structures (Zacks

& Tversky, 2002). For example, the transparency event is part of a larger daily activity that is a

19
part of a lesson that has particular event outcomes in the context of a larger instructional unit and

course, all associated with learning goals and expected outcomes for students.

Although event structures are perceptually cued by objective physical properties and are often

individuated on the basis of physical properties of motion and causality, they are also inferential

constructions cued by prior knowledge. Teachers interpret events and give them meaning,

drawing inferences about classroom plots, students’ understandings and motives, problem solving

processes, subgoals, goals, long-term ambitions, expectations of principals and parents, etc. Not

surprisingly, the ordinary language that people use to talk about, and draw inferences regarding

the meaning of events, also, like the physical properties of events, has hierarchical causal

structure. Zacks and Tversky (2001) suggest that the basic syntactic and semantic structures of all

languages reflect the hierarchical and causal nature of events. Both informal and professional

discourse about practice reflects a universal human tendency to impose structure, including causal

and hierarchical order, upon event perception.

This fact has several implications for professional education. One implication is that

practitioners’ comfort with a new conceptual discourse such as the learning sciences, which is

intended to provide top-down constraints on teachers’ classroom event perception and

interpretation, will likely be tied to how realistically instruction is able to map that language onto

naturally occurring event structures in the world of work. Those involved in teacher education are

well aware that teacher learners often expect and gain little from the theories to which their

curricula expose them (e.g., Simon, 1992). One source of that resistance may derive from the fact

that scientific theory talk is too far removed from the events of classroom practice, either because

the foundations of educational science were not historically grounded in classroom practice in the

first place, or because even grounded theories have evolved away from practice through out-of-

classroom professional discourse (e.g., talks and discussions at academic conferences, in journals,

20
in university classrooms). If theory isn’t well matched to student teachers’ perceptions of the

kinds of events it is supposed to explain, it won’t be bought, especially if teacher-learners are

immersed in classroom experiences while studying. Is it surprising that teacher’s everyday

practical talk about their practice may seem to them more informative than scientific theory?

However, an event that can informally be described as the teacher talking and students

responding by listening can also be seen as the teacher talking in a way that uses instructional

tools and artifacts that help direct students’ attention to what is most important for them to

process. An instructional objective is characterized, not merely in everyday terms as a goal to

cover material or pass a test, but as a goal of helping students construct understanding of

important concepts in a way that will help them transfer that understanding into new situations

within communities of practice outside of school. These sentences are examples of using learning

sciences terms (underlined) in a way that expands and deepens event perception in a useful way.

And teachers who understand these terms in depth can unpack them further for even deeper

insights into the causal nature of instructional goals and actions, including hidden “thinking”

actions that facilitate goal achievement. For example, knowledge construction is a process of

helping activate students’ relevant prior knowledge and use it to interpret new experience. In sum,

learning sciences (and other professional discourses) can and should be taught in ways that

expand natural language and perception, not work in conflict with it. Thus, the selection of what

concepts to teach teachers should honor both the causal and hierarchical aspect of natural event

perception in classrooms, as well as the causal and hierarchical aspect of what the science of

learning indicates is good instructional activity.

The goals of teaching learning science through video case study include expanding and deepening

teachers’ perceptions of practice, meshing practice with theory, and promoting flexible use of

learning science concepts in considering multiple interpretations and hypotheses about how to

21
achieve learning goals. An event-perception analysis allows us to understand more precisely what

this means and how to do it. Such an analysis has implications for selecting and presenting the

content of the conceptual domain (e.g., learning science) curriculum, the design and structuring of

cases of teaching practice to which that conceptual domain is mapped, and the design of the

instructional activities that mesh the domain of practice with the conceptual domain.

Meshing Practice with a Conceptual Domain

Meshing conceptual domains with practice is not a straightforward matter, for both the world and

the perceptual systems that see it are complex, dynamic systems, never at rest. Even while the

human perceptual system strives to impose order on a complex and changing world, that order

itself is an emergent and shifting phenomenon, so no particular and predictable pre-fabricated

order can be learned and imposed. At any given moment in the classroom, multiple events are

occurring simultaneously, and the teacher is selecting actions, from among many possibilities,

with multiple goals in mind (Lampert, 2001). These actions have consequences that help mark

each event as a unit of perception, but there are both intended and unintended consequences,

planned and serendipitous successes, and failures. Together, the environment and the human

perceptual system shape the happening, episodic world into shifting patterns representing

temporally-bounded, goal-oriented, hierarchically-structured and nested activity systems. These

patterns impose structure and constrain what would otherwise be too much interactive confusion

to manage.

Instructional Routines as Goal-Based Events

In this sea of complexity, the ability to perceive causal mechanisms is helpful and should be

taught and developed in teacher preparation programs. Classroom events, and perceptions of

events, influence and are influenced by teachers’ and students’ goals and the actions they initiate

in pursuit of those goals. One type of goal-based system that can help focus perception and

facilitate classroom management is the instructional routine. As documented by Leinhardt &

22
Greeno (1991) the instructional routine is a socially scripted schema for guiding classroom

activity so that important goals are achieved. Routines are normative classroom procedures, often

designed and shared by teachers, who adapt them to their purposes, that are taught to students and

that represent standardized practices that both teachers and students understand. After a

classroom community acquires them, routines manifest themselves in daily activity with little

need for verbal communication regarding what is expected. Pinups and gallery walks are

examples of types of routines frequently used in design-based classrooms, in which students learn

subject matter by designing and/or creating artifacts (e.g., Kolodner & Guzdial, 2000). During

pinups and gallery walks, student design teams display work in various stages of completion and

take turns explaining their ideas and viewing and commenting on others’ work. Another example

is scripting during collaborative learning (e.g., O’Donnell, 1999). Capable, experienced teachers

have perfected many detailed routines (Leinhardt & Greeno, 1991) and use them frequently,

adaptively and interchangeably to accomplish a variety of learning and assessment goals. As

learning sciences are incorporated into teacher education, they should help teachers understand

how to design and manage instructional environments, and their perceptions of them, through

instructional routines that promote multiple learning outcomes for multiple students.

Instructional Planning

Routines are “plugins” for flexible instructional planning. Teachers who enter the classroom with

well-conceived plans and a repertoire of routines for implementing them are oriented toward

particular learning goals and goal-related activities for students, toward seeing evidence regarding

goal achievement, and toward perceiving events as things that contribute to or interfere with goal

achievement. However, constructing an environment to achieve even one goal (much less

multiple ones, as required) is a non-determinant task, a matter of assembling a satisficing plan

from among many possible plan components (e.g., Simon, 1996). Also, plans are not writ in

stone; they must be adapted in practice and hence place only soft, anticipatory constraints on

event perception and development (e.g., Suchman, 1986). An unpredictable outcome for one or

23
more students may require a rethinking of next steps. A discipline problem may arise and cause

temporary suspension of the current goals and plans. Yet long-range and short-range instructional

planning helps pull order from a potentially unmanageable chaotic sea, placing (heavenly)

constraints on how events are chunked and viewed in classrooms.

An important role for learning sciences, then, is to furnish a rich language of cause and effect, of

goal-based activity structures, for instructional planning. That is, the learning science knowledge

taught in education courses should be that which helps define goals and the tendency and ability

to see actions that lead to or thwart those goals. For example, the goal of achieving an “enduring

understanding” (e.g., Wiggins and TcTighe, 1998) of satire or proportional reasoning or

photosynthesis (etc.) must be considered simultaneously with learning science topics that shed

light on issues of what varied combinations of classroom management, problem design,

instructional materials, forms of discourse, instructional routines and assessment, etc. are possible

within existing institutional, physical and social constraints and can be brought together in a plan

to help afford and cause the signs of understanding. Thus, preparing to teach requires learning the

skill of adaptive plan construction, in which goals and assessments and activities as routines are

flexibly assembled and reassembled in the process of achieving a satisficing instructional design.

This is an important form of cognitive flexibility, and it is the goal of the instruction that takes

place in the STEP pbl environment, described next.

The STEP pbl System

The STEP pbl system scaffolds instruction in which small groups of students deepen their

understanding of learning sciences as they work together on instructional design problems. The

system permits a course manager to create and modify problems in various ways, so the example

provided below merely illustrates a characteristic instantiation representing one way that

instruction in the STEP pbl environment has proceeded.

24
Problem based learning (PBL) is an instructional method developed for medical education (e.g.,

Barrows 1989) that has made its way into many other types of classrooms, including K-12

classrooms. The purpose of PBL is to help students acquire domain knowledge, usually scientific

knowledge, in the context of a solving a real-world problem that is based on a real-world case.

What we know about this method is largely gleaned from wisdom of practice, both from its

developer, Howard Barrows, and from the widespread PBL community that uses this method and

has begun to conduct research on it. In this community there is a standard PBL procedure, and

some members feel strongly that it shouldn’t be messed with. This procedure takes students

through a facilitated small-group, student-centered process in which students discuss and “solve”

a problem case (e.g., a case of medical diagnosis) as they fill out a whiteboard that has been

structured to facilitate their inquiries. In filling out the whiteboard, students proceed through

stages in which they observe facts in the case, formulate hypotheses, identify learning issues for

further investigation, conduct research, and revisit and discuss hypotheses until a problem

solution is reached. The main purpose of this activity is to learn about a conceptual scientific

domain that underlies a real-life case.

Several years ago, PBL was introduced into the STEP course as a face-to-face small-group

method that was practiced once a week in the classroom. It was introduced simultaneously with

the STEP KWeb (which did not at that time have a site supporting online problem-based

learning). PBL presented cases of instruction to be improved, and the instructional design projects

lasted several weeks. Between classes, students conducted research using the STEP web site and

other resources. One group that we have studied and reported on previously (e.g., Derry,

Seymour, Feltovich & Fassnacht, 2001; Derry, Seymour, Steinkuehler, Lee, & Siegel, in press)

consisted of five science education majors and their tutor. Their four-week task was to use

25
learning science concepts to help them redesign a traditional science unit on static electricity that

was based on video and other case materials available in the STEP KWeb.

Our studies of this group and our observations of the STEP course as a whole led us to several

conclusions. First, students solving a case together in a PBL format do examine and argue the

case from many points of view, and their arguments incorporate learning science concepts from

our course as well as life experiences and ideas from other courses. And, some of the same

learning science ideas do resurface in multiple cases over time, giving students chances to see

multiple examples of key concepts in action. So PBL does in fact help our course meet some

important conditions of CFT.

Second, students do appear to learn from the course experience. A preliminary evaluation by

Siegel (2002) and another by Stampen (unpublished) provided evidence that students may

increase their tendency to use learning science concepts when viewing instruction, activate more

of the concepts that experts think are relevant to cases, and from the learning science perspective,

build more sophisticated situation models and theories to describe what is happening in videos of

classrooms. Whether it improves them as teachers we cannot say with certainty, although students

themselves believe the program is relevant to their careers.

A third observation is that PBL is a difficult instructional method, unlike any that most people

have experienced. The facilitation of PBL is important to its success and the unseasoned TA’s

who served as facilitators for PBL groups struggle with it. Student teachers are often initially

resistant to the unfamiliar method, and there are larger institutional and program contexts that

make it likely that conflicts will arise (Derry et al., in press). Also, PBL is resource intensive

since on facilitator is required for every 7-10 students. In large courses, tutors must monitor

26
multiple groups. This is a substantial burden for TA’s, but we felt this was a problem that might

be addressed through technology.

Hence we created and added to our website the STEP pbl system for online support of problem

based learning. We use lower-case letters (pbl) to distinguish our version from the standard

approach that has evolved from Barrow’s work, since, through use and experience, we have

evolved and changed the approach significantly. The STEP pbl system is a collaborative

environment that is integrated with other eSTEPWeb.org resources. We decided that pbl should

go on line for a number of reasons (Steinkuehler, Derry, Hmelo-Silver, & DelMarcelle, 2001).

First, we knew that an online system could be designed to distribute some of the tutoring

responsibility to the system and the students themselves, lightening the responsibility for tutors.

Second, the STEP community is growing, and an online system facilitates larger course

management. Finally, an online system would enable us to eventually offer the course as distance

education, making the site more scalable for professional development.

The STEP pbl system supports either online small-group instruction or a hybrid model in which

students meet face-to-face in small groups during class and then extend their work outside of

class through online interaction. In both online and hybrid models, students are guided by a

human facilitator, typically a teaching assistant, and are required to complete and submit

individual and group artifacts, products related to and documenting various stages of instructional

design, through the online system. The online system collects and displays data on student

performance and affords detailed monitoring of work by individuals and small groups, permitting

detailed (and powerful) formative assessment of individuals and groups throughout the course.

This is both a bane and benefit of online courses, for this monitoring capability makes them

powerful learning tools that place substantial performance demands on both students and

instructors (e.g., O’Donnell, 2002).

27
Although they can be set up in different ways by the course manager, STEP instructional design

activities in pbl typically involve a phase of individual study and preparation, followed by a phase

of facilitated small-group design work, followed by a final phase in which the individual

analyzes, extends, and reflects upon the group’s work and how much the individual gained from

it. These phases are scaffolded online by the STEP system, which guides students through a series

of steps. The number of steps and required activities for each step may vary from problem to

problem, as desired by the course manager/designer. Here we describe one of the activities

created for the fall, 2002, course at UW Madison, which took students through a four-week, nine-

step design challenge. All students in this course completed two pbl activities on line. The

example to be described is the second activity that was completed by the English majors. Similar

design problems were created for students in secondary science, mathematics, social studies, and

foreign language.

When students entered the STEP pbl system to start their design challenge, they saw a “sidewalk”

with nine steps, each step associated with a particular due date for completion (insert Fig 6 about

here). Each student began the task by mouse-clicking on step 1, which opened a page of

instructions and a design problem appropriate to that student’s academic teacher certification

area. The design problem for English majors, Huckleberry Finn at Midwest High, is shown in

Figure 7. Previously, each student had been assigned to a small discipline-based work group that

was maintained in this problem activity and throughout the course.

Like design problems for other disciplines, the English problem referenced and linked to a

particular video case in the STEP video library, a classroom story that students were asked to

analyze in preparation for their design work. Students were asked to draw lessons and ideas from

the case under study and then apply those lessons and ideas by working with their group to design

28
or redesign a similar type of instruction. All design problems for all disciplines required students

to warrant their instructional designs through learning science research. This research was

facilitated by availability of the STEP KWeb and other online research resources integrated with

the STEP pbl system.

Pbl activities required students to apply a process of “backward design” leading to the creation of

a “group product,” a plan for an instructional unit. The unit to be developed in the English

teachers’ groups (see their problem) was to employ a controversial text of the students’ choice

and address instructional objectives related to themes of equity and diversity (similar to the case

the students were required to study), as well as literary topics, such as satire. Students learned

about backward design through readings in the KWeb and an assigned text, Understanding by

Design (Wiggins & McTighe, 1998). Adhering to the steps in their pbl “sidewalk,” students first

completed their reading assignments and studied their case, entering their thoughts and reflections

about the case into an online notebook (see Figure 6). After completing initial assignments by the

date due, students were at Step 4, where they joined their group of 4-5 other students and began

group design work. Groups were allowed to choose whether to work online at all times or

whether to supplement online work with face-to-face meetings during class. Most groups

continued to meet face to face.

During group design, steps 4-6, the pre-service teachers were scaffolded though a process in

which they first carefully considered what “enduring understandings” their unit would teach.

Next, they developed ideas for how they would assess their students, to determine whether goals

for understanding were being acquired. Finally, they worked together to design goal-related

activities. Each step in the process involved submission and discussion of the various teacher-

learners’ ideas for goals, assessments, and activities. Ideas were refined online through discussion

and voting, with ideas receiving strongest group support becoming part of the final group product.

29
Group activity was supported online by a group whiteboard (Figure 8) and a supplementary

discussion board. As shown in Figure 8, the whiteboard contained sections (marked by “tabs”) for

each stage of the groups’ work. For example, during the design of assessments the group

members were in the assessment section of the group whiteboard. During each major phase of the

group activity, such as the assessment or activities design phases, students entered their

“proposals” for what the group’s design should include, plus a learning-science justification for

their proposals, onto the group whiteboard. Students also used the group whiteboard to view and

comment on others’ proposals and justifications, read comments about their own proposals, and

modify their own proposals in response to group feedback. Students controlled what the system

put into their final product with a voting mechanism through which proposals receiving group

support were included.

Thus, the group whiteboard “forces” students to design an instructional unit, thinking about

assessments and activities in a certain order and in terms of how they would lead to enduring

understandings. Thus the group whiteboard represented the course manager’s epistemological

commitments (e.g., Suthers, 1999) regarding what kinds of goals, knowledge and evidence the

course manager wanted students to consider and discuss during learning. Parenthetically it is

noted that the group whiteboard in STEP pbl is a general tool that allows course managers to

change these commitments from problem to problem and course to course, by altering the number

of tabs, tab headings, and instructions to learners within each tab.

Upon completing the group product -- a justified plan for an instructional unit specifying goals,

assessments and activities -- individual students completed steps 7-9 individually. In step 7

individual students wrote their own critique and analysis of the group product. In step 8 students

reflected on their learning, and in step 9 provided anonymous feedback on the activity and site. A

30
TA facilitated all steps online. Each TA in the fall course managed four small groups of about

five or six students each.

Students Ratings of STEP pbl Activities and Site Tools

There were two instructional design pbl activities in the fall course, similar to the one described

above. Based on a class size of 60 and a response rate of about 97%, the following ratings of

components of the two pbl activities (Table 1), and the system tools used during the pbl activities

(Table 2), indicate that while there is room for improving the design of STEP pbl, the activities

were valued and generally well received. Several patterns in these responses can be observed.

First, from pbl-1 to pbl-2, there was a substantial increase in student satisfaction. This is likely

due to a number of factors, including student and TA experience with the method and system, as

well as the use of discipline-specific cases and problems in pbl-2 (rather than the generic

instructional design scenario grounded in a case on design-based instruction in a science class,

which was employed in pbl-1). Second, a decrease in ratings for TAs likely reflects a deliberate

fading of scaffolding. Since TAs were trying to decrease their involvement in student work, it is

not surprising that students perceived their input as less important. Finally, it is notable that the

most rewarding activities for students were those involving collaboration rather than

individualized work. Also, a highly rated tool was the hypertext information resource, the STEP

KWeb.

Table 1. Students’ Ratings of Phases of Two STEP pbl Activities

“How much did you learn from the following activities?”

Scale: 1 (Nothing) – 5 (A lot)

Activity pbl 1 Rating pbl 2 Rating

Overall pbl activity 3.8 4.35

Initial Proposal 3.35 3.70

31
 View others’ proposals 4.24 4.33

Establish goals 3.84 4.19

Group Design 3.94 4.38

Final group product 3.73 4.08

Individual explanation 3.91 3.84

Reflection 3.72 3.88

Interaction with group 4.28 4.37

Interaction with tutor 3.06 2.87

How well did the following tools work?

Scale: 1 (Very poorly) – 5 (Very well)
Cases 3.58 4.02

Case related concepts 3.75 4.22

Notebook – Initial proposals 3.64 4.05

Notebook – Research notes 3.66 4.02

Notebook – Individual evaluation 3.91 4.16

Notebook -- Reflections 3.84 4.18

Group Whiteboard 3.75 4.16

Group Discussion 3.65 4.07

Research Library 3.58 4.21

Knowledge Web (hypertextbook) 4.29 4.50

Pbl Help 3.27 3.45

Worked Examples 3.07 3.46

The following are some characteristic quotes from students, taken from their reflections about the

experience:

32
 . . . this lesson that we have designed as a group is definitely something I could see

myself using down the road when I have my own classroom. I feel it is a well thought

out lesson that can be easily modified to meet the needs of whatever type of class

“make-up” that I may have.

I will attempt to use this method when creating lessons plans for next semester. I

think it is a valid model that helps the teacher keep objectives clear and plan

meaningful activities which cater to the objectives.

I would use the unit itself. It was a good final product. I would also use this method

of creating lesson plans.

The plan that we made up as a group will be something that will be extremely useful

for me as a teacher. I also learned the value of input from others’ viewpoints on the

same unit because you are able to see different perspectives that can give you some

new and different ideas.

With future refinement, we are confident that online design activities in the STEP pbl site will

prove to be an important and successful addition to current approaches to teacher preparation.

Future Directions

Research with eSTEPWeb is proceeding in two general directions: Experimental and Evaluative.

In the experimental program we are conducting controlled studies that closely examine learning

and learning processes promoted by alternative instructional methods that can be set up,

supported, and tracked through the web site. One interesting contrast to the CFT instructional

33
approaches, discussed earlier, is suggested in recent work by Schwartz and Bransford (1998),

which implies that novices should participate in detailed study of contrasting cases in the domain

of practice prior to introducing them to related topics within the conceptual domain. Their

findings show that contrasting cases treatments prepare students for future meaningful learning

from direct instruction. “Knowledge differentiation” was the learning mechanism Schwartz and

Bransford proposed as responsible for the mental preparation. In STEP, the analog of the

Schwartz and Bransford treatment involves using the pbl environment to set up controlled, in-

depth comparative examinations of small parts of the landscape of practice (e.g., contrasting the

minicases on assessment or comprehension, as examples). The analog argument for effectiveness

is that giving contrasting-cases pbl exercises before providing lectures or assigning KWeb pages

on conceptual topics, would enhance novice teachers’ ability to learn ideas that relate to that

practice because it would give them a more differentiated knowledge structure to build upon. This

approach can be contrasted with the one based on CFT theory described earlier, the approach of

providing students with basic themes in the conceptual domain though direct instruction with

symbol systems, followed by exploration within the domain of practice through video study

involving criss-crossing and small multiples treatments enhanced by metaphorical experiential

symbol system (MESS) overlays. In fact, given the flexible nature of the human cognitive system,

it is likely that both approaches, if perfected, will be successful. But they might be successful in

different ways, enabling the testing of theory. And, given different contexts and learners, one

might be more feasible than the other to implement. The important point for now is that the STEP

system can support in-depth exploration and detailed comparisons of these and other related

instructional approaches, which will enable us to perfect them and uncover important details

about their underlying cognitive mechanisms.

Evaluative Research

Our other line of research is more evaluative, informed by design experiments conducted in situ

as courses are offered through STEP. Assessments in these contexts require a framing definition

34
of what each STEP course is attempting to accomplish. In this regard, it is helpful to discriminate

between a. the goal of developing teachers as thinkers about and planners of instructional

environments, versus b. that of also developing their ability to manage and productively interact

with students within the classroom environments they design or adapt from externally imposed

curricula. This distinction is made because, in the context of pre-service teacher education, it

more feasible to focus on the first goal, since pre-service teachers often do not have sufficient

control over the instructional environments in which they observe and practice teach and thus

may not be able to implement their designs. In these courses, therefore, the primary goals relate to

creating competent instructional planners and analyzers of instructional plans and practices who

flexibly and reflectively use learning sciences to inform their designs and analyses while still

operating within constraints imposed by local standards and curricula. Assessments for these

criteria include judging the course in terms of whether it improves students’ abilities to reflect on

and analyze the teaching they observe in video cases, as well as their ability to craft good

instructional plans that specify goals, assessments and activities that are warranted by learning

science and reflect a usable understanding of it.

In actual professional practice it is impossible to draw such a clear-cut distinction between design

as planning and design as implementation. Teachers’ instructional plans presuppose knowledge of

and commitment to certain forms of classroom interaction. And, the instructional design cycle is

iterative, so knowledge drawn from interactive experience within a certain design should lead to

teacher learning, modifications and improvements to that design. Clearly, planning and

implementation knowledge interact with one another and are important, mutually implicative

aspects of teaching as a design “science.” In extending STEP to in-service professional

development, which we are in the process of doing, we want teacher professional development

courses in STEP to focus on broader, more ambitious objectives. Not only do we wish to improve

in-service teachers’ abilities to thoughtfully plan and design good learning environments, given

35
standards and other social, institutional, physical and curricular constraints, but we also want to

help them operate intelligently within them “on the fly,” and collect and use data from

implementations to iteratively improve design and classroom interactions.

Practice as constraint-based cognitive flexibility

A recurring theme in both pre-service and in-service professional development is adaptive

flexibility within constraints, based on the insight that cognitive flexibility, including flexible use

of the learning sciences, occurs within a complex, constraint-based classroom system. Constraint-

based cognitive flexibility includes the ability to flexibly assemble satisficing plans to achieve

goals even in difficult environments, and to demonstrate an ability to adapt those plans as events

unfold. This requires a rapid, intuitive, flexible “seeing” of the classroom landscape at different

stages in instructional design and implementation and at multiple levels of abstraction and

specificity.

For example, in challenging circumstances, such as when discipline problems occur in the

classroom, the teacher’s event perception must switch temporarily from higher-level plans related

to long-term instructional goals, to the immediate concerns of maintaining classroom order, but

without losing site of the long-term goal. The relationship between immediate actions and long-

term goals is also important. For example, a successful inquiry science classroom is not an

accident but the result of weeks, perhaps months of preparing students to work productively in

learning communities with appropriate routines and norms that structure group activity even

when the teacher’s gaze is averted. Thus planning requires that teachers’ perception switch

flexibly between visions of what is possible later and what must be done first (and second and

third) to set that up. The ability to adaptively see and causally connect events at multiple levels of

abstraction is yet another form of cognitive flexibility required for teaching. Assessments of

teachers and teacher development programs should include assessments based on the concept of

constraint-based cognitive flexibility, requiring teachers to demonstrate competence in planning

36
and plan adaptation in a range of environmental conditions. The primary focus of professional

development programs and systems should be to help teachers develop constraint-based cognitive

flexibility grounded in scientific understanding of the learning sciences, an ambitious form of

transfer. We believe that STEP illustrates the kind of research-based development project that

will eventually make possible this type of professional development goal and programming.

References

Barrows, H. (1988). The tutorial process. Springfield IL: Southern Illinois University
Press.

Bartlett, F. C. (1932). Remembering: A study in experimental and social psychology.

New York: Cambridge University press.

Derry, S. J. (1996). Cognitive schema theory and the constructivist debate. Educational
Psychologist 31,163-17

Derry, S. & Lesgold, A. (1996). Toward a situated social practice model of instructional
design. In D. C. Berliner & R. C. Calfee (Eds.) Handbook of Educational Psychology (pp. 787-
806) New York: Macmillan.

Derry, S. J., Seymour, J. Feltovich, P., & Fassnacht, C. (2001). Tutoring and knowledge
construction during problem-based learning: An interaction analysis. Paper presented at the
Annual Conference National Association for Research in Science Teaching (NARST), St. Louis
MO.

Derry, S. J., Seymour, J., & Steinkuehler, C., Lee, J. & Siegel, M. (in press). From
ambitious vision to partially satisfying reality: Community and collaboration in teacher education.
To appear in S. Barab, R. Kling & J. Gray (Eds.) Designing Virtual Communities in the Service of
Learning. Cambridge: Cambridge University Press.

Dias, P. Gomes, M. J. & Correia, A. P. (1999). Disorientation in hypermedia

environments: Mechanisms to support navigation. Journal of Educational Computing Research,
20, pp. 93-118.

Gibson, J. J. (1979). The ecological approach to visual perception. Boston: Houghton

Mifflin.

37
 Goldstone, R. L. & Barsalou, L. W. (1998). Reuniting perception and conception.
Cognition, pp. 230-262.

Kolodner, J. L., & Guzdial, M. (2000). Theory and practice of case-based learning aids.
In D. Jonassen & S. M. Land (Eds.), Theoretical foundations of learning environments. Mahwah,
NJ: Erlbaum.

Lampert, M. (2001). Teaching problems and the problems of teaching. New Haven &
London: Yale University Press.

Leinhardt, G. & Greeno, J. G. (1991). The cognitive skill of teaching. In P. Goodyear

(Ed.) Teaching knowledge and intelligent tutoring (pp. 233-268). Norwood, NJ: Ablex.

Merseth, K. K. (1996). Cases and case methods in teacher education. In J. Sikula (Ed.),
Handbook of research on teacher education (pp. 722-744). New York: Macmillan.

Meyer, B. J. F. (1975). The organization of prose and its effects on memory. Amsterdam:
North-Holland.

O’Donnell, A. M. (1999). Structuring dyadic interaction through scripted cooperation. In

A. M. O’Donnell & A. King (Eds). Cognitive perspectives on peer learning. Mahwah, NJ:
Erlbaum.

Putnam, R. T. & Borko, H. (2000). What do new views of knowledge and thinking have
to say about research on teacher learning. Educational Researcher, 29(1), 4-15.

Salomon, G. & Perkins, D. N. (1989). Rocky road to transfer: Rethinking mechanisms of

a neglected phenomenon. Educational Psychologist, 24, 113-142.

Schank, R. P. (1982). Dynamic memory: A theory of reminding and learning in

computers and people. New York: Cambridge University Press.

Schwartz, D. L. & Bransford, J. D. (1998). A time for telling. Cognition and Instruction
16, 475-522.

Siegel, M., Derry, S. J., Steinkuehler, C. A., Kim, J.-B. & Seymour, J. (2001). What and
how preservice teachers learn: designing a course that fosters development of useful theoretical
knowledge and the assessment methods to capture it. Paper presented at the Annual Meeting of
the American Educational Research Association (AERA), Seattle WA.

Simon, H. A. (1996). The sciences of the artificial, 3rd edn. Cambridge: MIT Press

Simon, R.L. (1992). Teaching against the grain. Texts for a pedagogy of possibility. New
York: Greenwood Publishing Group.

Spiro, R. J., Zaritsky, R., Feltovich, P. J., Coulson, R. L., Theta, J. (April, 2001).
Teaching for transfer with digital video cases. Symposium paper presented at AERA Annual
Meeting, Seattle, WA.

38
 Spiro, R. J., Feltovich, P. J., Jacobson, M. & Coulson, R. L. (1991a). Cognitive
flexibility, constructivism and hypertext: Advanced knowledge acquisition in ill-structured
domains. Educational Technology, 31(5), 24-33.

Spiro, R. J. Feltovich, P. J. Jacobson, M. & Coulson, R. L. (1991b). Knowledge

representation, content specification, and the development of skill in situation-specific knowledge
assembly: some constructivist issues as they relate to cognitive flexibility theory and hypertext.
Educational Technology, 31(9), 22-25.

Spiro, R. J., Feltovich, P. J., & Coulson, R. L. (1992). Cognitive flexibility,

constructivism, and hypertext: Random access instruction for advanced knowledge acquisition in
ill-structured domains In T. M. Duffy & D. H. Jonassen (Eds.), Constructivism and the
technology of instruction: A conversation (pp. 57-75). Hillsdale, NJ: Lawrence Erlbaum.

Steinkuehler, C. A., Derry, S. J., Hmelo-Silver, C. E. & DelMarcelle, M. (2002) Cracking

the resource nut with distributed problem-based learning in secondary teacher education. Journal
of Distance Education.

Suchman, L. (1987). Plans and situated actions. New York: Cambridge University Press.

The STEP Research Group* (2000). Promoting teachers’ flexible use of the learning
sciences through case-based problem solving on the WWW: A theoretical design approach. In B.
Fishman & S. O’Connor-Divelbiss (Eds.), Proceedings of the Fourth International Conference of
the Learning Sciences (pp. 273-279). Mahwah, NJ: Erlbaum. (*M. Siegel, S. Derry, J. Kim, C.
Steinkuehler, J. Street, N. Canty, C. Fassnacht, K. Hewson, C. Hmelo, & R. Spiro)

Wiggins, G. & McTighe, J. (1998). Understanding by design. Alexandria, VA: ASCD.

**Zacks, R. & Tversky, D. (2002). Event perception. Psychological Review,

39
Figure 1: Criss-crossing in CFT instruction

Figure 2: Small multiples in CFT instruction

Figure 3: Knowledge reorganization for cognitive flexibility in real-world contexts

Figure 4: A web page from the Theories library

Figure 5: A web page from the Case library

Figure 6: A web page from the pbl site (the sidewalk)

Figure 7: Design challenge for English majors

40
 CFT: APPLY COURSE IDEAS TO MULTIPLE
CASES

SUBJECT MATTER STRUCTURE

LIBRARY OF (AS TAUGHT)
REAL-WORLD
“CASES”
LEARNING SCIENCES

COGNITIVE X-THEORY SOCIO

CASE 1 THEORY IDEAS CULTURAL
THEORY

CASE 2

IP SOCIO
VIEW COGNI
CASE 3 TIVE

CASE 4

CASE 5

Figure 1. Criss-crossing in CFT instruction

41
 CFT: COMBINE MULTIPLE COURSE IDEAS IN
MULTIPLE CASE ANALYSES

LEARNING SCIENCES

COGNITIVE X-THEORY SOCIO

THEORY IDEAS CULTURAL
CASE 1 THEORY

IP SOCIO
VIEW COGNI
CASE 1 TIVE

CASE 1

Figure 2. Small multiples in CFT instruction

42
 CFT: Reorganization of Knowledge
(A Knowledge Representation “Mess”)

CASE 5 CASE 4 CASE 3 CASE 2 CASE 1

IP VIEW

COGNITIVE
THEORY

SOCIO
COGNITIVE

X-THEORY
IDEAS

SOCIOCUL
TURAL
THEORY

Figure 3. Knowledge reorganization for cognitive flexibility in real-world contexts

43
Figure 4. A Web Page from the Theories Section

44
Figure 5. A Web Page from the Case Library

45
Figure 6. The pbl Site at Step 2, Showing Part of Individual Student Notebook With
Prompts

46
 Teaching "Controversial" Texts
Teachers who address controversial and sensitive topics through literature studies take a risk and sometimes meet disapproval
from parents, administrators, students, even colleagues. The case you will study as part of your STEP pbl activity is based on
the English classroom of a popular high school teacher, Mr. H, who is now in his fifth year of teaching. For the past several
years Mr. H has gone out on a limb in teaching a controversial seven-day unit based on The Adventures of Huckleberry Finn.
A "classic" in American literature, this book is taught in many high schools. However, over the years it has also appeared on
several lists of banned books, due to its controversial racial content.
The video case, Huck Finn at Midwest Hi, is based on a high school honors English class that is being audited by
one student who is the only African-American student in the classroom. You will see video based on the first four days of
instruction, although the unit continued for about seven days. In this unit, Mr. H poses the question for the class, "Should the
book be banned?" Throughout the unit, Mr. H and his students approach the issue from multiple perspectives, connecting what
they are discussing to their own environment and simultaneously delving deeply into important literary concepts, such as
satire, which are fundamental to intelligent reading and social criticism. Thus, the instructional goals for the unit pertained to
critical thinking about a controversial social topic, as well as development of domain-specific expertise in the field of English
literature.
Mr. H's discourse style, teaching and assessment methods, and choices for how to handle various topics and issues
in this classroom environment should be of interest. Through reading, questioning and discussion Mr. H and his students seem
to reach a higher level of understanding, although (as noted in Mr. H's interview), not all ideals are achieved. However, by the
end of the book, students ironically see Mark Twain as one of history's greatest opponents of racism, and a brilliant writer who
was able to deal with the issue at a time when few other writers would or could.
Notice there are yellow "keywords" placed throughout the video. These keywords are based on Mr. H's own description of
how he thought about his teaching when he viewed and helped edit the video.
Your Group's Task
Carefully study the video case of Mr. H's classroom plus the inquiry materials related to the case, in order to develop ideas for
how you would teach a similar unit. Your group's task is to follow the online pbl steps (based on the Wiggins and McTighe
approach to instructional design) in planning a 1-2 week unit focusing on race and diversity issues present in this novel, or a
different novel of your own choosing which also focuses on race or diversity issues. You may also choose to plan your
instruction for a very different teaching context. For example, you may develop ideas for teaching in a rural high school, in
middle school, in a racially heterogeneous classroom, or in a non-honors classroom. The choice of teaching context, as well as
which novel you use, is your group's decision.
Watching the video case closely is important. As you are watching, identify some learning science principles that are (or
are not) at work in the video and that may be influencing the success of the instruction. Some questions you will wish to think
about while watching the video case and in discussions with your group include:
-What do you think the teacher is trying to accomplish?
-How does he know his students are achieving what he intends for them?
-Why is he teaching as he does and how successful are his instructional activities?
Following examination of the video case, your group will follow the pbl steps to generate proposals for goals,
assessments and instructional activities for a similar unit of instruction. You may use the case to inform your own unit plan, or
you may also have different ideas that you would like to put forward. During your pbl activity, your group will articulate: 1.
the enduring understandings you hope your students will achieve, 2. how you will use assessment to insure those
understandings are acquired, and 3. the activities and methods of instruction and classroom/discourse management you will
employ. Your approaches should be framed and justified in learning science terms when possible.
While designing your unit, think about the following in relation to the case:
-Can you build on what seemed to work for Mr. H?
-What changes would you make and why?
-What types of assessments did Mr. H use and what would you use to evaluate your own teaching and your
students' progress?
-Which of Mr. H's instructional activities would you include and what would you do differently?
The information you enter into your STEP pbl online notebook and the Group Whiteboard should provide a synopsis of
your individual and group thinking on these kinds of issues.

Figure 7. A Design Challenge for English Majors

47
Figure 8. Part of a Group Whiteboard (Viewed from the Facilitator’s System)

48