Establishing Project-Enhanced Classrooms Through Design

Louis M. Gomez, Douglas N. Gordin
2115 N. Campus Drive, School of Education and Social Policy
Northwestern University, Evanston, IL, USA 60208,


The Learning Through Collaborative Visualization Project is attempting to establish open-inquiry collaborative science projects as a classroom activity structure. This efforts are described from a design experiment perspective where social and material supports are strategically arranged to encourage this classroom evolution. Supports are provided to help teachers form community, teachers manage projects, and students accomplish projects. The community is geographically distributed and draws on widely distributed expertise. Logistical difficulties of coordinating the community are mainly addressed through the use of multi-media network applications. Through the network student mentors and collaboration facilitators are able to interact with remote classrooms.


The Learning Through Collaborative Visualization (CoVis) Project (Pea 1993; Pea, Gomez & Edelson, 1995) is attempting to recast the activity of science classrooms. The inspiration for its approach comes from many sources, including Piaget's metaphor of the child as scientist, Dewey's (1938) clarion call for students to engage in projects sparked by their interest , Latour's (1982) tracing the success of scientific inquiry to the adoption of facile representational technologies (e.g., the calculus, the table of elements, and the electron microscope), and Bruner's (1963) challenge that an intellectually honest form of any subject can be taught to a child of any age. Inspired by such accounts the CoVis project is engaging in action research or an 'design experiment' to establish open-inquiry collaborative science projects as an activity structure. (Cole, 1993; Leont'ev, 1979). By open-inquiry science collaborative projects (hereafter, referred to simply as projects) we mean projects:

By activity structure we mean a behavior that is commonly and repetitively performed and has social and material support (examples of activity structures are mid-term tests, lectures, and lunch). As such, activity structures are supported and defined by communities that name the activity, give them value, regulate its function, and assess success or failure. Projects as a primary vehicle for instruction are rare in most schools.
Most schools do not have established practices to allow independent inquiry or collaboration at a distance as a primary mode of student and teacher activity. Many schools, of course, do projects. But, these are usually special capstone events, apart from day-to-day classroom activity. Even more rare are classrooms that work with other classrooms in the normal course of events. The CoVis design goal is to encourage activity structures that support the accomplishment of inter-school projects as a regular part of classroom life.

In this paper we take a design perspective (Brown, 1992; Brown & Duguid, 1992; Salomon, 1994; Collins, 1990) to report CoVis's efforts to establish projects as an activity structure. For us, a design perspective means that we present the strategic elements conjectured to be necessary to reach a complex design goal that has organizational, cognitive, and behavioral components. In this case the goal is creating school-based communities (including multiple classrooms and connections to neighborhood, professional, and governmental communities) that use projects as a key instructional vehicle. Four strategic elements characterize what we have done to support the adoption of projects as a new inter-school activity structure:

  1. Provide social and material supports that enable the new activity structure.
  2. Change the surrounding environment(s) to support the activity structure.
  3. Involve the participants as co-designers.
  4. Keep the design process alive by recording decisions and providing means for continuing evolution.
Next we describe how each strategic design element unfolds in the context of the CoVis Project.

Social and Material Supports

In designing supports for projects we made decisions about what functions are best supported socially and which are best supported materially. In general, all activity structures are supported by a combination of social and material supports. Consider shopping as an example. Each shopping outlet (e.g. catalog or store) supports the shopper in their task through personnel (e.g. taking the order on the phone) and through material or technological means (e.g., exchange of currency). The following categories represent our provisional or heuristic policies on the division of social and material supports for inter-school project activity.

Social Supports

Material Supports
These assignments are not exhaustive nor are they permanent. It represents our use of the resources we have available to solve our design problems. If, for example, we had scheduling software usable by schools, that coordination support would become material. Like any other set of design decisions (e.g., designing an airplane), this process of supporting an activity structure involves making the most informed set of judgments possible at the time. In the case of an aircraft, the designer consults previous planes. Here we have consulted other design experiments (e.g., Ruopp, 1993; Songer, 1994) and tried to adopt their choices and techniques to our purposes.

Changing the surrounding environment

Environments and systems occur in levels (e.g., a person is part of a family which is part of a neighborhood which is part of a city). Changing activity at one level often requires or can be substantially aided by also changing activities at other levels. Changes in the classroom can be aided or stymied by factors beyond the classroom. Within the educational reform movement these interrelationships are expressed as the need for systemic reform (Cuban, 1986) or the recognition that social reforms are required in order to achieve changes in science pedagogy. For example in the domain of assessment, D'Amico et al (1994) recognized that project-based curricula in the classroom need to be supported by assessment policies by at the school, department, district, and national level. Simply put an open-end inquiry reform will fail if the larger context demands the teacher present and learners acquire "just the facts". Therefore, CoVis has not only concentrated on student projects but also on teacher development, and building cross-school teacher community. We required that schools apply to CoVis project as a school community. In addition, we ask each CoVis school to hold open houses for parents to explain how school work will change. In this way school teachers, administrators, and parents are involved and active in determining what changes the CoVis intervention will bring.

Participants as co-designers

Whether it is acknowledged or not, recipients of design are always, in practice, co-designers. Their design participation manifests in selectively using certain parts of the design, choosing when and where the design is practiced, and in the performance or actual use of the design (what Suchman calls situating the action; Suchman, 1987; Schuler and Namioka, 1993), and in originating redesigns in functionality or form. Capturing and disseminating this 'in-practice' design can have substantial benefits, not only in improving functionality, but also in fostering a sense of ownership and participation by users. This latter point is particularly important for sustaining innovation. CoVis is attempting to involve teachers and students in its design efforts at every phase, thereby fostering community and producing products that foster growth or generativity through use.

Keeping the design process alive

The CoVis project is familially linked to a host of other design projects (Brown, 1992; Songer, 1994; Riel, 1990; Ruopp, et al., 1993). These efforts are 'design experiments' in that they seek to achieve some practical change through design and to assess the success or failure of the effort. The longer-term significance is also measured in extracting design principles of what works. Towards this end it is important to capture the design heuristics employed. This paper aims to provide some documentation of this sort on the CoVis project, so that similar efforts can be compared and that future efforts can draw upon its experience. While, it is still too early to provide evaluation of the design attempts described here, the evaluation will incorporate the following types of data:

  1. Demographic data will be collected on all students.
  2. Weekly surveys describing the class activities (e.g., project work, lecture, laboratories, films, guest speakers) will be collected from all teachers. In addition, the use of novel media, such as scientific visualization, will be reported.
  3. Studies of individual classes will be performed, both longitudinal and case studies around specific curricula.
  4. Usage of computational and communication tools will be monitored.
  5. The on-line group discussions by teachers (described below) will be used as a source of data (though anonymity will be preserved).
  6. Change measures in students inquiry ability, in particular their skill at planning a scientific investigation, and change measures on student attitude toward science, including interest and self-efficacy.
Our goal is to have a kind of "organic" data set, in that it can inform several constituencies and can change as design goals evolve.

CoVis Design Space

The classroom activity structures CoVis is attempting to design, and the means to accomplish them, are summarized in Table 1. The changes are described in terms of proposed activity structure and the social and material supports designed to facilitate its achievement. The rest of the paper will follow the structure of this table by discussing each level in turn. In particular this account will emphasize, the central role of networked computational tools, particularly for their ability to provide support for structured communication at a distance (e.g., tele-mentoring and networked curriculum) and to provide access to new representational mediums (e.g., scientific visualization).

Community Building

The CoVis project can most simply be described as an effort to build community. Currently, this community includes some 140 teachers at 45 schools; the over 4000 students they teach; more than a dozen graduate students, support staff and professors in education, computer science and meteorology; several high technology computer and communication companies; and dozens of mentors. This community has only recently grown to be this size. When the CoVis project began in 1991 it included only six teachers at two nearby schools. The explosive growth occurred only recently (summer '95) as part of an effort to test whether the CoVis methods, curriculum and principals could be scaled up, thus providing benefits to a more sizable and diverse constituency. The current schools span sizable portions of the United States, includes mostly high schools and some middle schools and encompasses all typical science classes. Clearly, networked computer communication means are essential if the assembled group of people are ever to live up to the term 'community.'

The requirements for these schools to join CoVis were a commitment to try project pedagogy, willingness to focus on cross-disciplinary topics in the geosciences, at least a 56 kilobyte connection to the Internet, a one to five ratio of computers to students in the classroom to be used, and a dedicated half-time technology administrator. These requirements neatly exemplify a recurring theme: Achieving a new activity structure (like projects) requires material and social supports.

Table 1: Example CoVis activity structures aided by social and material supports

Change Level Activity Structure Social Support Material Support
Teacher Community Cross-talk groups In-person Meeting at workshops Email
Share and refine curricular resources Web server coordination [Editor] Inter-school activities on Geosciences Web Server

Teacher in the Classroom Teacher leads project-based curriculum Workshop discussions Curriculum, and assessment rubrics
Teacher development specialist Mentor data base
Virtual field trips & weather briefings Video conferencing

Student Work Projects Adult Mentors Collaboratory Notebook
  • Negotiating scope
  • Finding Resources
  • Performing analysis
  • Logical argument
  • Convincing Presentation
Remote peers Scientific Visualization
Staging Activities & Examples of Projects
Project Schemata

Building the Initial Connections

A variety of means were used to approach the schools including posting to relevant listservs, negotiating with city and state public school administrators, holding open house demonstrations, and sharing the schools of other technology education projects (especially BBN's CONNECT schools). Teachers and schools that applied were screened through an application process. Four day workshops were then held to acquaint teachers with CoVis pedagogy, curriculum, and technology. The pedagogy was encountered through discussion groups, while the curriculum and technology portions were combined into hands-on activities followed by evaluation and critique. In addition, technology administrators attended a parallel strand of the workshops where they studied network hardware, communications programs and other software, and administering school local area networks. Part of the time was dedicated to investigating the special circumstances of their school, including, where they would obtain Internet service and how the network and computers in their schools could best be laid out.

Community Activity Structures

The activity structures through which teachers (and technical administrators) interact as a community are dialoguing over 'cross-talk' groups and building and annotating shared curriculum on the Geosciences Web Server ( Cross-talk groups provide networked communication between the teachers, while the Geosciences Web Server provides a living storehouse of curriculum, data sets, and other resources that teachers can annotate with suggestions, minor changes, and major additions. The two media are intended to support one another --one provide a means for dialogue among small groups (cross-talk) and the other (web server) server establishes a community wide storehouse of knowledge. Each of these is described more fully below, by enumerating their social and material supports and detailing design decisions in their formation.

Cross-talk Groups

Cross-talk groups consist of twenty or so teachers, who exchange email about wide variety of topics. The groups were selected to maximize geographical and socio-economic diversity. However, all the middle schools are grouped together, so as to provide for the differences based on age.

A email reflector (an special address that causes a message to be sent to all participants) is used to provide a shared email exchange. The exchange is moderated by a CoVis researcher who facilitates the conversation. The goals of the cross-talk are to foster professional growth, be a source of new ideas, provide an outlet for questions and comments, be a support system for teachers, and, in general, provide a place for the teachers to engage in dialogue. On occasion the facilitator invites experts on-line so as to provide a focal point and privileged voice for discussions and disputes concerning current hot topics or common curricular units. Aside from these expert 'visits' the cross-talk groups are designed primarily as an egalitarian medium.

The cross-talk groups are the primary way that dialogue occurs between cross-sections of the teacher community. The size of each group was limited in order to maximize participation. Many listservs are dominated by a few frequent talkers with many listeners. Incorporating a group facilitator further capitalizes on the group's small size. The facilitator will keep track of participation and contact members that seem to have withdrawn.

Generative Curricula on Geosciences Web Server

A Geosciences WWW server is being created in order to provide a central place for the distribution of the curriculum and data sets and a place where teachers can record comments, suggestions, revisions, and new curricula. The Web Server serves as a bridge between the various cross-talk groups and the greater community of teachers by establishing a common storehouse for the entire community.

Our intention is to build a generative resource where teachers can adapt projects and other curricular resources to their own objectives and the constraints of their situation. The Geoscience Web server is dynamic in several ways. Dynamic web based media being developed include a bulletin board of events and teacher pages of favorite web resources. The bulletin board allows teachers to post upcoming events of interest to the CoVis community. This medium is especially well suited to teachers recruiting others classes to participate in a collaborative activity (e.g., the famous Eratosthenes project where classes at different longitudes but the similar latitudes compute the earth's circumference using the angle of the sun at their different locations but at the same time). Using dynamic web pages teachers can add pointers to favorite web servers annotating the link with a name and description, thereby providing a resource for their students and others to use.

For collaboratively building the curriculum on the web a human editor is essential. An editor or broker is needed to preserve the coherence of the curriculum and prevent it from being drowned in alternative project designs. The editors goal is not to limit project options but to organize and standardize contributions. Further, an editor can provide indexing schemes to appropriately link new curriculum in with the existing corpus. Several of the curricula in the Geosciences web server are designed to be conducted across multiple schools at the same time, thereby providing common ground for networked discussions and coordinated inter-school work. The inter-school curricula brokered centrally at Northwestern which coordinates teacher interactions, disseminates innovations, connects the teachers with domain experts, and provides troubleshooting expertise for the curricula. A goal for the CoVis project is for teachers or some other agency, like a publisher, to assume this role of inter-school curriculum coordinator. How to encourage this and thereby support the longevity of inter-school collaborations is a key research question. Today schools and the surrounding culture have well worked ways to sustain single classroom activities. These structures have to evolve for inter-school events if they are to survive.

Permeable Classrooms conducting project-enhanced curricula

Our goal is for teachers to develop inter-school project curricula as a well entrenched activity structure. In this section we briefly outline some of the design decisions we have made to give projects a key role in classrooms.

We started with workshop discussions where we modeled student projects for teachers . Workshop discussions were designed to help teachers consider the role that projects could play in their curriculum. Workshops also provided a way for teachers to share their experiences with each other, and provide a forum for CoVis research staff to build common ground with the teachers. Material support for this activity are provided in the form of curricula and associated assessment rubrics packaged into a loose-leaf resource guide and on the Geosciences web server. The curriculum topics are:

  1. 'Land Use Management Plan' where students investigate the environmental decisions made by a particular city over the past hundred years and retrospectively consider alternative decisions.
  2. City in a Box where students investigate the properties of soil with respect to erosion, environmental pollution, and stability. These properties are investigated in isolation and then combined into a scale model. The scale models from different schools compete with one another.
  3. 'Student Conference on Global Warming' has students evaluate the evidence for global warming and consider possible trends and consequences. Students then investigate either a global issue or the point of view of a single country. The results of their investigations are shared and then the entire class considers current results of international policy in light of their earlier projects.
  4. 'Its Weather Man' has students look at real-time and historical data to explore principles of ordinary and severe weather. This learning is used to understand various historical scenarios, thereby providing students the means to build scenarios of their own.

The curricula units follow a similar structure:

  1. Short-term staging activities (e.g., laboratory experiments and computer simulations) occur first to provide students with the principles and techniques that underlie work in the area.
  2. Group project work follows where students 'do science' by engaging in open-inquiry that they partially structure. Outside resources are drawn into the classroom to support these projects, ranging from tele-mentoring (or mentoring discussions over the Internet) to obtaining scientific data sets.
  3. Finally, students present, evaluate, and synthesize their project work both within and across classrooms.

These curricula are based on CoVis's experience to date on how projects can be integrated into the life of a classroom and made feasible. They are by no means finished products. The primary role of the curricula is to provide rough schemata that teachers can use in designing a class where projects play an important part. Associated with each curricula are opportunities for teachers to bring outside resources into their class. A primary goal of CoVis is to connect outside mentors and experts with classrooms through activities. In this way CoVis is looking to make interaction with disciplinary experts a regular part of classroom activity -- in this sense the classroom walls become more permeable.. Consider the following examples of networked access to the world outside the classroom supported by CoVis activities:

Bringing video-based interactive experts into the classroom is an approach shared by other projects (Feldman, 1994). Our goal here is two-way interaction between teachers and learners and the world beyond rather than one-way lectures.

A concerted effort has also been made to line up domain experts and other mentors to participate with students and teachers during the inter-school curricular activities. A data base of mentors has been provided for teachers to help contact mentors. This data base, which is another part of the Geoscience web server, supplies means for teachers to selectively search and "check-out" mentors whose expertise match their students interests. The data base is designed to be used by teachers, not students. Our belief is that the teacher needs to form a preliminary relationship with the mentor and provide a filter before student interaction. This filter is both to protect the mentor from inappropriate access by students and also to help the mentor in understanding the context (through the teacher's eyes) in which the student is working.

Students doing projects in the spirit of scientists

Students doing projects is the culmination of the curricula. Before the open-ended segments of projects begin, students engage in staging activities. These activities are designed to acquaint the students with the principles and techniques that underlie work in that area.

We will use global warming to illustrate (Gordin, et. al.). In the global warming curriculum staging activities students are acquainted with natural variation in climatic temperature, human caused increases in atmospheric carbon-dioxide, the use of spreadsheets and scientific visualization. Further, the staging activities specifies themes for the open-ended projects to follow. In particular, students focus on consequences and causes by country or explore a global change. The goal is to provide project schemata that ease the students task of deciding on the focus and extent of a project, while still providing ample room for students interests to dictate the specific investigation. In laying out typical questions and data useful to investigate global warmings potential impact on a country and how a country might be contributing to global warming occurring, a general framework or schemata has been established. Students specialize the framework by choosing a country, its specific data, and the particular issue they want to focus on (e.g., rise in carbon-dioxide emissions due to recent growth, deforestation, flooding due to rising sea levels).

Students and mentors often have difficulty finding common ground on which to talk. One solution to this problem is provided by the Collaboratory Notebook (O'Neill and Gomez, 1994) that combines the functionality of a word processor with the ability to organize a scientific investigation on a series of semantically typed pages (e.g., question, supporting evidence, conclusion). These pages can alternatively be viewed as the steps in constructing an argument or dialogic exchanges in a conversation. Further, the pages are kept in a centralized data base that support multiple levels of privacy and single-writer multiple-readers access. As such, students can record their emerging lines of reasoning and their mentors and teachers can insert dialogic rejoinders that provide in-context aid to their investigations.

Another way of forging common ground with scientist mentors is the use of the same data sets. Use of scientist data sets and their specialized representations systems has become one of CoVis's primary operationalizations of its goal to have students learn science by doing science. CoVis curricula and software have incorporated scientific visualization as a primary representational system to aid classroom activity. In the global warming curriculum visualizations of the Earth's global energy balance are investigated using monthly means of surface temperature, incoming sunlight, albedo (i.e., earth-atmosphere reflectivity to incoming sunlight), energy being emitted by the earth-atmosphere system, and net energy received. While students often encounter difficulties in appropriating these sophisticated representational tools there are sound reasons for encouraging their use, even at the middle and high school levels. New representations often herald the investigation of new topics and phenomena in science. In particular, scientific visualization is linked with the modern study of complex phenomena (e.g., the storage of energy in the earth's atmosphere called the greenhouse effect). Similarly, visualization can bring the study of complex multi-level phenomena to science classrooms. In addition, synergistic advantages we predict will be obtained by the combination of students using contemporary data and representations and having scientists as mentors, since the common data and tools being used can forge common conceptual ground on which meaning negotiation can occur.

Learning from Design Experiments

Design experiments aim to change the world through substantive action and also to discover the means to enact those changes more broadly. A difficulty is in extracting the necessary parts of the design experiment from the incidental. This is problem exists for all experiments, even those in the traditional laboratory hypothesis proving models. Unlike traditional factorial experiments, design experiments do not make efforts to reduce the variables manipulated to a minimum and perform controlled manipulations. Their first goal is to find a confluence of elements 'that work.' In the CoVis case, for example, we first want to find that set of elements that will cause a community to 'do' projects. If we succeed then the question becomes: How then can principles be extracted from the results of design experiments? We have no specific solution for this problem, but believe two directions are promising. First, a scale-up of the design experiment can be attempted so that its interventions involve different magnitudes and diversity of participants. Assessment of those populations where the design is successful and those where it is not may be a window on the key elements of the design experiment. Second, meta-studies of many design experiments can be compiled and regularities in success and failure observed. Obviously, CoVis is attempting a verification of the first kind. Helpful to both efforts is documenting the intentions, methods, and design goals, and planned proof procedures. These documents will be useful in gauging what was tried, and what was achieved. Documentation of this sort, at the beginning, middle, and end of the design experiment will serve the design community in learning from its work and shaping future designs.


Brown, A.L. (1992). Design experiments: Theoretical and methodological challenges in evaluating complex interventions in classroom settings. The Journal of the Learning Sciences 2(2), 141-178. In P.S. Adler and T.A. Winograd (Eds.), Usability: Turning technologies into tools. Oxford: Oxford University Press.
Brown, J.S. & Duguid, P. (1992). Enacting design for the workplace.
Bruner, J.S. (1963). The process of education. Cambridge, MA: Harvard University Press.
Cole, M. & Engeström, Y. (1993). A cultural-historical approach to distributed cognition. In G. Salomon (Ed.), Distributed cognition. New York: Cambridge University Press.
Collins, A. (1990). Towards a design science of education (Technical Report No. 1). New York: Center for Technology in Education, Bank Street College of Education.
Cuban, L. (1986). Teachers and machines: The classroom use of technology since 1920. New York: Teachers College Press.
D'Amico, L., Gomez, G., & McGee, S. (1994, May). A case study of student and teacher use of projects in a technology-supported distributed learning environment. Paper presented at the Educational Testing Service Conference on Natural Language Processing techniques and Technology in Assessment and Education, Princeton, NJ.
Dewey, J. (1938). Experience and Education. New York: Macmillian Publishing Company.
Feldman, G.C. (1994). The jason project. [Online] Available at For more information send email to, phone number (301) 286-9428.
Gordin, D.N., Edelson, D.C., Gomez, L.M., Lento, E.M., Pea, R.D. (1996). Student conference on global warming: A collaborative network-supported ecologically hierarchic geosciences curriculum. To appear in the Preprints, Fifth Symposium on Education, American Meteor. Soc., Boston, MA.
Leont'ev, A.N. (1979). The Problem of Activity in Psychology. In Wertsch, J.V. (Ed), The Concept of Activity in Soviet Psychology (pp. 37-71). Armonk, New York: M. E. Sharpe, Inc.
Latour B. (1988). Drawing things together. In M. Lynch & S. Woolgar (Eds.). Representation in scientific practice. Cambridge, MA: MIT Press.
O'Neill, D.K. & Gomez, L. M. (1994). The Collaboratory Notebook: A Distributed Knowledge-Building Environment for Project Learning. Proceedings of ED-MEDIA 94, Vancouver, Canada, 1994.
Pea, R.D., Gomez, L.M., Edelson, D.N. (1995). Science Education as a Driver of Cyberspace Technology Development. INET '95, Honolulu, Hawaii. [Online] Available at
Riel, M. (1990). Cooperative learning through telecommunications. Instructional Science, 19, 445-466.
Ruopp, R., Gal, S., Drayton, B., & Pfister, M. (1993). LabNet: Toward a community of practice. Hillsdale, NJ: Erlbaum.
Salomon, G. (July, 1994). Not just the individual: A new conception for the future of educational psychology. Presidential address of the division of educatioinal, instructional and school psychology presented at the International congress of applied psychology. Madrid, Spain.
Schuler, D., & Namioka, A. (1993). Participatory design: principles and practices. Hillside, New Jersey: Lawrence Erlbaum Associates, Inc.
Songer, N.B. (1994). Knowledge construction through global exchange and dialogue: A case of kids as global scientists.
Suchman, L. (1987). Plans and situated actions: The problem of human-machine communication. Cambridge: Cambridge University Press.


We are grateful for research support of the Learning through Collaborative Visualization (CoVis) Project by the National Science Foundation Grant RED-9454729, and the Illinois State Board of Education/Eisenhower Program. We would also like to thank our colleagues from the CoVis Project and community of users for extended discussions of these issues, and continual useful feedback on design, rationale, and pedagogical issues.