A101 Edelson Etal 99 PDF

Addressing the Challenges of Inquiry-Based Learning
Through Technology and Curriculum Design

Daniel C. Edelson, Douglas N. Gordin, Roy D. Pea
To cite this version:

Daniel C. Edelson, Douglas N. Gordin, Roy D. Pea. Addressing the Challenges of Inquiry-Based
Learning Through Technology and Curriculum Design. The Journal of the Learning Sciences, 1999,
8(3-4), pp.391-450. �hal-00190609�
HAL Id: hal-00190609

https://telearn.archives-ouvertes.fr/hal-00190609
Submitted on 23 Nov 2007
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
THE JOURNAL OF THE LEARNING SCIENCES, 8(3&4), 391450
Copyright O 1999, Lawrence Erlbaum Associates, Inc.
Addressing the Challenges of

Inquiry-Based Learning Through
Technology and Curriculum Design
Daniel C. Edelson
Institute for the Learning Sciences and School of Education and
Social Policy
Northwestern University
Douglas N. Gordin and Roy D. Pea

Centerfor Technology in Learning
SRI International
Menlo Park. CA
Inquiry experiences can provide valuable opportunities for students to improve their
understanding of both science content and scientific practices. However, the imple-
mentation of inquiry learning in classrooms presents a number of significant chal-
lenges. We have been exploring these challenges through a program of research on
the use of scientific visualization technologies to support inquiry-based learning in
the geosciences. In this article, we describe 5 significant challenges to implementing
inquiry-based learning and present strategies for addressing them through the design
of technology and curriculum. We present a design history covering 4 generations of
software and curriculum to show how these challenges arise in classrooms and how
the design strategies respond to them.
Students at all grade levels and in every domain of science should have the opportu-
nity to use scientific inquiry and develop the ability to think and act in ways associated
with inquiry. (National Science Education Standards. National Research Council
[NRC], 1996, p. 105)
Correspondence and requests for reprints should be sent to Daniel C. Edelson, Institute for the
Learning Sciences and School of Education and Social Policy, Northwestern University, 1890 Maple
Avenue, Evanston, IL 60201. E-mail: d-edelson@nwu.edu
392 EDELSON, GORDIN, PEA
Recent years have seen a growing call for inquiry to play an important role in sci-
ence education (e.g., American Association for the Advancement of Science
[AAAS], 1994; Blumenfeld et al., 1991; Linn, diSessa, Pea, & Songer, 1994; NRC,
1996). This call for inquiry-based learning is based on the recognition that science
is essentially a question-driven, open-ended process and that students must have
personal experience with scientific inquiry to understand this fundamental aspect
of science (Linn, Songer, & Eylon, 1996; NRC, 1996, Project 2061). Furthermore,
inquiry activities provide a valuable context for learners to acquire, clarify, and ap-
ply an understanding of science concepts. At the same time, computer technologies
are receiving increased attention from the science education community because of
excitement about their potential to support new forms of inquiry. These two reform
trends are coming together in the form of numerous projects to create designs for
technology-supported, inquiry-based science learning. For example, a large num-
ber of educational research and development projects are currently exploring the
use of computers and networks to collect, exchange, and analyze scientific data. In
the earth sciences alone, scores of educational projects have been initiated in the
last decade to provide data and analysis tools to the educational community. Other
projects engage learners in the collection and exchange of scientific data. Many of
these projects in the United States are the direct result of National Aeronautics and
Space Administration (NASA) and National Science Foundation fbnding pro-
grams designed to facilitate synergies between the scientific research and educa-
tion communities. The goal of this article is to address the needs of designers of ed-
ucational experiences like these.
Over the last 6 years, the authors have been engaged in the investigation oftech-
nology-supported inquiry learning (TSIL) through the design, implementation,
and evaluation of scientific visualization tools for learners. Our goal has been to
understand the opportunities and obstacles presented by scientific visualization as
a technology to support inquiry-based learning. In the pursuit of this objective, we
have developed a series of scientific visualization environments and inquiry-based
curricula and have studied their use in both laboratory and public school classroom
settings. In the course of this research, we have confronted a number of significant
challenges to the implementation of inquiry-based learning and explored design
strategies for responding to them. In this article, we discuss these challenges and
strategies in the context of a design history covering four generations of software
and curriculum. We begin with a discussion of the opportunities for learning pro-
vided by TSIL in general and how they apply to the specific domain of climate sci-
ence and the specific technologies of scientific visualization. We then present a
summary of the challenges for implementing inquiry learning that this design his-
tory addresses, followed by the design history itself. Following the history, we
highlight the design strategies we pursued and describe how they address the chal-
lenges to inquiry-based learning. We conclude with a brief discussion of the next
steps for design and evaluation in this research.
CHALLENGES OF INQUIRY-BASED LEARNING 393
THE POTENTIAL OF TSlL
Inquiry, the pursuit of open questions, is fundamental to the practice of science. In-
quiry-based science learning is based on the idea that science learning should be au-
thentic to science practice, an idea advocated by Dewey (1964a, 1964b). Modem
support for inquiry-based learning comes from research in cognitive science that
provides evidence for the importance of activity and authentic contexts for learning
(e.g., Greeno, Collins, & Resnick, 1996). Authentic activities provide learners with
the motivation to acquire new knowledge, a perspective for incorporating new
knowledge into their existing knowledge, and an opportunity to apply their knowl-
edge. In contrast to the passive reception of knowledge associated with conven-
tional science learning, inquiry is active. As an authentic scientific practice, inquiry
also provides a valuable context for science learning.
In this section, we provide a brief rationale for incorporating inquiry into sci-
ence education and for using technology to support inquiry-based learning. To do
so, we highlight the opportunities for learning provided by inquiry and the oppor-
tunities to support inquiry-based learning provided by computer technologies. We
discuss these opportunities both in general and in the specific case of climatology
and scientific visualization that we have explored in our research.
Opportunities for Learning
Scientists' knowledge of scientific concepts, understanding of scientific tools and

media, and inquiry skills are inextricably intertwined. Our goal in attempting to im-
plement inquiry-based science learning is to help students develop this same sort of
integrated understanding. Participation in inquiry can provide students with the op-
portunity to achieve three interrelated learning objectives: the development of gen-
eral inquiry abilities, the acquisition of specific investigation skills, and the under-
standing of science concepts and principles.
The first opportunity for learning provided by inquiry is the opportunity to de-
velop general inquiry abilities. In its prototypic form, inquiry-based science learn-
ing involves the pursuit of open-ended questions and is driven by questions
generated by learners (Blumenfeld et al., 1991; Linn et al., 1996; Welch, Klopfer,
Aikenhead, & Robinson, 1981). In practice, posing researchable questions and
pursuing them through open-ended investigations are abilities that must be
learned. General inquiry abilities include posing and refining research questions,
planning and managing an investigation, and analyzing and communicating re-
sults. Inquiry activities provide the opportunity to develop and exercise these gen-
eral inquiry abilities.
Inquiry also provides the opportunity to acquire specific investigation skills.
Different disciplines of science use different methods of investigation and differ-
ent standards of evidence. Some forms of inquiry that have been explored by edu-
cational researchers include controlled experimentation (Schauble, Glaser,
Duschl, Schulze, & John, 1995), modeling (Jackson, Stratford, Krajcik, &
Soloway, 1996; Penner, Giles, Lehrer, & Schauble, 1997; Resnick, 1994;
Wilensky & Resnick, 1999), synthesis of primary sources (Linn, Bell, & Hsi,
1998; Wallace et al., 1998), and exploration of quantitative data (Hancock, Kaput,
& Goldsmith, 1992; Tabak, Smith, Sandoval, & Reiser, 1996). Each ofthese forms
of investigation has its own specific procedures and skills. Within these forms,
there are even further differences between scientific disciplines. For instance, the
design of controlled experiments in chemistry is very different from the design of
controlled experiments in psychology. The inquiry-based approach to science
learning provides learners with the opportunity to learn these scientific practices
by engaging in them themselves.
The third opportunity for learning provided by inquiry is the chance to develop
an improved understanding of science concepts. Inquiry activities can contribute
to this knowledge acquisition process by providing a meaningful context for learn-
ing. However, it is important to distinguish inquiry learning from discovery learn-
ing (Bruner, 196 1; de Jong & van Joolingen, 1998). In our conception of learning
from inquiry, students can discover scientific principles through their inquiry ac-
tivities, but discovery is not the only mechanism for learning from inquiry. Inquiry
can contribute to the development of science content understanding in all of the
following ways:
Problematize (Hiebert et al., 1996). Inquiry activities can lead learners to con-
front the boundaries of their knowledge or recognize gaps in that knowledge. The
limits of one's knowledge are often revealed by the failure of an expectation about a
particular situation (Schank, 1999) leading to curiosity. The curiosity elicited by
such problematic situations creates a focused motivation to learn (Berlyne, 1966).
Demand. Successfilly completing a scientific investigation requires science
content knowledge. The design of an inquiry activity can, therefore, place a de-
mand for knowledge on the part of a learner that will require him or her to acquire
it to complete the investigation successfully.
Discover and refine. By providing learners with the opportunity to pursue an-
swers to questions, inquiry activities can enable learners to uncover new scientific
principles and refine their preexisting understanding of scientific principles in the
answers that they construct.
Apply. Inquiry activities can give learners the opportunity to apply their scien-
tific understanding in the pursuit of research questions. The need to apply scientific
knowledge can require a learner to reorganize and re-index it in ways that will sup-
port its future use. The application of existing knowledge can also reinforce it and
enrich its connections to other knowledge.
Technological Support for Inquiry-Based Learning
Computing and networking technologies offer dramatic, new opportunities to sup-

port inquiry-based learning. For instance, in their analysis of technology as a sup-
port for project-based science learning, Blumenfeld et al. (1 99 1) identified six con-
tributions that technology can make to the learning process:
Enhancing interest and motivation.

Providing access to information.
Allowing active, manipulable representations.
Structuring the process with tactical and strategic support.
Diagnosing and correcting errors.
Managing complexity and aiding production.
All of the fundamental properties of computing technologies offer benefits for in-
quiry-based learning-the ability to store and manipulate large quantities of infor-
mation, the ability to present and permit interaction with information in a variety of
visual and audio formats, the ability to perform complex computations, the support
for communication and expression, and the ability to respond rapidly and individu-
ally to users. With regard to the inquiry process in particular, a variety of promising
technologies have been developed for learners, including tools for modeling phe-
nomena and processes from the real world (e.g., Model-It, Jackson et al., 1996;
STELLA, Mandinach & Cline, 1996; ThinkerTools, White, 1993),visualizing and
analyzing quantitative data (e.g., TableTop, Hancock et al., 1992; WW2010,
Plutchak et al., 1998; GLOBE, Rock, Blackwell, Miller, & Hardison, 1997;
Blue-Skies, Samson et al., 1994), exchanging data and ideas across distances
(GLOBE, Rock et al., 1997; Kids As Global Scientists, Songer, 1995), structuring
and supporting discussion (e.g., CoVis Collaboratory Notebook, Edelson, Pea, &
Gomez, 1996a; Gomez & O'Neill, 1994; CaMILE, Guzdial, Turns, Rappin, &
Carlson, 1995; SpeakEasy, Hoadley, Hsi, & Berman, 1995;CSILE, Scardamalia &
Bereiter, 1994), and providing access to information in the form of digital collec-
tions and libraries (e.g., Knowledge-Integration Environment, Linn et al., 1998;
ARTEMIS, Wallace et al., 1997). In our approach to inquiry-based learning, com-
puter technologies are used to provide investigation tools, knowledge resources,
and record-keeping tools.
An Example: Climatology and Scientific Visualization
The field of climate science and the technologies of scientific visualization both
provide rich opportunities for inquiry-based learning. In combination, they provide
opportunities to exploit technology as a support for inquiry to achieve the three cat-
egories of learning objectives described previously.
Climatology. The science of climate is an important area for study because,

like the other geosciences, it integrates the sciences of physics, chemistry, biology,
and geography. It provides an opportunity to apply a systems approach to science
learning that is increasingly practiced in the earth sciences (Ireton, Manduca, &
Mogck, 1996; NASA Advisory Council, 1988). Climatology also provides an op-
portunity for students to develop an understanding of processes on scales ranging
from local to global. Climate topics, including the cause of the seasons, energy
transfer from sun to earth, and the role of sunlight in photosynthesis, are included in
the secondary science education content standards for earth science published by
both the American Association for the Advancement of Science (AAAS, 1994) and
the National Research Council (NRC, 1996).
Because the study of large-scale atmospheric processes does not permit experi-
mentation, climatology has historically been an observational science requiring
the collection and analysis of data (Frodeman, 1995), although it increasingly in-
volves the generation of data through computational models. The availability of
large-scale data collection mechanisms, including remote-sensing satellites and
networked data collection stations, has led to an explosion of data available to sup-
port scientific investigation of climate (NASA, 1998). The techniques of
two-dimensional geographic visualization and data analysis that we investigate in
this research have played an important role in the practice of climatology in the
years since these large databases and analysis tools have been available. The avail-
ability of the data and the impact of visualization technologies on the practice of
science in this field provided the initial inspiration for this research.
Scientific visualization. The term scientzfic visualization covers a class of

techniques for rendering data that take advantage of the computer's powerful capa-
bilities for graphical and animated display (McCormick, DeFanti, & Brown, 1987).
Scientific visualizationtechniques are used to display data visually through the sys-
tematic variation of color, shape, orientation,and position. The techniques of scien-
tific visualization have had an enormous impact on many fields of science because
they exploit the ability of the human visual system to identify patterns in visual im-
agery, where previously scientists could only search for such patterns through com-
plex, analytical processing. For example, atmospheric scientists have made great
leaps in their understanding of processes as small as individual thunderstorms and
as large as global climate change by collecting and modeling enormous quantities
of data and rendering them in visual displays (Kaufmann & Smarr, 1993). Scien-
tists use visualization not just for inquiry but also for communication, to report
findings, and to convince others. As a glance at virtually any journal in the

geosciences reveals, scientific visualizations now play an important role in com-
munication within the scientific community.
Previously, we have argued that scientific visualization offers great promise as
a support for science learning (Gordin & Pea, 1995). The same properties that have
made visualization a powerful technology for scientific discovery by scien-
tists-the ability to render complex data for visual interpretation-make it a po-
tentially powerful tool for science learning by students. Furthermore, as a
technology for investigation and a medium for communication, scientific visual-
ization can provide students with a means to participate in the practice of science,
thereby benefiting from the inquiry-based learning that is increasingly being rec-
ognized as a critical element of science education. Finally, as a set of visual repre-
sentations, scientific visualization formats are likely to take a place alongside
charts and graphs as representations that students must master to become compe-
tent decisionmakers in a world that is increasingly influenced by science and tech-
nology.
In our research, we have focused primarily on the use of scientific visualization
for the analysis of large collections of quantitative data created by the scientific
community. The form of visualization that we have focused on is the display of
rectangular arrays of geographic data that use colors to represent ranges of numeri-
cal or categorical values. This "color map" geographic representation (Figure 1) is
widely used in the atmospheric sciences community and appears increasingly in
popular media, such as the daily weather maps in USA Today. The research re-
ported here has taken place concurrently with a number of efforts exploring the use
of visualization and data analysis in a variety of disciplines (e.g., Asbell-Clarke &
Barclay, 1996; Barstow & Berenfeld, 1996; Greenberg, Kolvoord, Magisos,
Strom, & Croft, 1993; Rock et al., 1997).
Learning objectives for climatology and inquiry through visualization. The

combination of visualization and climatology has provided us with a set of inte-
grated learning objectives coveringboth inquiry and science content. At the level of
general inquiry skills the objective that has guided our design efforts is to help
learners develop the skills of formulating and refining researchable questions,plan-
ning and conducting coherent investigations,and reporting and applying results. In
the specific domain of visualization and analysis of quantitative data, our objec-
tives for students include the ability to use the techniques of visualization to iden-
tify and quantify patterns and anomalies in data, to employ arithmetic and statistical
techniques to explore the relations between varying quantities, and to present re-
sults and support arguments. In the content area of atmospheric science, our objec-
tive is to help learners develop an integrated understanding of the radiative energy
transfer processes of emission, absorption, and reflection and how those physical
processes interact with the physical geography of the earth and atmospheric chem-
istry to determine climatic conditions on Earth.
THE CHALLENGES OF IMPLEMENTING

INQUIRY-BASED LEARNING
Although inquiry offers compelling opportunities for science learning, there are
many challenges to the successful implementation of inquiry-based learning. For
example, researchers have documented that children have difficulties conducting
systematic scientific investigations (e.g., Kraijcik et al., 1998; Schauble et al.,
1995).Data gathering, analysis, interpretation, and communication are all challeng-
ing tasks that are made more difficult by the need for content-area knowledge. Al-
though we entered this design process with some specific ideas of how technology
could be used to address the challenges of inquiry-based learning, we found that
these challenges appeared in many forms and that responding to them effectively al-
most always required the use of both technological and curricular design strategies.
In this article, we focus on five of the most significant challenges to the successful
implementation of inquiry-based learning. The experiences described in the follow-
ing sections demonstrate that the failure to address any ofthese challenges success-
fully can prevent students from successfully engaging in meaningful investigations
and therefore undermine learning. The five challenges are as follows:
1. Motivation. For students to engage in inquiry in a way that can contribute to

meaningful learning they must be sufficiently motivated. The challenging and ex-
tended nature of inquiry requires a higher level ofmotivation on the part of learners
than is demanded by most traditional educational activities. To foster learning, that
motivation must be the result of interest in the investigation, its results, and their
implications. When students are not sufficiently motivated or are not motivated by
legitimate interest, they either fail to participate in inquiry activities or they partici-
pate in them in a disengaged manner that does not support learning. Motivation is
recognized by Soloway, Guzdial, and Hay (1 994) as one of three primary chal-
lenges for Learner-Centered Design.
2. Accessibility of investigation techniques. For students to engage in inquiry,
they must know how to perform the tasks that their investigation requires, they
must understand the goals of these practices, and they must be able to interpret
their results. Scientific investigation techniques such as data collection and analy-
sis can be complicated and typically require a level of precision and care that are
not required of students in their everyday experiences. If students are not able to
master these techniques, then they cannot conduct investigations that yield mean-
ingful results. The need for tools to be accessible to learners across the full diver-
sity of abilities and prior experiences is another challenge of Learner-Centered

Design raised by Soloway et al. (1994).
3. Background knowledge. The formulation of research questions, the devel-
opment of a research plan, and the collection, analysis, and interpretation of data
all require science content knowledge. In designing inquiry-based learning, the
challenge is providing opportunities for learners to both develop and apply that
scientific understanding. If students lack this knowledge and the opportunity to de-
velop it, then they will be unable to complete meaningful investigations.
4. Management of extended activities. To achieve the ultimate goal of
open-ended inquiry, students must be able to organize and manage complex, ex-
tended activities. A scientific investigation requires planning and coordination of
activity and the management of resources and work products. Students are not typ-
ically asked to manage extended complex processes as part of traditional educa-
tional activities. If they are unable to organize their work and manage an extended
process, students cannot engage in open-ended inquiry or achieve the potential of
inquiry-based learning.
5. The practical constraints ofthe learning context. The technologies and activ-
ities of inquiry-based learning must fit within the practical constraints of the learn-
ing environment, such as the restrictions imposed by available resources and fixed
schedules. Although this challenge may not have the same theoretical importance
as the other four for advancing our understanding of the learning sciences, it has
enormous practical implications for design. A failure to work within the available
technology or fit within the existing schedule in a school will doom a design to fail-
ure. Therefore, meeting the constraints of the environment is a critical consider-
ation in design that must be considered alongside learning needs in the design of
cumculum and technology.
In the sections that follow we present the history of our efforts to address these
challenges in the area of climate science. In the course of this process, we em-
ployed a variety of technological and curricular design strategies, some drawn
from the work of others, some novel. These strategies include
Meaningfiulproblems. A meaningful problem is one that has implications that

matter to students. As many advocates of project- and inquiry-based science have
argued (e.g., Barron et al., 1998; Blumenfeld et al., 1991), a meaningful problem
can be used to establish a motivating context for scientific inquiry.
Staging activities. Staging activities are sequences of structured investigations
that introduce learners to investigation techniques, help to develop background
knowledge, and establish a motivating context. They can be used to set the stage
for open-ended inquiry activities. The progression from staging activities to
open-ended inquiry resembles the progression from problem-based to pro-
ject-based learning described by Barron et al. (1998).
Bridging activities. Bridging activities are designed to bridge the gap between
the practices of students and scientists. Bridging activities employ practices that
are familiar to students as a means of introducing unfamiliar scientific practices.
Supportive user-interfaces. A supportive user-interface provides scaffolding
for learners by embedding the tacit knowledge of an expert in the user-interface.
Supportive user-interfaces can make investigation tools more accessible to learn-
ers. Our approach to supportive interfaces resembles the use of scaffolds in
Learner-Centered Design (Jackson et al., 1996; Soloway at al., 1994).
Embedded information sources. An embedded information source is a library
of resources that is linked directly to an investigation tool. They can provide leam-
ers with just-in-time access to information that can provide the background knowl-
edge necessary to complete an investigation. The use of information sources in this
way resembles the use of Ask Systems (Ferguson, Bareiss, Bimbaum, & Osgood,
1992) in the Goal-Based Scenario architecture (Schank, Fano, Bell, & Jona,
1993/1994).
Record-keeping tools. Record-keeping tools allow learners to record the pro-
cess and intermediate products of an extended activity. They can be used to support
the management and organization of inquiry activities.
In the design history presented below, we describe the selection and use of each
of these strategies in the course of the iterative design process. Following the de-
sign history, we elaborate on these strategies and their value for addressing the
challenges of inquiry-based learning.
INTRODUCTION TO THE DESIGN HISTORY
Since we started this research in 1992, we have developed four generations of sci-
entific visualization environments for archival climate data.' This research began
under the Learning through Collaborative Visualization (CoVis) Project (Pea,
1993) and has continued under the Supportive Scientific Visualization Environ-
ments for Education Project and the Center for Learning Technologies in Urban
Schools (LeTUS). All are National Science Foundation-sponsored, educational
technology research efforts targeted at the reform of science education at the mid-
dle and high school levels. This particular research on scientific visualization for
learners is one component of a broad effort to provide authentic science experi-
ences for learners (Edelson et al., 1996b; Pea, 1993).
'We have also developed visualization environments for the investigation of real-time weather data.
That work, in collaboration with researchersin the Department of Ahnospheric Sciences at the Univer-
sity of Illinois at Urbana-Champaign,is described elsewhere (Fishman & D'Amico, 1994; Plutchak et
al., 1998).
Our adaptation of scientific visualization technologies for learners began with
an examination of the way that researchers use scientific visualization in their
work and an evaluation of the potential for scientific visualization as a tool to sup-
port learners (D'Amico et al., 1994; Gordin & Pea, 1995). Our earliest design ef-
forts focused primarily on technological issues, such as designing user-interfaces
to scientists' tools that provided scaffolds to enable learners to participate in the
practices of scientists. These efforts were guided by the principles of
Learner-Centered Design (Soloway et al., 1994). As we recognized the limits of
technology to fully address the challenges of inquiry-based learning, we broad-
ened our efforts to include curriculum design.
The design and evaluation described here were conducted using an informal,
collaborative approach. Design was done by teams at Northwestern University
that included faculty in education and computer science, graduate students, profes-
sional programmers, and practicing teachers. Input and feedback were sought fre-
quently from content area scientists. Use of the software and curriculum were
observed both in classroom and laboratory settings. Records of these uses were
collected through a combination of direct observation by the research team, video-
taping, interviews and journals of teachers and students, and unsolicited feedback
from teachers. With a few exceptions, the evaluation was informal because we
were involved in a rapid cycle of iterative design and implementation. The issues
discussed here were all revealed by clearly identifiable problems in classroom ex-
periences, such as the failure of learners to engage in inquiry or their inability to
complete an inquiry successfully with the time and resources available. As we be-
gin to see evidence that these obvious problems have been resolved, we are shift-
ing our efforts increasingly from identifying problems to characterizing impacts.
We are therefore launching some more formal evaluation efforts, which we dis-
cuss in the final section of this article.
Although the following discussion focuses primarily on the issues that were
raised as a result of our observations of students in classrooms, there were many
other influences on the design process. These influences included (a) scientific re-
searchers, through observation of their practices and advice they offered; (b) edu-
cational researchers, through writings, examples of software and curriculum, and
conversation; and (c) teachers, through observation of their practice and advice
they offered.
In the following sections, we describe the four generations of visualization en-
vironments and the curriculum we developed to accompany them. The four visual-
ization tools are the Climate Visualizer, Radiation-Budget Visualizer, Greenhouse
Effect Visualizer, and Worldwatcher. Throughout the design history, certain basic
characteristics of the visualization tools have remained the same. They all display
data in the same two-dimensional, gridded color map format, and they were all de-
signed to provide an integrated suite ofdata and tools. For each design, we summa-
rize the considerations that were most important in the design, data, architecture,
and features of the software, the cumculum that was designed to go with the soft-
ware, and its use. We then discuss how the design attempted to address the chal-
lenges of fostering inquiry-based learning and the issues that were raised by the
use of the design by teachers and student.
THE CLIMATE VISUALIZER
The Climate Visualizer (Gordin, Polman, & Pea, 1994) was the first visualization
environment we developed. It was first distributed to teachers participating in the
CoVis Project in the fall of 1993.Its design was heavily influenced by the research
and teaching practices of our collaborating climatologist. In fact, we conceived of
the Climate Visualizer as a direct adaptation of a scientist's tool for use by learners.
The goal of the Climate Visualizer was to put the powerful features of scien-
tists' visualization environments into the hands of high school students. Because
of the complexity of their user-interfaces and the challenges they posed to inter-
preting visualizations, we concluded that the actual tools used by scientists were
inaccessible to mainstream high school students. Therefore, the primary design
goal of the Climate Visualizer was to respond to the challenge of making investi-
gation techniques accessible to learners. At this stage in the design process, we
viewed accessibility as a technological challenge, which we addressed by design-
ing a supportive user-interface for the Climate Visualizer. This supportive inter-
face provided significantly more structure and information in its user-interface
than a comparable tool designed for scientists.
Data Library
The Climate Visualizer's data library contained basic weather data for theNorthern
Hemisphere above 20" latitude, covering a period from the early 1960s to the late
1980s. The data included measurements of temperature, wind speed and direction,
and atmospheric pressure at several levels in the atmosphere, all recorded at 12-hr
intervals. These data are used by researchers to investigate general climate patterns
and specific large-scale weather events in the Northern Hemisphere.
Software Architecture, Features, and Interface
In contrast to the general-purpose visualization environments used by scientists,

the Climate Visualizer's interfacewas customized to the data in its data library. The
Climate Visualizer was implemented as a "front-end" to Transform, a powerful sci-
entific visualization environment that is widely used in the research community.
Users selected data to view in the Climate Visualizer by specifying a date, time of
day, and one or more variables from a menu of available data. The data were then
downloaded over the Internet from a server, rendered by the Transform visualiza-
tion engine, and transferred to the Climate Visualizer for display. To enhance per-
formance, data sets were cached on the local computer for future use once they had
been downloaded.
Like the visualizers that succeeded it, the Climate Visualizer presented visual-
izations in the form of gridded, two-dimensional color maps in a rectangular pro-
jection (Figure 2). Temperature was displayed as the main color map, atmospheric
pressure as contour overlays, and wind speed and direction as vector overlays. The
Climate Visualizer provided users with the ability to customize their visualizations
by varying the colors that were used to represent data values and by changing the
numerical range of the color palette. The Climate Visualizer supported compari-
son of data sets by allowing users to view them simultaneously side by side or by
subtracting them from each other.
Curriculum Design
We did not create any cumculum to accompany the Climate Visualizer initially. At
the time we did not feel predesigned cumcula were appropriate because we were
hoping to foster an entirely student-driven form of inquiry-based learning in which
students would generate and pursue their own research questions. We designed the
Climate Visualizer with the expectation that the data and tools in the Climate Visu-
alizer would lead students to pose questions and pursue them through open-ended
investigations.To the extent that we engaged in curriculum design at all, that design
consisted of recommending to teachers that they have their students explore the
data in the Climate Visualizer in an open-ended fashion to identify interesting ques-
tions to investigate and then pursue those questions with the guidance of the teacher
through open-ended inquiry.
Use
The Climate Visualizer was first used by students in a workshop led by CoVis
teachers in the summer of 1993. In the 1993-1 994 school year, it was used for a pe-
riod of 2 weeks in the classes of one of the six teachers participating in the CoVis
project at that time. It was used for a few days in the classes of two of the other
teachers that year.
Issues Addressed by the Design
The Climate Visualizer was designed to address the challenge of making the visual-
ization technologies used by scientists accessible to learners. One of the most sig-
nificant obstacles to the use of scientists' tools by learners is the lack of contextual
knowledge that they provide. Because scientists bring an extraordinary amount of
tacit knowledge about geography, the data being displayed, and the phenomena
that the data represent, they are able to work in a relatively information-poor envi-
ronment. For example, Figure 3 shows a screen from Transform, the popular and
extremely powerful two-dimensional visualization environment designed for re-
searchers that we used as the visualization engine for the Climate and Radia-
tion-Budget Visualizers. To a nonexpert, the visualization is almost completely un-
interpretable. The variable being displayed is encoded in the file name that serves
as the title for both windows. A careful reading of "T850-1-188-OZ" that appears
in the title bar reveals to a researcher who is familiar with this data that he or she is
viewing the average midnight Greenwich Mean Time for January 1988 at the 850
millibar equal-pressure surface in the atmosphere. The geographic range and scale
of the displayed region are not identified, and the numbers that appear in the win-
dow at the side have no units or any indication ofhow their values map into colors in
the visualization. Finally, the operations that are available to the user are all con-
cealed in menus. As lacking in support as this interface is, expert researchers have
no difficulty working with data and visualizations in this form because of the exten-
sive knowledge that they bring to the task.
In contrast to the interface displayed in Figure 3, the supportive scientific visu-
alization interface in the Climate Visualizer provided scaffolds for learners in the
form of contextual information about geography (through latitude and longitude
markings, as well as continent overlays), about units (through labels), and about
the mapping of numerical values to colors (through a color palette display). The
Climate Visualizer allowed users to specify the units in the display, so that stu-
dents who were more familiar with Fahrenheit could interpret visualizations using
values with which they were comfortable. The color palette display in the Climate
Visualizer tracked the movement of the cursor and displayed the numerical value
for the current cursor location over its corresponding color in the color bar. Finally,
all of the operations that the user could perform were displayed as permanently
visible buttons in the Climate Visualizer interface.
A second place that the Climate Visualizer provided support that was not avail-
able in general-purpose scientists' tools was in the interface for selecting the data
to visualize. Comparable tools for scientists typically require users to identify data
sets by specifying file names, which in most data libraries are abbreviated and ob-
scure, as in the "T850-1-1-88-02'' in Figure 3. In contrast, the Climate Visualizer
allowed users to select data from the data library by choosing variable names,
dates, and times from a menu of available options. This required that we customize
the Climate Visualizer to the specific data in its library, which would compromise
the generality that scientists rely on in their tools.
The third place that the Climate Visualizer provided additional support for
learners was in the form of default settings for the display parameters. The Climate
d File Edit lmaae Numbers Templates Tables Ulindows 'fiiP 12:~
?r
> FIGURE 3 A visualization of temperature data for the Northern Hemisphere displayed by Transform, a powerful, general-purpose visualization environment
widely used by scientific researchers.
Visualizer provided learners with access to a number of powerful operations for

the analysis of data, such as the ability to customize the display of visualizations,
sample data values for individual points in an image, select regions of interest, and
coordinate the simultaneous interpretation of multiple data sets. As we began to
work with these operations during the design process ourselves, we recognized
that access to these p o w e h l operations can be a double-edged sword. For exam-
ple, the ability to modify the color palette means that with a simple change to one
display parameter, a visualization can be transformed from a multicolored, highly
patterned color map to a blank, monochromatic field. We learned that an important
component of a scientist's expertise is knowing how to set display parameters so
that a data set will display in a meaningful form. Without adequate support, a nov-
ice who does not yet understand the units or range of values in a data set can spend
a great deal of time working with a visualization just to display it in a way that will
allow him or her to begin to interpret it. To protect learners from undue frustration
in the early stages of interpretation,we created a scaffold in the form of default dis-
play settings for data. For each variable in the data library, we established default
display settings for the color scheme, its range, and the units displayed. For tem-
perature, for example, the default color scheme was made up of five colors ar-
ranged in the order of the rainbow, and the default units were degrees Fahrenheit.
Any time a user selected a data set to view, it was displayed using these defaults.
The defaults served as a scaffold that enabled learners to begin to interpret visual-
izations without needing to set display parameters that they may not yet under-
stand. However, the ability to customize the display of visualizations still allowed
learners to progress to more sophisticated interpretations.
When we observed the Climate Visualizer's use in schools, we found that stu-
dents at a wide range of ability levels were able to use its supportive interface to se-
lect data and to display and interpret visualizations. The first activity that many
students did with the Climate Visualizer was to research the weather conditions on
the day they were born. With little or no instruction, students were able to retrieve
the data for the date they were born, locate the cell on the visualization covering
their birthplace, and use the display to identify temperature. In each of these steps,
students took advantage of the contextual information embedded in the
user-interface. For example, the map overlay and latitude/longitude markings en-
abled students to locate the area they were interested in, and the readout that tracks
the movement of the cursor enabled them to access the underlying data values.
These steps would have been extremely difficult if not impossible using an infor-
mation-poor interface like the one in Figure 3. Unfortunately, we had few opportu-
nities to see whether the supportive interface made more sophisticated analyses of
data more accessible because of other challenges discussed in the next section.
In addition to the issue of accessibility, our design for the Climate Visualizer
also addressed the challenge that students face in managing extended activities in
inquiry-based learning. Based on the advice of experienced inquiry-based science
teachers and prior experience with Project INQUIRE (Hawkins & Pea, 1987), we
were concerned that students would have difficulty planning, organizing, and co-
ordinating the inquiry process when they used the Climate Visualizer. Teachers
with experience supervising open-ended projects told us that teaching students to
manage open-ended activities was the biggest challenge of implementing this ap-
proach, and experienced teachers who were considering this approach for the first
time consistently cited this issue as one of their greatest concerns. To help address
the challenges of managing the inquiry process, we developed the Collaboratory
Notebook. The Collaboratory Notebook is a structured hypermedia environment
that supports asynchronous collaboration across a network (Edelson, Gomez, &
Pea, 1996a; O'Neill & Gomez, 1994). To support inquiry using the Climate Visu-
alizer, the Collaboratory Notebook and the Climate Visualizer were bundled to-
gether as one application. Like other structured hypermedia, collaboration
environments designed for learners, such as CSILE (Scardamalia & Bereiter,
1994; Scardamalia, Bereiter, McLean, Swallow, & Woodruff, 1989), Inquire
(Hawkins & Pea, 1987), and CaMILE (Guzdial et al., 1995), the Collaboratory
Notebook uses structure as a form of scaffold. The Collaboratory Notebook pro-
vides a structure based on a task model of scientific inquiry that encourages stu-
dents to organize their work into such components as questions, conjecture,
evidence, plans, information, and commentary. By combining it with the Climate
Visualizer, we sought to create a place for students to plan and record their investi-
gations, coordinate work efforts among collaborative teams, and receive feedback
from teachers and mentors.
Issues Raised by Classroom Experiences
Although our initial experiences with the Climate Visualizer in classrooms showed
that we had increased the accessibility of visualization as an investigation tech-
nique, we were confronted by unanticipated obstacles to accessibility as well as is-
sues for motivation and compatibility with classroom constraints.
Accessibility. One concern we had about the use of the Climate Visualizer in
classrooms was that it was not being used for the full range of investigation tech-
niques it was designed to support. The Climate Visualizer was designed as an envi-
ronment for the visual interpretation of data. However, in classrooms it was primar-
ily used as an interface for making queries about data values in specific locations. In
the visual interpretation of data, a user looks for patterns in a visualization that are
revealed by the variations in color across an image. For example, an expert looks
for both smooth trends and sharp discontinuities in color. This process plays a role
in meaning-making by revealing patterns that require explanation or that provide
4 10 EDELSON, GORDIN, PEA
evidence for a hypothesis. For example, a novice looking at the visualization of

temperature in Figure 2 might see the darker regions in the middle of the North
American and Eurasian continents as a curious pattern that needs explaining. An
expert would see those same patterns as evidence that proximity to the oceans has a
moderating effect on temperature. The features of the Climate Visualizer that sup-
ported visual interpretation were the ability to customize the display by changing
the color mappings, to display the quantitative difference between two images, to
view multiple variables in the same visualization, and to view multiple visualiza-
tions simultaneously. However, the activity described previously, in which stu-
dents used the Climate Visualizer as an interface for locating data about a particular
place and time, was characteristic of most of the use of the Climate Visualizer. In
this use, each visualization is serving as avisual interface to a database of individual
values, rather than as an interactive image to support visual interpretation of data.
This use failed to take advantage of the real power of visualization and was, in fact,
a questionable use of the particular data in the Climate Visualizer.
In reflecting on the discrepancy between our intended use of the Climate Vi-
sualizer and the actual use, we recognized that we had only partially removed
the barriers to accessibility of visualization. Although we had responded to the
technological barriers through a supportive interface, we had not addressed an
important conceptual barrier. We had failed to provide teachers and students
with an understanding of the full range of visualization techniques that the Cli-
mate Visualizer was designed to support and the role those techniques could
play in facilitating interesting investigations. An illustration of the difference be-
tween the teachers' conceptions of the use of the Climate Visualizer and the de-
signers' is provided by an activity that one teacher designed and conducted in
his earth science classes. In this activity, groups of students were each assigned
a major city in the area covered by the data. They were asked to graph the
weather experienced by that city for a specific month over a 25-year period. To
conduct this investigation, students called up 25 visualizations, sampled the data
point that included their assigned city, entered the data into a spreadsheet, and
then graphed it. In addition to the fact that this activity only used the Climate Vi-
sualizer as an interface to point data, it was also a problematic use of the Climate
Visualizer's data. Because of the coarse resolution of the data (2" latitude by 4"
longitude) and the fact that much of the data was based on models not measure-
ments, the data set was not actually a reliable source of values for individual cit-
ies. When the data set was created, it was intended to support investigations of
processes extending over large regions, not individual cells.
Although we were disappointed that we had not obtained the full use of the Cli-
mate Visualizer that we had sought, we learned an important lesson from this ex-
perience. It demonstrated the role a curriculum consisting of structured inquiry
activities could play in providing students with models of investigation techniques
and their value for pursuing research questions. In retrospect, we realized that we
had relied too much on the tool itself to provide teachers and students with an ac-
tivity model for visualization. Unfortunately, when the teacher stepped into the
breach, he too lacked appropriate models of use to guide his design. Thus this ex-
perience also taught us about the need to provide teachers with good conceptual
models of the tools' uses as well. In later efforts, we addressed this conceptual bar-
rier to accessibility by creating curricula incorporating structured inquiry activities
to accompany the software to provide teachers and students alike with experience
showing the range of available inquiry techniques and their value for conducting
investigations.
Motivation. A related concern raised in these classroom experiences was that

the Climate Visualizer saw little use in extended or open-ended investigations.
Only 7 of more than 353 students chose to use it in their student-directed research
projects. There are a number of reasons for this lack of adoption, including the in-
complete understanding of the Climate Visualizer's capabilities and value that we
discussed previously. However, our observations and conversations with teachers
and students led us to recognize that another part ofthe problem was that the design
failed to address student motivation adequately. The intersection between the data
sets in the Climate Visualizer's library and the interests of students was small and
hard for students to see. The result was difficulty with both creating initial engage-
ment and maintaining engagement over the longer term.
The problem with initial engagement showed up when students were given
the opportunity to "play with the data" in an exploratory fashion, our recom-
mended initial use of the software. The value of the data for investigating gen-
eral weather patterns in the Northern Hemisphere and specific large-scale
weather events was not particularly compelling to students with no additional in-
troduction. They did not have sufficient background in weather to find the ob-
servable weather patterns interesting, and they had no memory of the major
weather events covered by the data or other means to identify them. As we re-
ported above, the most common exploratory activity that students performed
was to look for data about the date and place they were born. Answering this
question was clearly motivating. Almost universally when students were invited
to "explore" the data, they started with this activity, spontaneously "inventing" it
when it wasn't suggested to them. However, this investigation was almost al-
ways followed by a visible letdown when students found the answer. After they
had compared with their friends, there was no place to go with the information.
In general, this investigation turned out to be a dead-end activity for students be-
cause it did not lead students to notice or be interested in other aspects of the
data. To our great disappointment, it was extremely rare for students to sponta-
neously pose other questions or to browse through data sets to look for interest-
ing or curious patterns.
The problem of maintaining engagement with the Climate Visualizer showed

up in those cases when its use did lead students to raise their own questions, be-
cause those questions, more often than not, could not be pursued within the tool.
The fact that the data were limited to meteorological variables and they only cov-
ered a portion of the Northern Hemisphere meant that many of the questions raised
by students were not researchable within the tool. For example, some students ex-
pressed interest in investigations of Africa but were thwarted because the data only
covered the Northern Hemisphere. Others asked about the causes and effects of the
weather phenomena that the Climate Visualizer displayed and about the relation
between climate and human activities, but they were unable to access any data
about physical or human geography to explore these questions. Finally, both
teachers and students repeatedly asked if it were possible to investigate global
warming, a scientific controversy that was receiving considerable press coverage
at the time. This failure to maintain engagement was particularly disappointing be-
cause in these instances the Climate Visualizer was serving its desired goal of fos-
tering interest on the part of learners but was then frustrating them by not allowing
them to pursue the interests it had raised. Although we were disappointed by the
inability of the Climate Visualizer to capitalize on these expressions of interest,
they helped to convince us of the ability of climate science to connect to the inter-
ests of students, particularly by the large number of students and teachers who ex-
pressed interest in investigating global warming.
With the benefit of hindsight, the reason the Climate Visualizer failed to engage
students as we had hoped was clear. The tool was designed by adapting a set of sci-
entific practices without taking adequate account of the differences between the
interests of scientists and high school science students. The recognition of this
challenge to motivation led us to purse two important strategies in subsequent de-
signs. First, we selected data sets to support investigations that we had reason to
believe would interest our population of students. Second, we created cuniculum
and activities designed to create initial engagement on the part of students that cap-
italize on that engagement to motivate extended, open-ended investigations.
Practical constraints. The final design problem with the Climate Visualizer
was a practical failure to meet the constraints of the classroom.The software, as we
had designed it, took too long to display visualizationsto be practical in a classroom
setting. The need to transfer data across the network and the computation involved
in generating visualizations on the high-end personal computers of the time2meant
that students could experience a delay of up to a minute between requesting an im-
age and viewing it. In a 50-min class, this reduced the productivity of students sig-
ZTheClimate Visualizer and theRadiation-Budget Visualizer were used on Apple Macintosh Quadra
700 computers, which were high-end personal computers at the time.
nificantly, particularly because their attention typically wandered during the delay.
In subsequent designs, we experimented with several different strategies to im-
prove performance in an effort to address this problem.
THE RADIATION-BUDGETVISUALIZER
The design of the Radiation-Budget Visualizer3 addressed several of the issues ex-
posed by our experiences with the Climate Visualizer. For example, in the develop-
ment of the Radiation-Budget Visualizer, we assembled a data library around the
motivating issue of global warming. We also developed a curriculum to introduce
visual interpretation and data analysis activities to students and teachers. Finally,
we modified the software architecture to improve its performance.
Data Library
In contrast to the Climate Visualizer's regional data, the Radiation-Budget Visual-

izer's data sets covered the entire world. These global data sets were selected to
support investigations of energy transfer in the earth-atmosphere system and the
influence of the sun, Earth's physical geography, and the atmosphere on tempera-
ture. The Radiation-Budget Visualizer 's variables included incoming solar energy
(insolation), earth-atmosphere reflectivity (albedo), terrestrial absorbed solar en-
ergy, outgoing long-wave radiation, net energy balance, the greenhouse effect, and
surface temperature. The data showed monthly averages for one year, 1987. In ad-
dition to the data based on measurements of the actual Earth, the Radiation-Budget
Visualizer also included data showing two hypothetical scenarios, an Earth with no
atmosphere and an Earth with an atmosphere but no clouds.
Like the Climate Visualizer, the Radiation-Budget Visualizer was implemented as

a front-end to Transform. It displayed visualizations in the same two-dimensional
grid format, dynamically displaying the latitude and longitude position of the cur-
sor and the data value at that position. It improved on the Climate Visualizer's per-
formance by storing all of its data on the local computer, removing the need to
transfer data across a network before rendering them.
'The term radiation-budget is used by scientists to refer to the balance between incoming and outgo-
ing energy in the earth-atmosphere system, which determines whether the earth experiencesa net cool-
ing or warming over a specific period of time.
The interface to the Radiation-Budget Visualizer's data library was spare. It
consisted of a menu for selecting a month of the year and a table for selecting the
variables to view. Each column in the table represented one of the three models of
the earth (i.e., the earth as observed, the earth modeled without clouds, and the
earth modeled without an atmosphere), and each row corresponded to a variable.
Thus a user could request a visualization of a variable under a specific model by se-
lecting the appropriate cell of the table.
Curriculum Design
The Radiation-Budget Visualizer was designed to support activities in which learn-

ers compare data for the two hypothetical scenarios to the actual earth system val-
ues. In particular, the scenarios enabled students to investigate the role of clouds
and the atmosphere in determiningglobal climate. To help foster appropriate use of
the Radiation-Budget Visualizer, we worked closely with an environmental sci-
ence teacher to develop a structured, inquiry-based activity for it. The activity de-
sign took the form of a two-page handout with instructions for tasks to do with the
Radiation-Budget Visualizer and nine questions to answer. The handout was very
similar to the hands-on laboratory activities to which the teacher's students were
accustomed. The activity began with very specific instructions for simple
data-querying tasks (e.g., identify the incoming solar radiation, albedo, and surface
temperature for Chicago in February 1987), then moved to questions that asked for
visual interpretation (e.g., identify the relation between solar radiation and surface
temperature), and then open-ended questions with no specific instructions (e.g.,
what is the influence of clouds?). The progression of this activity from simple activ-
ities to complex and from specific instructionsto open-ended tasks was designed to
expose students to the techniques of visualization in a guided fashion initially and
then allow them to gain experience deciding which techniques to use through more
open-ended tasks.
Use
The Radiation-Budget Visualizer was released to participants in the CoVis Project

in the winter of 1994 and was used by one environmental science teacher that year.
It was also used to support research projects by a few students in the classes of two
Earth science teachers that year.
The design of the Radiation-Budget Visualizer and its accompanying cumculum

activity attempted to address motivation, accessibility, and practical constraints.
The greenhouse investigation activity was intended to address challenges to

both motivation and accessibility. The activity was designed to capitalize on the
global warming controversy as a motivating context by introducing the green-
house effect and leading students through an investigation of it. It addressed the
accessibility issue raised by the Climate Visualizer by introducing students to the
range of investigation techniques supported by the tool. In practice, we observed
that the early questions with more specific instructions helped introduce students
to the process of looking for patterns in data. Although the depth of their analyses
and the quality of their answers varied widely, we did observe students using the
Radiation-Budget Visualizer to support the sort of visual interpretation we had
missed with the Climate Visualizer.
The change in data sets was designed to address the challenge of motivation.
The incorporation of data showing the greenhouse effect was designed to capital-
ize on the motivating aspects of the global warming controversy.
The practical constraints of the classroom were addressed in the design of the
Radiation-Budget Visualizer by removing the need to transfer any data across the
network. This improved performance, but, as we also note in the following, did not
adequately address the challenge of meeting classroom constraints.
Our classroom experiences with the Radiation-Budget Visualizer were limited, but
they helpedus to focus on an important issue ofbackground knowledge that we had
observed but not addressed previously, and they revealed that we had not ade-
quately met the challenge of practical classroom constraints.
The need for background knowledge to support investigations became a clear
issue with the Radiation-BudgetVisualizer. Although the Climate Visualizer inad-
equately addressed the need for background knowledge as well, it was not as
clearly apparent in classrooms because of the limited way in which it was used. In
the class that used the Radiation-Budget Visualizer, students were often confused
by the interface to the data library because it relied on their understanding what
each of the variables and the hypothetical models represented. Understanding
what the data sets in the tool represented and their relation to natural processes was
the goal of the activity, but the design of its user-interface made them prerequisite
knowledge for using the tool. As a result, it fell to the teacher to provide explana-
tions for the variables and models in the user-interface. However, because activi-
ties were conducted in self-paced groups, the teacher was faced with the dilemma
of explaining things to the whole class at one time when the information was rele-
vant to some groups and not others, or to be available for each group when they re-
quired it. One of the reasons that the other teachers cited for not using the
Radiation-Budget Visualizer in their classes was that they were concerned about
their ability to provide their students with the background knowledge that they felt
the tool required.
The final obstacle to inquiry-based learning with the Radiation-Budget Visual-
izer was that the new architecture had only partially achieved the objective of
meeting the practical constraints of the classroom. As with its predecessor, the
computational requirements of rendering visualizations from large data sets
strained the capacity of the classroom computers. The slow performance remained
a problem for use in standard-length class periods. In addition, the storage require-
ments of the data sets and the cost of the commercial application Transform meant
that the Radiation-Budget Visualizer was not suitable for broad distribution to
schools.
THE GREENHOUSE EFFECT VISUALIZER
The Greenhouse Effect Visualizer (Gordin, Edelson, & Pea, 1995) succeeded the
Radiation-Budget Visualizer and was the first of the visualizers that was suitable
for broad use. The Greenhouse Effect Visualizer was created by simplifying the
Radiation-Budget Visualizer and adapting it to the World Wide Web. In develop-
ing curriculum to accompany the Greenhouse Effect Visualizer, we attempted to
addressed the issue of motivation even more completely by creating an expanded
curriculum unit that connected the science more directly to the social and political
implications of global warming. The issue of background knowledge was ad-
dressed by incorporating documentation and scientific explanations into the
user-interface, and the constraints of the classroom were addressed by moving to a
web-based interface. In addition, with the Greenhouse Effect Visualizer, we intro-
duced a new aspect to the visual interpretation of data by providing data sets at mul-
tiple temporal and spatial resolutions.
Data Library
The Greenhouse Effect Visualizer displayed the same variables showing the trans-
fer of solar energy through the earth-atmosphere system as the Radiation-Budget
Visualizer. However, the data showing the hypothetical models ofthe earth with no
clouds and no atmosphere were eliminated to simplify the interface and reduce the
storage requirements. A new feature of the Greenhouse Effect Visualizer's data li-
brary was the inclusion of data for a variety of temporal and spatial resolutions.
Each variable was available at monthly, seasonal, and yearly temporal resolutions,
and at 2.5 x 2.5,s x 5, 10 x 10,20 x 20,60 x 60, and 360 x 180" spatial resolutions.
The objective of making these different resolutions available was to enable stu-
dents to explore patterns and averages over different temporal and spatial scales.
This is particularly important for investigating global warming, for which the im-
pact is characterized by a change in an annual, global average, whereas its impacts

and causes are distributed throughout the world and across the year. As with the Ra-
diationBudget Visualizer, the data only covered 1987.
The architecture for the Greenhouse Effect Vis alizer departed dramatically from
its predecessors. In the Greenhouse Effect Visualizer, all the available visualiza-
tions were rendered in advance and stored on a web server. Users accessed data by
selecting a variable, time interval, and spatial resolution from a list (Figure 4). Vi-
sualizationswere displayed in interactive web pages that allowed the learner to se-
lect a point in a visualization and view its latitude, longitude, and data value (Figure
5). The use of the World Wide Web greatly reduced the computational require-
FIGURE 4 The interfacetothe data library in the web-based GreenhouseEffectVisualizer.

FIGURE 5 Data displayed in the web-based Greenhouse Effect Visualizer
ments needed to access the visualizations, although it added the requirement of an

Internet connection. In addition, the performance was improved over the previous
visualizers by generating all of the images in advance and caching them on the
server. Because of the need to prerender all the visualizations, however, the ability
to customize the display ofdata by modifying the color palette or mapping ofcolors
to values was eliminated.
The other major addition to the Greenhouse Effect Visualizer was the inclusion
of explanatory materials linked to the visualization interface. This material pro-
vided explanations of the individual variables available in the Greenhouse Effect
Visualizer as well as a description of the basic energy balance cycle that was illus-
trated by visualizations from it.
Curriculum Design
In the fall and winter of 1995-1 996, we developed a 5-week project-based curricu-
lum entitled The Student Conference on Global Warming. In contrast to the narrow
focus on the greenhouse effect of the activity developed for the Radiation-Budget
Visualizer, the Global Warming Conference covered a wide range of scientific is-
sues associated with the global warming controversy. The development of the
Global Warming Conference was part of a new focus in the CoVis project on the de-
velopment of extended curriculum. Unlike the Radiation-Budget Visualizer activ-
ity, it was not developed specifically around the opportunities provided by scien-
tific visualization techniques. The conference was designed to be an engaging,
coherent, inquiry-based unit to achieve learning objectives in earth and atmo-
spheric sciences. In doing so, it took advantage of the global warming controversy
as a unifying, motivational context. As an extended inquiry-based unit, it drew
upon a number of investigation techniques and sources of data, only one of which
was the Greenhouse Effect Visualizer. The overall curriculum consisted of a 2-day
introduction to the global warming controversy, a 2-week sequence of structured 1-
to 2-day inquiry activities, and a 2-week open-ended investigation of a specific
global warming issue or country. The project unit culminated in a mock interna-
tional conference on global warming that could be conducted either within a single
classroom or by classes in multiple schools over the Internet.
The structured activities introduced learners to the science and controversies
surrounding global warming. One of these activities took advantage of the Green-
house Effect Visualizer to explore seasonal changes in solar radiation and their in-
fluences on Earth's surface temperatures. In the first part of this activity, students
looked at incoming solar radiation at several different spatial resolutions to help
develop an understanding of the concept of spatial averaging. They then were
asked to characterize the way that incoming solar energy varied by latitude and by
season and to explain this variation using knowledge they had gained through a
prior activity in which they looked at the tilt of the earth's axis relative to the sun in
different seasons. In the final part of the activity they compared surface tempera-
ture visualizations with incoming solar energy to characterize the relation between
solar energy and temperature (Figure 6).
The curriculum materials for this activity included a handout for students that
provided instructions for each stage of the activity and questions for students to an-
~ w e rThis
. ~ activity called for students to employ all of the visualization operations
of the Greenhouse Effect Visualizer. It called for them to use the variable spatial
resolutions to explore the relations between local variation and global average. It
called for them to engage in visual interpretation to characterize large-scale pat-
terns and to compare energy visualizations with temperature. Finally, it called for
41n practice, most teachers modified the handout to suit their own objectives.
them to use the data sampling capability to find specific numerical values for spe-
cific locations. The sequencing of this activity among the structured activities and
the introduction to it that was provided for teachers and students were designed to
place it in context within the larger framework of the global warming controversy.
The structured activities were presented as an exploration of natural and unnatural
variation to answer the larger question, "Are human emissions of greenhouse
gases contributing to global warming?" The different activities looked at natural
and "unnatural" variations in a number of greenhouse-related variables over a va-
riety of time scales, and this one focused on natural variation over seasonal and an-
nual intervals.
Use
The Greenhouse Effect Visualizer was first used by the environmental science
teacher who had previously used the Radiation-Budget Visualizer in the spring of
1995. She used it for a unit on global warming that helped shape the design of the
Global Warming Conference curriculum (Gomez, Gordin, & Carlson, 1995). The
Greenhouse Effect Visualizer was then made available for a broader audience of
schools in the 1995-1 996 school year, a year in which the CoVis project expanded
from 2 local schools to approximately 25 schools scattered throughout the eastern
United States. In that year, the Greenhouse Effect Visualizer and its accompanying
curriculum were used by 24 middle and high school teachers in the CoVis Project.
In addition, we received reports from many non-CoVis teachers, including commu-
nity and 4-year college instructors,who had integrated the Greenhouse Effect Visu-
alizer into classes on earth, atmospheric, and environmental science.
The design of the Greenhouse Effect Visualizer and the cuniculum for the Global
Warming Conference addressed several issues raised by our prior experiences, spe-
cifically, motivation, accessibility, background knowledge, and practical class-
room constraints.
The Global Warming Conference curriculum was designed with the explicit
goal of connecting the climate science to the societal implications of global warm-
ing. This cumculum was designed to introduce the scientific controversy over
global warming and contextualize the investigations of the greenhouse effect us-
ing the Greenhouse Effect Visualizer as part of understanding how global warm-
ing might occur and how it might be affect energy and temperature data. The role
of the mock international conference in the curriculum was to create a demand for
this understanding of the greenhouse effect to deal with the larger social, eco-
nomic, and political implications of global warming. Although students found the
issue of global warming motivating and were engaged by the Global Warming
Conference, we still observed problems with the alignment of the motivation and
the learning objectives, which we discuss in the next section.
The Global Warming Conference curriculum was designed using a framework
that was intended to address both the challenge of making investigation techniques
accessible and the need for background knowledge. This framework, which be-
came the backbone for all curriculum units developed by CoVis, took advantage of
a class of activities that we call staging activities. The curriculum framework be-
gins with an introductory, or hook, activity that was designed to create interest in
the question or controversy covered by the curriculum. This is followed by staging
activities, which are short, structured investigations designed to build understand-
ing of the relevant investigation techniques and help develop scientific under-
standing. In an approach that is similar to the progression from problem-based to
project-based learning described by Barron et al. (1 998), staging activities are de-
signed to provide the background necessary for students to engage in an extended,
open-ended investigation that follows them. This open-ended activity builds to a
culminating, performance-based activity in which students apply the knowledge
gained through the curriculum sequence. Thus staging activities address the chal-
lenge of making investigation techniques accessible by introducing them in a
structured, guided activity, and they address the need for background knowledge
to conduct open-ended inquiry by providing its acquisition in the context of indi-
vidual investigations.
The need for background knowledge was also addressed in the Greenhouse
Effect Visualizer by linking explanatory materials directly to the Greenhouse Ef-
fect Visualizer interface. These explanations were added to provide support for
teachers and students in understanding the nature of the data, how it was gener-
ated, what it measures, what the units are, and how it can be used to understand
climate processes. Explanations of the fundamental processes of energy transfer
in Earth's energy cycle were also included, such as reflection and absorption of
solar shortwave radiation, emission of longwave radiation by the earth, and the
absorption and emission of longwave radiation by the atmosphere (greenhouse
effect). Teachers reported that these explanatory materials were valuable both
for preparing themselves and for supporting students during their use of the
Greenhouse Effect Visualizer.
As for practical constraints, the adoption of the web-based interface allowed us
to make a supportive visualization environment available to a broad audience of
learners for the first time. Rather than requiring high-end personal computers with
large amounts of memory and disk space, the Greenhouse Effect Visualizer could
be accessed from any Internet-connected computer with a color monitor and
enough memory to run a web browser. It was, in fact, used in classrooms with a
wide variety of technology infrastructures.
Although the Greenhouse Effect Visualizer represented a breakthrough in the num-

ber of schools who used it, observations and reports of use still raised important
concerns about motivation, accessibility, background knowledge, and practical
constraints.
Motivation. The Global Warming Conference curriculum was intended to

provide a motivating context for learning about climate based on concern over the
implications of global climate change. This context was designed to motivate not
just the visualization activities using the Greenhouse Effect Visualizer but also
the other structured investigations involving data, graphing software, modeling
software, and hands-on laboratory experiments. In practice, we found that teach-
ers struggled to present the curriculum to their students as a coherent whole. In
several cases, teachers chose to focus on the structured investigations, treating
them like a traditional curriculum unit organized around a topic, not a contro-
versy. Other teachers chose to focus on the societal implications and
de-emphasized the inquiry activities and the climate science. Only a handful of
teachers actually conducted a mock conference, and those that did found it diffi-
cult to connect the science content to the social, economic, and political issues
that were the subject of the conference. Although teachers responded quite favor-
ably to the curriculum, and most who conducted it in 1996 repeated it in 1997,
their positive feedback was based on individual activities that they found valu-
able rather than the overall motivating structure. Based on these results, we de-
cided that global warming was, in fact, a motivating context but that the curricu-
lum had to be more carefully constructed to align the science investigations with
the culminating activity.
We became aware of a new challenge for accessibility and motivation
through our observations of the Greenhouse Effect Visualizer in use. We
started to become aware of populations of students who were following in-
structions and identifying patterns in data but did not show any curiosity about
the phenomena in the patterns they observed. Their responses contrasted with
other students we had observed who found the same patterns to be surprising
and intriguing and were easy to engage in the process of constructing explana-
tions for their observations. One of the places we observed this was in a struc-
tured investigation of seasonal and geographic differences in temperature that
was similar to the seasons activity described previously. The activity was de-
signed to present students with discrepancies in the patterns of incoming solar
energy and surface temperature in a way that would elicit curiosity about the
sources of differences. For example, although New York and Madrid receive
approximately the same average solar radiation in January, their average tem-
perature differs by approximately 20 OF. In fact, many students did find the
differences between the neat horizontal bands of incoming solar energy and
the much messier patterns of surface temperature (Figure 6) to be intriguing,
and they would spontaneously offer possible explanations. However, we be-
came increasingly aware that there were many students who recognized the
difference in the patterns and could describe these differences without finding
them remarkable or feeling a need to explain them. These students were partic-
ipating in the activity in a mechanical fashion. They were not actively resistant
or uninterested; the activity just failed to engage them. Because the students
were noticing curious phenomena without becoming curious about them, we
conjectured that they were not entering the activity with sufficiently strong ex-
pectations about the phenomena they were observing.
To address this problem, we looked for a design solution that would be more ef-
fective in activating expectations on the part of students. The solution we devel-
oped was an introductory activity in which students drew their own visualizations
of temperature with crayons on blank maps of the world. In this activity, students
were asked to draw on any prior knowledge they might have to produce their best
guess of what the average temperatures are like for a particular month of the year
(Figure 7). Students frequently describe their experience of this activity as demon-
strating both how much they know about temperature that they didn't realize and
also how much they don't know. In effect, the activity is an exercise in articulating
prior conceptions, which Hunt and Minstrel1 (1 994) have shown to be effective in
enabling learners to connect new knowledge to old. However, the critical role it
was playing in our activities was activating expectations and eliciting curiosity.
We found that after creating these visualizations representing their preconcep-
tions, students approached the investigation of temperature variation with open
questions and active expectations that produced noticeably more interest in the
subsequent seasons investigation.
After introducing this activity, we recognized it has other benefits as well. It in-
troduces the color map representation by having learners use it expressively before
they use it intepretively. It also creates a bridge to the highly technical activity of
visualization through an activity that learners are already familiar with and enjoy,
drawing with crayons. In fact, we have come to call this type of activity a bridging
activity. Bridging activities are designed with the explicit goal of providing them
with the context or knowledge they require to engage successfully in more scien-
tific activities. Unlike the investigation activities that are adapted from scientific
practices, bridging activities are invented specifically to suit the needs of learners
in bridging the gap to scientific practices. Bridging activities can address motiva-
tional, accessibility, and background knowledge issues. Following on the success
of this activity, we incorporated a visualization-drawing activity into the Global
Warming Conference curriculum and incorporated a drawing feature into the de-
sign of subsequent visualization environments.
Accessibility. An important issue for the accessibility of investigation tech-

niques became apparent after the Greenhouse Effect Visualizer was in use for some
time. Although the Greenhouse Effect Visualizer and its curriculum had made
short, structured investigations accessible to learners within the constraints of
classrooms, it was not being used for extended or open-ended inquiry. In retrospect,
we realized that the reduction in functionality that resulted from the decision to
adopt a web-based architecture with prerendered visualizations precluded the use
of the Greenhouse Effect Visualizer for open-ended inquiries. The loss of the abil-
ity to customize the display of visualizations and the elimination of arithmetic oper-
ations on data sets meant the Greenhouse Effect Visualizer lacked the support re-
quired for in-depth exploration of its quantitative data. These limitations, coupled
with the reduction in the available data, restricted the Greenhouse Effect Visualizer
to support for a small number of short (though pedagogically important) investiga-
tions. This limitation was addressed in subsequent designs by returning to an archi-
tecture that rendered visualizations in real time and provided a large number of op-
tions for the customization of visualizations and the quantitative analysis of data.
Background knowledge. Although we found that the explanatory materials

in the Greenhouse Effect Visualizer helped learners to interpret the visualizations,
we found students still lacked important background knowledge for their investiga-
tions. Even with the explanations, many students had difficulty understanding the
more abstract variables and their units of measurement, such as reflectivity and en-
ergy flux, which are measured in percent and in watts per square meter, respec-
tively. As an example, we often observed students referring to areas with high in-
fluxes of radiation as "hot" because they appeared red in visualizations, even
though the temperature might actually be quite cold, as in Antarctica during the
Southern Hemisphere summer. In subsequent designs, this issue was addressed by
designing activities that provided students with physical experiences involving
these quantities and their units of measurements.
Practical constraints. Finally, in trying to meet the constraints of the class-

room, we found that the dependency on an Internet connection was a problem for
many schools because they had unreliable or low bandwidth network connections.
WORLDWATCHER
Worldwatcher, the most recent visualization environment in this line, continues

to be the subject of active research and development. It represents an attempt to
integrate all the lessons learned from the previous visualizers and curriculum.
WorldWatcher is the first of the visualization environments to be developed by a

professional programmer and, not coincidentally, represents a significant step
forward in terms of software functionality. It has many more of the operations
found in commercial, general-purpose scientific visualization tools, but it retains
the approach to data-specific, supportive interfaces of its predecessors. The first
version of WorldWatcher was released in 1996 with the name Climatewatcher.
ClimateWatcher had a data library assembled specifically for investigations of
global climate change. The software was subsequently enhanced to include an
open data architecture that allows users and cuniculum designers to create their
own data libraries. It was then renamed WorldWatcher to reflect its broader ap-
plicability.
Data Library
WorldWatcher retained the data sets from the Greenhouse Effect Visualizer
showing Earth's energy balance. These data sets were supplemented by addi-
tional data displaying physical geography and human activities. The new data
include Earth surface elevation, dominant ground cover and vegetation, precipi-
tation, human population, and carbon emissions from industry. These data sets
were added, in part, to allow students to examine potential causes and impacts of
climate change. Although the transition to an open data architecture in
WorldWatcher has made it easy to add new data sets, the climate data sets con-
tinue to be the core of the data library.
With WorldWatcher we returned to a standalone, nonnetworked application to pro-

vide a level of functionality that was not possible on the World Wide Web at the
time that the development began. Unlike the three previous visualizers,
WorldWatcher renders its own images rather than relying on a third-party visual-
ization application.
With its enhanced interactivity, WorldWatcher is more like the Climate Vi-
sualizer than its immediate predecessor, the web-based Greenhouse Effect Vi-
sualizer. Like all of the previous visualizers, WorldWatcher displays
two-dimensional false-color visualizations in an interface that provides geo-
graphic referents, including continent overlays and latitude/longitude markings
(Figure 8). It allows users to select preferred units (e.g., Fahrenheit or Kelvin),
and it provides a constant readout showing the location of the cursor (both lati-
tudellongitude and countrylprovince) and the data value at that point. The two
innovations of the Greenhouse Effect Visualizer are also retained in
1 8 0 ~ 1 3 590W
~ 45W 0 4 5 ~ 90E 135E 180E
Longitude
FIGURE 8 A visualization window from the WorldWatcher software displaying surface tem.
perature for January 1987.
WorldWatcher. These are the ability to examine visualizations at varying spa-

tial resolutions (now computed dynamically) and the inclusion of supporting
explanations and materials. In addition, WorldWatcher enables users to cus-
tomize the display of their visualizations by zooming and panning, changing
the palette of colors used to represent data values, and changing the mapping
of colors to numbers.
Besides the functions for tailoring the display of data, WorldWatcher also con-
tains operations for analyzing and creating data with mathematical operations, for
creating data using a painting metaphor, for viewing data using other visual repre-
sentations, and for creating documents incorporating visualizations and
hypermedia links to data sets.
Curriculum Design
For the initial release of WorldWatcher, we adapted the Global Warming Confer-
ence to take advantage of the new capabilities of the software. In the spring of 1998,
as part of the activities of the newly formed LeTUS Center, we developed a new
Global Warming Curriculum, tailored to the middle school science standards for
the Chicago Public Schools. The Global Warming Curriculum is a 6-week project
unit in which learners prepare science briefings for the participants in a fictitious in-
ternational conference on global warming. In addition, as the community of users
and the range of data available in WorldWatcher grow, we have assembled a library
of structured activities and open-ended investigation topics. One example is a sin-
gle-session investigation of an elevation data set for Mars. A second example is a
10-hr structured investigation called the Create-A-World activity, in which stu-
dents create a fictitious world and generate temperature data for it based on an in-
vestigation of Earth's climate (Edelson, 1998a).
Use
The first version of WorldWatcher (Climatewatcher) was released in the spring of

1996. It was used by CoVis Project schools as part of the Global Warming Confer-
ence curriculum in the 1996-1997 and 1997-1998 school years. WorldWatcher
was used in pilot tests of the LeTUS Global Warming Curriculum in three Chicago
Public Schools in the spring and fall of 1998. It has also been used in a variety of un-
dergraduate geosciences courses at Northwestern and elsewhere, and in several
workshop-setting studies conducted by the authors at Northwestern and SRI Inter-
national. More than 2,000 copies of WorldWatcher have been distributed over the
Internet or on CD-ROM since 1996.
WorldWatcher and its accompanying curricula contain new elements to address all
five of the challenges of implementing inquiry-based learning.
Motivation
By enhancing Worldwatcher's data library and curriculum, we have attempted to

resolve the challenges of motivating inquiry more completely than its predeces-
sors. The objective of many of these enhancements is to take better advantage of
global warming as a motivating context. For example, we enlarged the data library
to include human and physical geography data sets to support investigations of the
causes and implications ofglobal climate change that go beyond the cycle of energy
transfer in the earth-atmosphere system. For example, the population and carbon
emissions data sets support investigations of the population impact of predicted
global warming and the contributions of different nations to the buildup of green-
house gases in the atmosphere. The ability to investigate these causes and implica-
tions of visualization makes it easier for curriculum designers and teachers to con-
nect the motivating social aspects of global warming to the climate science. In
addition, the broadened data library also supports the goal of maintaining engage-
ment by allowing students to pose and pursue their own research questions. This is-
sue was raised by our experiences with the Climate Visualizer but not effectively
addressed by the Radiation-Budget Visualizer or Greenhouse Effect Visualizer.
Worldwatcher's larger library of related data and the ability to import new data
make it possible for WorldWatcher to accommodate a much broader range of stu-
dent-generated investigations than its predecessors could.
The curriculum was also revised to address some of the problems of coherence
that showed up in the CoVis Global Warming Conference cuniculum. In the
LeTUS Global Warming Curriculum developed in 1998, the motivating context is
still provided by an international global warming conference, but rather than play-
ing the role of participants in that conference, the new curriculum establishes their
role as scientific advisors to the participants. The difference is important because it
directs students' attention away from the diplomatic issues of an international con-
ference back toward the scientific issues and the potential implications of global
warming and the prevention measures being considered. At each stage in the 6- to
8-week curriculum, students must prepare briefings designed for representatives
of the United Nations or individual countries. These briefings motivate investiga-
tions of topics such as, "How would we know if Earth were warming?" "How
could changes in the atmosphere lead to global warming?'and "Predictions and
solutions."
Accessibility of Investigation Techniques
The design of WorldWatcher addresses the challenge of accessibility at both ends

of the inquiry spectrum. It incorporates new features to increase the initial accessi-
bility of scientific visualization as a technology for learning, and it introduces more
powerful investigation techniques to make more extensive and sophisticatedinves-
tigations possible.
Interpreting visualizations. Beginning in 1996, the CoVis Project ex-

panded to include middle school science teachers. In conversations about the
Greenhouse Effect Visualizer, they raised concerns about their students' abilities to
interpret visualizationsas gridded representations of data, aiopposed to pictures or
maps with colored regions. To make the basic interpretation of these visualizations
more accessible, we added several features to WorldWatcher.
Two features address the challenge of understanding a visualization as a grid of
numerical values. The first is an operation that allows the learner to magnify the vi-
sualization to a point where each cell in the image displays its numerical value su-
perimposed on its color (Figure 9). This feature provides a visual analogy to
spreadsheets, a gridded data representation that is familiar to many students. The
second is a data creation tool that allows users to input data by painting colors on
visualizations. The interface for painting data is modeled on conventional paint
programs with a paintbrush for drawing cell by cell and a paint can for filling in re-
gions. Users set a data value to paint by typing it in or by using an eyedropper to se-
lect the corresponding color from the color bar or visualization. The drawing
feature picks up where the crayon-drawing activity described previously leaves
off. The crayon drawing introduces the concept of color as value, but its spatial res-
olution is determined by the sharpness of the crayon, and the chromatic resolution
is determined by the number of crayons. In the WorldWatcher tool, students are re-
stricted to the gridded spatial layout of cells and the spatial resolution associated
with a particular visualization. Similarly, in u ~ i n gthe color bar as a paint palette,
they have the numerical resolution correspondingto the 254 colors in the color bar,
but they can bypass the 254-value restriction by typing in numerical data values at
a higher precision. Informal observationsand teacher reports indicate that working
with these drawing tools helped students to understand the relation of the visual
image to the underlying data when they began working with the actual data in
Worldwatcher's data library.
Surface Temperature January 82 (F)

Surface Temperature: The temperature at the surface of the earth.
Longitude
FIGURE 9 A visualization magnified to show the individual data values in each cell.
The second concern of teachers about interpreting visualizations that we ad-

dressed in WorldWatcher was the distortion of geography involved in projecting
the spherical Earth onto a rectangular grid of data. Several experienced teachers
complained of having struggled to teach geometric projections to students over
their whole careers. To help with this challenge to understanding, we added two
additional operations to WorldWatcher. The first displays data in a projection
that preserves the relative area of regions at different distances from the equator
more accurately than the default rectangular projection. The second displays vi-
sualizations as cut-and-fold diagrams that can be printed out and assembled into
polyhedra-resembling globes.5 Both projections are shown in Figure 10. Al-
though neither of these provides the interactivity of the default rectangular rep-
resentation in WorldWatcher, they provide opportunities for teachers to conduct
bridging activities in which learners understand the effects of projection.
Locating data in the data library. In the design of WorldWatcher, we

identified a new opportunity to embed the tacit knowledge drawn on by scientists in
the user-interface. In the Greenhouse Effect Visualizer and many online libraries of
scientific data, the interface for selecting data consists of list of variable names,
which are typically a challenge for learners because of the technical nature of the
names. The interfaces for selecting data in most general-purpose scientific visual-
ization tools are generally even less accessible because they force users to locate
data in files relying on the file system directory structure and file names to indicate
where a particular data set can be found. In the Greenhouse Effect Visualizer, we at-
tempted to address this problem by linking background knowledge and documenta-
tion directly to the user-interface. In WorldWatcher, we created a mechanism to al-
low designers to embed contextual information directly into the interface to the data
library. Specifically, WorldWatcher provides for the use of conceptual diagrams as
data library interfaces. For the radiation-budget data in WorldWatcher, the inter-
face to the data library takes the form of an energy balance diagram showing the
transfer of energy as it arrives from the sun and passes through the earth-atmo-
sphere system (Figure 11).
When a user selects a particular variable name in this diagram, he or she is pre-
sented with a list of available dates from which to choose. The diagram provides a
visual representation of the contents of the library and indicates how the variables
relate to each other. To allow for extensions to WorldWatcher, the software pro-
vides a mechanism for any user to create these diagrammatic data interfaces.
T h e inspiration for this feature was Michael Eisenberg's hypergami project (Eisenberg & Nishioka,
1997).
CHALLENGES OF INQUIRY-BASEDLEARNING 435
Access to powerful investigation techniques. The design of WorldWatcher

also addressed accessibility at the other end of the inquiry spectrum, that is, access
to the tools for interpretationand analysis of data to support extended, sophisticated
investigations. In this respect, WorldWatcher represents a large swing of the pen-
dulum away from the Greenhouse Effect Visualizer, which did not allow for any
manipulation of its pre-rendered visualizations and provided no operations that
could be performed on the underlying data. In contrast, WorldWatcher provides a
broad array of options for customizing the display of visualizations, performing
quantitative analyses of the data, and viewing data in other visual representations.
We highlight four categories of these features in the following.
Appearances. In WorldWatcher, an appearance is a data structure that de-

scribes a visualization's displayed units, color scheme, range of displayed values,
magnification, and spatial and chromatic resolution. WorldWatcher allows the
user to customize the appearance of any visualization and to apply one visualiza-
tion's appearance parameters to other visualizations to make them visually com-
parable. Appearance data structures provide the mechanism for establishing de-
fault settings for the display of specific variables, a scaffold for learners that we
first used in Climate Visualizer. The range of appearance settings in
WorldWatcher is comparable to those of general-purpose visualization tools used
by scientists, providing learners with the same sophisticated tools for visual inter-
pretation as scientists.
Selection regions. WorldWatcher provides a number of mechanisms for se-
lecting regions within visualizations. Selection regions enable learners to iden-
tify and focus on specific areas of interest. The mechanisms for creating
regions include a rectangular selection tool, a lasso for drawing region bound-
aries, and select-by-location (e.g., Australia or Utah) and select-by-value (e.g.,
between 32 O F and 72 O F ) operations. Region masks can be combined using
Boolean union, intersection, and subtraction operations; can be copied from
one visualization to another; and can be stored in files. Thus regions allow us-
ers to make sophisticated queries to identify areas of the world with specific
data attributes.
Mathematical and statistical analyses. WorldWatcher allows students to ana-
lyze and create data through simple statistical and mathematical operations. It pro-
vides summary statistics (max, min, and mean) both for an entire visualization and
for any selected region. WorldWatcher also allows students to manipulate the data
in an image by applying the basic arithmetic operations at every point in an entire
image or a selected region. A second set of arithmetic operations enable students to
perform binary operations on data sets, including addition, subtraction, multiplica-
tion, division, maximum, minimum, and a form of correlation. So, for example, a
learner can investigate seasonal differences by subtracting January's temperature
from July's.
Alternative visual representations. To give users more leverage in tracking pat-

terns in data over space and time, WorldWatcher provides a number of other visual
representations in addition to the standard color maps including animations, histo-
grams, line plots, and scatter plots.
Background Knowledge
The design of WorldWatcher addresses learners' needs for background knowledge

through both technology and curriculum design.
The technology design expands upon the knowledge resources that were first
integrated into a visualization environment in the Greenhouse Effect Visualizer.
WorldWatcher includes an extensive multimedia database of explanatory materi-
als that we call a MediaBase. The knowledge resources we have assembled cover
three areas: science content, the data library, and the techniques of visualization
and data analysis. The science content resources cover topics related to climate and
climate change. The data library documentation describes the sources and uses of
the data sets. The visualization and analysis resources provide documentation on
the tools and operations available within WorldWatcher and their use in investiga-
tions. The contents of the MediaBase are accessible from within the visualization
environment to enable users to access it on an as-needed basis.6
The curriculum design incorporates activities to help develop the scientific
knowledge required to engage in investigations of data successfully. For example,
one of the challenges exposed by the Greenhouse Effect Visualizer and its prede-
cessors was the difficulty that students had in understanding the quantities mea-
sured by the data and their units of measurement. The LeTUS Global Warming
Curriculum addresses this issue by integrating concrete, hands-on experiences
with computer-based activities. For example, in their investigation of the influence
of reflectivity on climate, students begin with a traditional science lab in which
they place thermometers in envelopes of different colors under a light source. They
track the temperature in each envelope to see the effect of reflectivity on tempera-
ture. This activity is designed to provide students with a physical experience of re-
flectivity on a familiar scale as preparation for a WorldWatcher activity in which
they explore the reflectivity of Earth on a global scale. In this investigation, they
look at the relation among ground cover, reflectivity, and temperature.
6TheMediaBase was still under development at the time ofwriting. It currently includes documenta-
tion to assist students and teachers with the use of WorldWatcher and background information about the
available data. Explanatory resources for science content are currently distributed on paper as part of the
Global Warming curriculum materials.
Organizing and Managing Open-Ended Inquiry
Although the Climate Visualizer included the Collaboratory Notebook as a sup-

port for managing open-ended inquiry, the simple, short-term activities students
conducted with the visualizer did not require the support of the Collaboratory
Notebook. However, as we returned to the goal of facilitating extended inquiry
with WorldWatcher, we felt it was necessary once again to provide support for
managing the process and products of short investigations. In discussions with
teachers about investigations their students had conducted in the past, many
complained that students would forget from day to day what they had done and
what they had left to do in an activity. These experiences support the conclu-
sions of others that learning to record and use the transient thoughts, actions, and
products of the inquiry process is an important part of learning to conduct suc-
cessful open-ended investigations (de Jong & van Joolingen, 1998; Loh et al.,
1997; Thornburg & Pea, 1991). Maintaining an accurate record is necessary to
support the activities of planning, monitoring, reflecting, and communicating
that are critical for conducting and learning from inquiry. Unfortunately, many
students do not develop the habits and skills of maintaining good records be-
cause they are not necessary to be successful at traditional school practices. To
have the opportunity to master these skills, students must have tools that support
them in maintaining records.
WorldWatcher incorporates support for record-keeping in the form of note-
book and annotation facilities. WorldWatcher notebooks are multimedia docu-
ments that can contain formatted text, graphics, and hypermedia links to
visualization data and other notebooks. Students create notebooks using a text
editor that allows them to embed graphics and links. Visualization links allow
users to store a visualization in a file with all of its current appearance settings.
This capability allows students to save the state of their inquiry to return to it
later or to allow others to recreate their investigation. The annotation capabili-
ties allow users to associate specific notebooks with a visualization and to doc-
ument visualizations using a title and brief description that displays in the
visualization window. To assist students in the tracking process,
WorldWatcher automatically annotates visualizations that are created through
mathematical operations with a documentation string that describes the source
data and the operation that was used in the creation of that visualization. In a
related research effort, we have developed a separate application called the
Progress Portfolio (Loh et al., 1997), which provides even greater support for
organizing, annotating, and communicating about the products of the inquiry
process. The Progress Portfolio was designed to be used in conjunction with
any investigation environment. It includes a "Data Camera" that allows leam-
ers to store and annotate images from an application like WorldWatcher in the
Progress Portfolio without disrupting the flow of their work by requiring them
to switch from one program to another. Because the Progress Portfolio has
only recently been released, we do not yet have any experience with students
using it and WorldWatcher together.
Practical Constraints
Like the Radiation-Budget Visualizer, WorldWatcher avoided the network perfor-

mance issues that were an obstacle to use of both the Climate Visualizer and the
Greenhouse Effect Visualizer in classrooms. Because its data sets are available on a
CD-ROM, it does not place extraordinary demands on storage either. Finally, al-
though it places demands on computational power that are similar to its predeces-
sors, the steady advance in technology has reduced that challenge. The least expen-
sive Macintosh computers that have been on the market since 1996 all provide the
computational power that WorldWatcher requires.
WorldWatcher has been used in a wide range of classrooms from middle school
through college.' Although little formal evaluation has been conducted on those
uses, our observations and reports do not reveal issues of the magnitude of those
raised by their predecessors for motivation, accessibility, background knowledge,
management of inquiry, or meeting the practical constraints of classrooms. Where
prior visualizers raised issues that were clear obstacles for the use of all or most stu-
dents in the classes we observed or were in contact with, the concerns that are being
raised by the use of WorldWatcher seem to be less severe or to affect specific popu-
lations of students. Because of the progress made through the intensive design pro-
cess leading to WorldWatcher, identifying the next set of issues will require a more
careful evaluation process than was required by the previous environments. We
provide a brief overview of these evaluation efforts in the next steps described in
the final section of this article. Here, we present a brief summary of
Worldwatcher's status with respect to the specific challenges of inquiry-based
learning.
'It has been difficult to track the actual use of WorldWatcher in classrooms. Several thousand copies
of WorldWatcher have been distributed over the Internet and on CD-ROM, many of those directly to
teachers. However, we have no data on how many of those copies were actually used in classrooms. The
most conservative estimate from the schools participating in the CoVis Project and other schools in di-
rect contact with our research group is that at least 50 ofthose teachers have used WorldWatcher in their
classrooms.
Motivation. The newest Global Warming Curriculum has demonstrated

promise in providing a coherent motivating context in pilot tests in three schools.
However, we have observed difficulties with students losing sight of the overall
context as they become involved in individual activities. The problem appears
when they complete an activity and have lost track ofwhy they conducted the activ-
ity and where they are going next. Reestablishing that motivating context is an im-
portant challenge, and we are making revisions to the curriculum in response.
Accessibility of investigation techniques. Evidence from dozens of class-

rooms and several workshop-setting studies shows that despite a number of minor
usability problems with Worldwatcher's user-interface, students at a wide range of
age and ability levels are able to master the major visualization and data analysis
operations in WorldWatcher within their first 45-min session. We have observed
all of the operations in WorldWatcher appropriated successfully by middle and
high schools students for use in student-directed investigations.8 Nevertheless, we
do not see uniform appropriation of features across student populations, nor do stu-
dents always use them in a careful, scientific fashion. We have been especially
struck by the strength of the confirmation bias in student investigations. Under-
standing the factors that contribute to these observations will require more careful
attention to individual differences in future research.
Background knowledge. We have received favorable responses from

teachers to the knowledge resources in WorldWatcher and the activities designed
to give concrete experience with the quantities measured by the data. However, we
are still awaiting evidence of their value from direct observation of students.
Organizing and managing open-ended inquiry. Teachers also respond

favorably to the WorldWatcher notebook facility, and we have observed students
using it successfully to record their work when asked to in guided activities. We do
not yet have a significant body of experience with students engaged in open-ended
activities to learn if and how students use it spontaneously. Our early impression is
that the technological support in the form of notebooks is only part of the solution
and that helping students to recognize the value of record-keeping and develop ef-
fective record-keeping habits is going to require additional supports, such as pro-
viding structures for students to work within and prompts for metacognition (Linn
et al., in press; Scardamalia et al., 1989; Tabak et al., 1996).
sThis excludes histograms and scatterplots, which were added specifically for use at the undergradu-
ate level.
Practical constraints of the learning context. The biggest current practi-

cal challenge to Worldwatcher's compatibility with typical learning contexts is its
restriction to the Apple Macintosh. This issue is being addressed through further
software development. In addition, although the activities in the Global Warming
Curriculum have been carefully designed to accommodate the restrictions of
45-min class periods, teachers have raised concerns about the total amount of time
the entire curriculum takes and balancing it against other curricular needs. This
raises the issue of trade-offs between breadth and depth of coverage that are com-
mon in the implementation of inquiry-based learning.
LESSONS LEARNED: A DESIGN FRAMEWORK

FOR TSIL
The most important lesson of the experiences reported here is that the implementa-
tion of TSIL requires an integrated process of both technology and activity design.
In particular, these experiences show that to address the challenges of in-
quiry-based learning successfully, the TSIL design process must coordinate four
interdependent components: the identification of a motivational context, the selec-
tion and sequencing of activities, the design of investigation tools, and the creation
of process supports. By exposing a wide range of challenges to inquiry-based learn-
ing, the design effort described here provides the opportunity to explore strategies
for each of these four components. In this section, we use these four components as
a framework to organize our discussion of the strategies we identified and the ways
those strategies address the challenges of inquiry-based learning.
Identificationof a Motivational Context
To address the need for motivation, designers of inquiry-based learning must iden-
tify a motivating context for inquiry early in the design process. Our early experi-
ences with the Climate Visualizer demonstrated the risks of not establishing a
meaninghl motivating context. In addition, we found ourselves in a position where
we could not retrofit a motivating context to the existing design. Because the moti-
vational context has critical implications for the other elements of a TSIL design, it
must be identified early in the design process. In the process described here, for ex-
ample, the selection of global warming as a motivating context influenced the na-
ture and sequencing of the investigation activities and determined the contents of
the data libraries and the knowledge resources.
Our strategy for the selection of a motivating context resembles those of other
researchers in inquiry and project-based science learning in its focus on a mean-
ingful, controversial, and open scientific issue (e.g., Barron et al., 1998;
Blumenfeld et al., 1991). Through our conversations with students and teachers,
we identified four elements that give the investigation of global warming the inter-
est and value that are necessary to initiate engagement. First, the topic is familiar.
Unlike many of the topics that students study in traditional science cumcula,
global warming is an issue that they have heard about. Second, it has potential di-
rect implications for students. They are concerned that they may experience the ef-
fects of global warming. Third, the possibility of global warming and the measures
that have been suggested to reduce it have important social and policy implications
that draw on students' senses of fairness and entitlement. Fourth, the scientific is-
sues associated with global warming are a matter of legitimate uncertainty and
controversy in the scientific community. Students are intrigued by the fact that sci-
entists do not understand the science of climate completely and that they even dis-
agree. In comparison to the settled science that they are used to learning about in
school, students find the open controversy appealing. These four properties of
global warming apply to many issues in science and can serve as guidelines for the
design of inquiry learning in general. However, they do not by any means consti-
tute a complete list of features of a motivating context, and as Blumenfeld et al.
(1991) point out, there is a frustrating lack of empirical data describing the inter-
ests of learners that could be drawn on in the design of inquiry-based learning.
The selection of global warming as a motivating context had important implica-
tions for the other technology and curriculum design elements. For example, we
developed an increasingly comprehensive library of data sets that supported inves-
tigations of global warming and its potential causes and implications. We also de-
veloped and refined curricula designed to establish global warming as a
compelling issue for learners and then draw on their concern over global warming
to motivate investigations of the science of global climate and climate change.
Selection and Sequencing of Activities
The selection and sequencing of activities provides an opportunity to address the

challenges of motivation, accessibility, and background knowledge. In this pro-
cess, we used staging and bridging activities to address these three challenges.
Having selected global warming as the motivational context, we were able to
create sequences of staging activities that developed motivation by exposing stu-
dents to the potential implications of global warming through discussions, read-
ings, and videos. These activities were designed to simultaneously develop
interest and raise the key scientific questions that students would investigate over
the course of the unit. The sequencing of staging activities in this way can help to
establish long-term motivation by connecting students' concerns and interests to
specific scientific questions. Within this larger context, we also used the bridging
activity of map drawing to elicit curiosity about specific issues. By asking students
to articulate their prior conceptions, these types of activities can help to establish a
focused interest in specific phenomena.
We also addressed an important challenge of accessibility through the selection
and sequencing of activities. This was the challenge of helping students to under-
stand the opportunities offered by the available investigation techniques. Begin-
ning with the Radiation-Budget Visualizer, we addressed this challenge by
creating sequences of structured investigations that introduced the capabilities of
the investigation tools in manageable chunks and demonstrated their usefulness
for answering questions within the motivating context. By guiding students
through short, well-structured investigations, staging activities can provide stu-
dents with conceptual models of investigation techniques that the students can
draw on in subsequent open-ended investigations. The development of these con-
ceptual models helps to make the investigation techniques more accessible.
The design of staging and bridging activities also enabled us to address the
challenge of background knowledge. Each of the staging activities in the curricu-
lum sequences described here was designed to help students develop and refine
their understanding of specific content. Sequenced appropriately, staging activi-
ties can build on each other in successive steps to help students develop the knowl-
edge and skills that will enable them to engage in open-ended investigations.
Design of Investigation Tools
The design of a scientific investigation tool is the process of selecting specific oper-
ations and creating user-interfaces to them. In the case of the visualization environ-
ments described here, the design of the investigation tools also included the assem-
bly of data libraries. In this work we found that the design of investigation tools
could play an important role in addressing the challenges of motivation, accessibil-
ity, and practical constraints.
In the design process reported here, we addressed motivation through the selec-
tion of data sets and through the design of operations to support bridging activities.
The selection of data sets enabled us to draw on the motivating context of global
warming. The visualizers prior to WorldWatcher had difficulty initiating and
maintaining engagement because of the limitations of their data libraries. The data
library in the Climate Visualizer was unable to initiate engagement because its
limited data did not support investigations that students found interesting. In the
design of the subsequent Radiation-Budget and Greenhouse Effect Visualizers,
the selection of data sets related to Earth's energy balance enabled us to draw on
global warming as a motivating context. However, it wasn't until the development
of WorldWatcher, when we added data addressing human contributions to possi-
ble global warming and the potential implications of global climate change, that
we were able to effectively initiate and maintain engagement for large numbers of
students. The design of operations also enabled us to address the challenge of mo-
tivation, specifically through the creation of operations to supporting bridging ac-
tivities. As we described previously, the bridging activities in which students draw
visualizations to articulate their initial conceptions can help to create interest. To
support these bridging activities, we designed a number of operations, modeled on
the interfaces of graphics programs, to enable students to create new
Worldwatcher data sets by "painting" them.
From our first designs, we recognized the need to address accessibility through
the design of investigation tools. Following the Learner-Centered Design ap-
proach, we began by creating scaffolds in the tools. We identified tacit knowledge
that scientists draw on in the creation and interpretation of visualizations and em-
bedded that knowledge in the form of contextual information in the user-interface.
We maintained this approach throughout the design history. We provided addi-
tional scaffolds in the form of default settings that enabled learners to create inter-
pretable visualizations without requiring any foreknowledge of the data. Finally,
in we created operations and representations that were specifically designed to
make the properties of the tools clear to learners. These included the spread-
sheet-like display of numerical values in magnified visualizations (Figure 9) and
the ability to project rectangular visualizations into alternative projections, includ-
ing cut-and-fold images that can be printed and assembled into a polyhedral globe
(Figure 10). All of these strategies address the problem that learners lack the back-
ground knowledge that scientists bring to their use of scientific investigation tools.
The supports in the software made investigation techniques accessible to learners
by embedding that knowledge in the interface for learners to draw on, by bypass-
ing the need for that knowledge, or by helping the learners to develop the knowl-
edge themselves.
The two practical constraints of the learning environment that we continually
confronted were limits on available time and technology infrastructure in class-
rooms. Although these issues draw designers' attention away from the fundamen-
tal processes of teaching and learning, they are nevertheless critical concerns for
the implementation of any approach to learning. In this design process, we were re-
peatedly confronted by the need to improve the performance of the software and
match its requirements to the technology infrastructure in schools. The rapidly
changing nature of technology means that the specific strategies we employed in
this process have little general value. However, these experiences do demonstrate
the critical importance of both speed and reliability for classroom technologies.
Creation of Process Supports
Process supports are the elements of a TSIL design that address the needs of in-
quiry-based learning beyond the direct support for the investigation techniques. In
the case of the designs described here, the process supports that proved necessary
were knowledge resources and record-keeping tools. These supports address the
challenges of background knowledge and managing extended activities.
In this design history, we addressed the need for background knowledge in the
Greenhouse Effect Visualizer and in by creating information resources within the
software environments. These resources provided a mechanism for learners to ac-
cess background information as the need arose in the course of an investigation.
By providing these resources, we attempted to reduce the demand on the teacher
and existing classroom resources, thereby simultaneously addressing practical
constraints of the leaming environment. By placing these resources within the
software environment, we took advantage of the opportunity to link them directly
to data or activities, making it possible for students to move between investigation
activities and knowledge resources with minimal interruption.
We addressed the need to support students in organizing and managing inquiry
activities by providing them with tools to record the process and the intermediate
products of their investigations. Record-keeping tools provided students with the
opportunity to create and monitor plans, articulate hypotheses, analyze evidence
carefully, and reflect on their progress. Our experiences indicate that these re-
cord-keeping facilities were helpful but that they were not sufficient to resolve the
challenges of managing extended activities.
Putting the Elements Together
Each of the elements of the TSIL design process can address different facets of the
challenges of implementing inquiry-based learning. Assembled into a coherent
whole, they have the potential to successfully achieve all three of the categories of
learning objectives for inquiry-based leaming that we presented earlier-specific
inquiry skills, general inquiry abilities, and scientific content. By addressing the
challenge of motivation, an effective TSIL design can create sufficient engagement
to bring about learning. By making investigation techniques accessible, a design
can engage students in inquiry activities through which they achieve mastery of
specific inquiry skills and can develop new content understanding through discov-
ery and refinement. Once engaged in inquiry, students often require additional
knowledge to complete their investigations and interpret their results. In success-
fully addressing the challenge of background knowledge, a design can meet this
need by supporting the further development of content understanding. By address-
ing the challenge of supporting extended, open-ended activities, a design can pro-
vide the opportunity to fblfill the motivating context and allow learners to develop
general inquiry abilities. Finally, addressing the practical constraints of the learn-
ing context makes it possible for such a design to achieve these learning objectives
in realistic settings. We do not claim to have fully achieved these objectives
through the designs presented here. However, the strategies we employed in these
designs have been successful at reducing the obstacles to implementing in-
quiry-based learning, bringing the designs substantially closer to achieving those
objectives.
NEXT STEPS
WorldWatcher and its accompanying curricula are the products of an extensive it-
erative design process. The advances made in the design of WorldWatcher and the
Global Warming Curriculum provide the opportunityto pursue new research direc-
tions. In the next stage of this research, we are initiating a new phase of evaluation,
and we are broadening our design scope to include support for the role of the
teacher.
Evaluation
The issues for inquiry-based learning discussed here were identified through an it-
erative design and evaluationprocess. This formative evaluation has focused on the
identification and resolution of critical obstacles to the implementationof effective
inquiry-based learning with scientific visualization. However, the success of
WorldWatcher and its accompanying curricula in addressing many of the chal-
lenges raised by its predecessorsrepresents an important transition in this design re-
search process. Having reached this point, we are currently initiating a more formal
evaluation process designed to take advantage of WorldWatcher and its accompa-
nying curricula to explore the theories and assumptionsthat have guided this design
process. The goal ofthis next phase of evaluation will be to create a rich description
of the enactment of TSIL in classrooms, with attention to the variation across and
within classrooms, and to use these descriptions to help us understand the contribu-
tions ofthe design strategies discussed here to the implementationof TSIL in class-
rooms and to attitudinal and learning outcomes.
Design for Teachers
In the discussion of the design process in this article, we have focused primarily
on the needs of learners. However, over the course of this effort, we have be-
come increasingly aware of the challenges that TSIL poses for teachers. For ex-
ample, at many points in this design process, we were told by individual
teachers that they could not use our software or curriculum because they did not
feel comfortable enough with the technology, the content, or the process of man-
aging TSIL. Our experiences have convinced us that supporting the role of the
teacher raises important additional challenges for the design of both software
and activities beyond the issues for students that we have discussed here. We
have begun to explore these issues through a new program of design research fo-
cusing on support for the enactment of inquiry-based science learning by teach-
ers (Brown & Edelson, 1998). In this new research, we are investigating the
issues that confront teachers in the enactment of curricula like the Global
Warming Cumculum and how to prepare materials for teachers that will support
them in planning, customizing, conducting, and assessing them within the prac-
tical constraints of their local environment.
ACKNOWLEDGMENTS
This work has been supported in part by the National Science Foundation under the
Grants MDR-9253462, RED-9453715, RED-9454729, ESI-9720687, and
REC-9720383. Additional support was provided by Apple Computer, Sun
Microsystems, and Spyglass, Inc.
We would like to acknowledge the crucial contributions of all of the re-
searchers and participating teachers and students in the and SSciVEE Projects
to the designs described here, as well as to the development and refinement of
the ideas in this article. We are also grateful to our reviewers and to Janet
Kolodner for helpful comments.
Important contributions to the development of the software and curriculum dis-
cussed in this article were made by Joseph Polman, Kevin O'Neill, Barry Fishman,
Laura D'Amico, Steven McGee, Joey Gray, Paul Forward, Patricia Carlson,
Eileen Lento, Matthew Brown, Duane Griffin, and Louis Gomez. Worldwatcher
was written by Brian Clark. Ray Pierrehumbert, David Hollander, and Brad
Sageman served as consultants on scientific content and practice.
REFERENCES
American Association for the Advancement of Science. (1994). Benchmarksfor science literacy. New
York: Oxford University Press.
Asbell-Clarke, J., & Barclay, T. (1996). Discovering the scientist within: Hands-On Universe. In
Hands-On! (Vol. 19, pp. 4 5 , 17). Cambridge, MA: TERC.
Barron, B. J. S., Schwartz, D. L., Vye, N. J., Petrosino, A., Zech, L., Bransford, J. D., &The Cognition
and Technology Group at Vanderbilt. (1998). Doing with understanding: Lessons from research on
problem- and project-based learning. The Journal of the Learning Sciences. 7, 271-3 1 1 .
Barstow, D., & Berenfeld, B. (1996, April). Data visualization as an essential component of
telecollaborative inquiry-based science learning. Paper presented at the annual meeting of the
American Education Research Association, New York.
Berlyne, D. E. (1966). Curiosity and exploration. Science. 153, 25-33.
Blumenfeld, P. C., Soloway, E., M a n , R., Krajcik, J. S., Guzdial, M., & Palincsar, A. (1991). Moti-
vating project-based learning: Sustaining the doing, supporting the learning. Educational Psycholo-
gist, 26, 369-398.
Brown, M., & Edelson, D. C. (1998). Software in context: Designing for students, teachers, and class-
room enactment. In A. S. Bruckman, M. Guzdial, J. L. Kolodner, & A. Ram (Eds.), Proceedings of
ICLS 98: International Confevence on the Learning Sciences. Atlanta, GA (pp. 6 3 4 9 ) . Charlottes-
ville, VA: AACE.
Bnmer, J. S. (1961). The act of discovery. Harvard Education Review. 31, 21-32.
D'Amico, L., Fishman, B., Gordin, D. N., McGee, S., O'Neill, D. K., & Polman, J. (1994). The atmo-
spheric sciences: Six views into a communiq ofpractice. Unpublished manuscript, Northwestern
University.
de Jong, T., & van Joolingen, W. R. (1998). Scientific discovery learning with computer simulations of
conceptual domains. Review ofEducationa1 Research, 68. 179-20 1.
Dewey, J. (1 964a). Progressive organization of subject matter. In R. D. Archambault (Ed.), John Dewey
on education: Selected writings (pp. 373-387). Chicago: University of Chicago Press.
Dewey, J. (1964b). Science as subject matter andasmethod. In R. D. Archambault (Ed.), John Deweyon
education: Selected wrirings (pp. 182-1 95). Chicago: University of Chicago Press.
Edelson, D. C. (1998a). Matching the design of activities to the affordances of software to support in-
quiry-based learning. In A. S. Bruckman, M. Guzdial, J. L. Kolodner, & A. Ram (Eds.), Proceedings
oflCLS 98: International Conference on the Learning Sciences, Atlanta, GA. December 16-19 (pp.
77-83). Charlottesville. VA: AACE.
Edelson, D. C. (1998b). Realising authentic science learning through the adaptation ofscience practice.
In B. J. Fraser & K. Tobin (Eds.), International Handbook ofscience Education (pp. 3 17-33 1).
Dordrecht, The Netherlands: Kluwer.
Edelson, D. C., Pea, R. D., & Gomez, L. M. (1996a). The collaboratory notebook: Support for collabora-
tive inquiry. Communications of the ACM, 39. 32-33.
Edelson, D. C., Pea, R. D., & Gomez, L. (1996b). Constructivism in the collaboratory. In B. Wilson
(Ed.), Constmctivisr learning environments. Case studies in instructional design. Englewood
Cliffs, NJ: Educational Technology Publications.
Eisenberg, M., & Nishioka, A. (1997). Creating polyhedral models by computer. JournalofCumputers
in Mathematics and Science Teaching, 16, 477-5 1I.
Ferguson, W., Bareiss, R., Bimbaum, L., & Osgood, R. (1992). ASK systems: An approach to the real-
ization of story-based teachers. The Journal of the Learning Sciences, 2, 95-1 34.
Fishman, B., & D'Am~co,L. (1994). Which way will the wlnd blow? Networked tools for study-
ing the weather. In T. Ottmann & I. Tomek (Eds.), Educational Multimedia and
Hypermedia. 1994 (pp. 209-216). Vancouver, BC: Association for the Advancement of
Computing in Education.
Frodeman, R. (1995). Geological reasoning: Geology as interpretive and historical science. GSA Bulle-
tin. 107, 960-968.
Gomez. L. M., Gordin, D. N., & Carlson, P. (1995). A case study of open-ended scientific inquiry in a
technology supported classroom. In .I. Greer (Ed.), Proceedings ofAI-Ed '95, Seventh World Con-
ference on Art~ficiallntelligencein Education (pp. 17-24). Charlottesville, VA: Association for the
Advancement of Computing in Education.
Gordin, D. N., Edelson, D. C., & Pea, R. D. (1995, January). The Greenhouse Effect Visualizer: A tool
for the science classroom. Preprints of the Fourth American Meteorological Society Education
Symposium, Dallas. Texas (547-556).
Gordin, D. N., & Pea, R. D. (1995). Prospects for scientific visualization as an educational technology.
The Journal of the Learning Sciences. 4. 249-279.
Gordin, D. N., Polman, J. L., & Pea, R. D. (1994). The Climate Visualizer: Sense-making through scien-
tific visualization. Journa[ of Science Education and Technology, 3, 203-226.
Greenberg, R.,Kolvoord, R. A., Magisos, M., Strom, R. G., & Croft, S. (1993). Image processing for
teaching. Journal ofscience Education and Technology, 2, 469480.
Greeno, J., Collins, A., & Resnick, L. B. (1996). Cognition and learning. In R. Calfee & D. Berliner
(Eds.), Handbook of educationalpsychology. New York: Macmillan.
Guzdial, M., Turns, J., Rappin, N., & Carlson, D. (1995). Collaborative support for learning in complex
domains. In J. L. Schnase & E. L. Cunnius (Eds.), Proceedings of CSCL 95: Computer Supportfor
Collaborative Learning (pp. 157-160). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
Hancock, C., Kaput, J. J., & Goldsmith, L. T. (1992). Authentic inquiry with data: Critical baniers to
classroom implementation. Educational Psychologist, 27, 337-364.
Hawkins, J., & Pea, R. D. (1987). Tools for bridging the cultures of everyday and scientific thinking.
Journal of Research in Science Teaching, 24, 291-307.
Hieben, J., Carpenter, T. P., Fennema, E., Fuson, K., Human, P., Murray, H., Olivier, A,, & Wearne, D.
(1996). Problem solving as a basis for reform in cuniculum and instruction: The case of mathemat-
ics. Educational Researcher, 25(4), 12-2 1.
Hoadley, C. M., Hsi, S., & Berman, B. P. (1995). The Multimedia Forum Kiosk and SpeakEasy. Pro-
ceedings ofACM Multimedia '95 (pp. 363-364). New York: ACM Press.
Hunt, E., & Minstrell, J. (1994). A cognitive approach to the teaching of physics. In K. McGilly (Ed.),
Classroom lessons: Integrating cognitive theory and classroom practice (pp. 5 1-74). Cambridge,
MA: MIT Press.
Ireton, M. F. W., Manduca, C. A., & Mogck, D. W. (Eds.). (1996). Shaping thefuture of undergraduate
earth science education: Innovation and change using an earth system approach. Washington, DC:
National Science Foundation.
Jackson, S. L., Stratford, S. J., Krajcik, J.. & Soloway, E. (1996). A learner-centered tool for students
building models. Communication ofthe ACM, 39(4), 4 8 4 9 .
Kaufmann, W., & Smart, L. L. (1993). Supercomputing and the transformation ofscience. New York:
Freeman.
Krajcik, J.,Blumenfeld, P. C., Marx,R. W., Bass, K. M.,Fredericks, J.,& Soloway, E. (1998). Inquiry in
project-based science classrooms: Initial attempts by middle school students. The Journal ofthe
Learning Sciences. 7, 3 13-350.
Linn, M. C., Bell, P., & Hsl, S. (1998). Lifelong science learning on the Internet: The knowledge integra-
tion environment. Interactzve Learning Environments, 6(1-2), 4-38.
Linn, M. C., disessa, A,, Pea, R. D., & Songer, N. B. (1994). Can research on science learning and in-
struction inform standards for science education? Journal of Science Education and Technolog):
3(1), 7-15.
Linn, M. C., Songer, N. B., & Eylon, B. S. (1996). Shifts and convergences in science learning and in-
struction. In R. Calfee and D. Berliner (Eds.), Handbook ofeducationalpsycho[ogy. New York:
Macmillan.
Loh, B., Radinsky, J., Reiser, B. J., Gomez, L. M., Edelson, D. C., &Russell, E. (1997, December). The
progress portfolio: Promoting reflective inquiry in complex investigation environments. Proceed-
ings of CSCL 97: Computer Support for Collaborative Learning, Toronto, Canada (pp. 169-1 78).
Mandinach, E. B., & Cline, H. F. (1996). Classroom dynamics: The impact of a technology-based cur-
riculum innovation on teaching and learning. Journal of Educational Computing Research, 14(1),
83-1 02.
McConnick, B. H., DeFanti, T. A,, & Brown, M. D. (Eds.). (1987, November). Special issue on
visualization in scientific computing. Computer Graphics, 21(6).
National Aeronautics and Space Administration. (1998). Understanding our changingplanet: Earth
Science Enterprise 1998 applicationsfact book. Washington, DC: Author.
National Aeronautics and Space Administration Advisory Council. (1988). Earth system science: A
closer view (Report ofthe Earth System Sciences Committee, NASA Advisory Council ). Washing-
ton, DC: National Aeronautics and Space Administration.
National Research Council. (1996). National science education standards. Washington, DC: National
Academy Press.
O'Neill, D. K., & Gomez, L. (1994). Thecollaboratory notebook: A distributed knowledge-building en-
vironment for project-enhanced learning. In T. Onmann & I. Tomek (Eds.), Educational Multime-
dia and Hypermedia. 1994: Proceedings ofEd-Media 94 (pp. 4 16-423). Charlottesville, VA: Asso-
ciation for the Advancement of Computers in Education.
Pea, R. D. (1993, May). Distributed multimedia learning environments: The collaborative visualization
project. Commztnication of the ACM, 36(5), 60-63.
Penner, D. E., Giles, N. D., Lehrer, R., & Schauble, L. (1997). Building functional models: Designing an
elbow. Journal of Research in Science Teaching, 34, 125-143.
Plutchak, J., Hall, S., Wojtowicz, D., Sridhar, M., Ramamurthy, M., & Wilhelmson, R. (1998). Weather
World 2010: A customizable, user-oriented WWW site. Proceedings of the 14th International Con-
ference on Interactive Information and Processing Systems (IIPS)for Meteorology, Oceanography,
and Hvdrology, Phoenix, Arizona (pp. 402403).
Resnick, M. (1994). Tztrtles, termites, and trafficjams:Explorations in massivelyparallel microworlds.
Cambridge, MA: MIT Press.
Rock, B. N., Blackwell, T. R., Miller, D., & Hardison, A. (1997). The GLOBE program: A model for in-
ternational environmental education. In K. C. Cohen (Ed.), Internet linksfor science education: Stu-
dent-Scientist partnerships. New York: Plenum.
Samson, P. J., Steremberg, A,, Ferguson, J., Kamprath, M., Masters, J., Monan, M., & Mullen, T. (1994,
January). Blue-Skies: A new interactive teaching tool for K-12 education. Proceedings ofthe Third
American Meteorological Society Education Symposium (pp. 59-514).
Scardamalia, M., & Bereiter, C. (1994). Computer support for knowledge-building communities. The
Journal of the Learning Sciences, 3, 265-283.
Scardamalia, M., Bereiter, C., McLean, R. S., Swallow, J., & Woodruff, E. (1989). Computer-supported
intentional learning environments. Journal of Educational Computing Research, 5(1), 5 1-68.
Schank, R. C. (1999). Dynamic memoly revisited. New York: Cambridge University Press.
Schank, R. C., Fano, A,, Bell, B., & Jona, M. (199311994). The design of goal-based scenarios. The
Journal of the Learning Sciences, 3, 305-346.
Schauble, L., Glaser, R., Duschl, R. A., Schulze, S., & John, J. (1995). Students' understanding of the
objectives and procedures of experimentation in the science classroom. The Journal ofthe Learning
Sciences. 4, 13 1-1 66.
Soloway, E., Guzdial, M., & Hay, K. E. (1994). Learner-centered design: The challenge for HCI in the
2 l st century. Interactions, 1, 36-47.
Songer, N . B. (1995). Knowledge construction through exchange and dialogue: A case study of Kids as
Global Scientists. The Journal ofthe Learning Sciences, 5(4), 297-328.
Tabak, I., Smith, B. K., Sandoval, W. A,, & Reiser, B. J. (1996, June). Combining general and do-
main-specific strategic support for biological inquiry. In C. Frasson, G. Gauthier, & A. Lesgold
(Eds.). Proceedings of the Thirdlnternationul Conference on Intelligent TutoringSystems (ITS '96),
Montreal (pp. 288-296). New York: Springer-Verlag.
Thomburg, D. G., & Pea, R. D. (1991). Synthesizing instructional technologies and educational culture:
Exploring cognition and metacognition in the social studies. Journal ofEducationa1 Computing Re-
search, 7(2), 12 1-164.
Wallace, R., Bos, N., Hoffman, J., Hunter, H. E., Kiskis, D., Klann, E., Krajcik, J., Peters, G., Richard-
son, D., Ronen, O., & Soloway, E. (1998). ARTEMIS: Learner-centered design of an information
seeking environment for K12 education.
Wallace, R., Soloway, E., Krajcik, J., Bos, N., Hoffman, J., Hunter, H., Kiskis, D., Klann, E., Peters, G.,
Richardson, D., & Ronen, 0. (1998, April). ARTEMIS: Learner-centered design of an information
seeking environment for K-12 education. Conference Proceedings ofCH198: Human Factors in
Computing Systems, Los Angeles. California (pp. 195-202). New York: ACM.
Welch, W. W., Klopfer, L. E., Aikenhead, G. S., &Robinson, J. T. (198 1). The role of inquiry in science
education: Analysis and recommendations. Science Education. 65, 33-50.
White, B. Y. (1993). ThinkerTools: Causal models, conceptual change, and science education. Cogni-
tion and Instruction, 10. 1-100.
Wilensky, U., & Resnick, M. (1999). Thinking in levels: A dynamic systems approach to making sense
of the world. Journal ofscience Education and Technologv, 8(1), 3-19.

A101 Edelson Etal 99 PDF

Uploaded by

Copyright:

Available Formats

A101 Edelson Etal 99 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

A101 Edelson Etal 99 PDF

Uploaded by

Copyright:

Available Formats

Addressing the Challenges of Inquiry-Based Learning

Through Technology and Curriculum Design

To cite this version:

HAL Id: hal-00190609

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

Addressing the Challenges of

Douglas N. Gordin and Roy D. Pea

THE POTENTIAL OF TSlL

Opportunities for Learning

Scientists' knowledge of scientific concepts, understanding of scientific tools and

Technological Support for Inquiry-Based Learning

Computing and networking technologies offer dramatic, new opportunities to sup-

Enhancing interest and motivation.

An Example: Climatology and Scientific Visualization

Climatology. The science of climate is an important area for study because,

Scientific visualization. The term scientzfic visualization covers a class of

findings, and to convince others. As a glance at virtually any journal in the

Learning objectives for climatology and inquiry through visualization. The

THE CHALLENGES OF IMPLEMENTING

1. Motivation. For students to engage in inquiry in a way that can contribute to

sity of abilities and prior experiences is another challenge of Learner-Centered

Meaningfiulproblems. A meaningful problem is one that has implications that

INTRODUCTION TO THE DESIGN HISTORY

THE CLIMATE VISUALIZER

Software Architecture, Features, and Interface

In contrast to the general-purpose visualization environments used by scientists,

Issues Addressed by the Design

Visualizer provided learners with access to a number of powerful operations for

Issues Raised by Classroom Experiences

evidence for a hypothesis. For example, a novice looking at the visualization of

Motivation. A related concern raised in these classroom experiences was that

The problem of maintaining engagement with the Climate Visualizer showed

In contrast to the Climate Visualizer's regional data, the Radiation-Budget Visual-

Software Architecture, Features, and Interface

Like the Climate Visualizer, the Radiation-Budget Visualizer was implemented as

The Radiation-Budget Visualizer was designed to support activities in which learn-

The Radiation-Budget Visualizer was released to participants in the CoVis Project

Issues Addressed by the Design

The design of the Radiation-Budget Visualizer and its accompanying cumculum

The greenhouse investigation activity was intended to address challenges to

Issues Raised by Classroom Experiences

THE GREENHOUSE EFFECT VISUALIZER

pact is characterized by a change in an annual, global average, whereas its impacts

Software Architecture, Features, and Interface

FIGURE 4 The interfacetothe data library in the web-based GreenhouseEffectVisualizer.

FIGURE 5 Data displayed in the web-based Greenhouse Effect Visualizer

ments needed to access the visualizations, although it added the requirement of an

Issues Addressed by the Design

Issues Raised by Classroom Experiences

Although the Greenhouse Effect Visualizer represented a breakthrough in the num-

Motivation. The Global Warming Conference curriculum was intended to

Accessibility. An important issue for the accessibility of investigation tech-

Background knowledge. Although we found that the explanatory materials

Practical constraints. Finally, in trying to meet the constraints of the class-

Worldwatcher, the most recent visualization environment in this line, continues

WorldWatcher is the first of the visualization environments to be developed by a

Software Architecture, Features, and Interface

With WorldWatcher we returned to a standalone, nonnetworked application to pro-

WorldWatcher. These are the ability to examine visualizations at varying spa-