Using virtual reality computer models to support student understanding of astronomical concepts.

        The purpose of this study was to examine how 3-dimensional(3-D)
        models of the Solar System supported student development of
        conceptual understandings of various astronomical phenomena that
        required a change in frame of reference. In the course described
        in this study, students worked in teams to design and construct
        3-D virtual reality computer models of the Solar System. Using a
        qualitative and methodological approach, we found that computer
        modeling supported students in learning about astronomy concepts
        that required a change in frame of reference or perspective.
        These findings are particularly important for astronomy
        education because many concepts in astronomy require changing
        one's perspective within a 3-D space.

INTRODUCTION

Many learning situations require students to mentally transform 2-D objects into dynamic 3-D objects within some particular process or state of being (Dixon, 1997). For example, conceptualizing scientific processes and phenomena in three dimensions is essential if one is to understand the scientific concepts of Earth's seasons, or various genetic and cellular processes (Gotwals, 1995; Windschitl, Winn, & Headley, 2001). In addition, to understand many science concepts, learners may need to translate among reference frames, to describe the dynamics of a model over time, to predict how changes in one factor influence other factors, or to reason qualitatively about physics processes that are best explored in 3-D space (Dede, 2000). These findings have encouraged educators to examine the use of 3-D technologies as a means to support students in constructing and visualizing scientific phenomena. For example, the CoVis project developed educational activities in which students analyzed and interpreted complex visual representations in atmospheric science (Pea, 1993). Other projects use computational modeling software that allows students to construct and view scientific models that can be used to visualize weather cells (Hay, Marlino, & Holschuh, 2000), gorilla behavior (Hay, Crozier, and Barnett, 2000), the dynamic and ecological processes of the Puget Sound (Windschitl & Winn, 2000), the Solar System (Yair, Mintz, & Litvak, 2001), and Newtonian physics phenomena from varying perspectives (Salzman, Dede, & Bowen-Loftin, 1995). In general, these studies have reported that 3-D computer modeling technologies can support students in understanding spatial relationships between objects and the relationships between variables that constitute a particular phenomenon.

At the college level, many students enroll in introductory astronomy courses because of their appeal and students' perceptions that astronomy is less mathematical than the alternative courses of physics or chemistry. However, to understand basic astronomical phenomena such as day and night, and the reasons for the seasons and lunar and solar eclipses, students must visualize relationships between objects (e.g., the Earth, Moon, and Sun) and events (.g., eclipses) from different 3-D perspectives (Yair et al., 2001). Hence, it is not surprising that students have great difficulty in developing conceptual understandings of astronomical phenomena (Barnett & Morran, 2002; Sneider & Ohadi, 1998; Stahly, Krockover, & Shepardson, 1999). In fact, over 100 research studies have been conducted that report on student difficulty in developing understandings of astronomical phenomena and show that students typically have poor or contrary explanations of such scientific phenomena--explanations that are in conflict with those currently accepted by the scientific community (Pfundt & Duit, 2004; Wandersee, Mintzes, & Novak, 1994). This difficulty in developing an understanding of astronomical concepts arises, in part, because the science of astronomy requires students to develop an understanding of the complex relationships and dynamics between objects in 3-D space as well as to examine objects and events from different perspectives (Parker & Heywood, 1998a).

During the past 2 years, we have been researching and developing an introductory undergraduate astronomy course in which students construct computer models of the Solar System using 3-D modeling software. In this paper, we focus our analyses on investigating the following research questions: (1) Does the use of computational models impact student understanding of astronomy concepts? and (2) Which astronomy concepts are best taught using 3-D computational models?

BACKGROUND

Use of 3-D Virtual Reality Computer Modeling to Support Learning

The 3-D modeling tools and manipulatives historically available to educators (e.g., orreries found in many science classrooms) may not engage students in the kind of concept-building activities necessary to promote the development of conceptual understandings (Dede, 2000). For example, most astronomy resources available to students are in the form of 2-D charts and images in textbooks that attempt to emulate 3-D scientific phenomena (Keating, Barnett, Barab, & Hay, 2002). Students viewing these images may lack the sense of depth or scale of these phenomena that is needed to understand the dynamics of the concept (Dede, 2000).

Recent research in examining 3-D computer technologies to support science learning has found an encouraging array of positive learning outcomes in a range of projects and domains. Findings include better symbolic retention of human cell organelle information (Gay, 1994), increases in spatial understanding of architectural spaces (Youngblutt, 1998), significantly higher scores on performance and achievement tests (Cobb, Crosier, Wilson, & Wilson, 2002), more accurate and complete understandings of engineering concepts (Bell & Fogler, 1995), increases in learning speed-of-life skills (Moshell, Michael, & Hughes, 2002), long-term retention of the atom (Byrne, 1996), increased ability to identify and draw perspectives of a pyramid (Ainge, 1996), and an increase in the ability of low achievers to draw mental models of ecology concepts (Osberg, Winn, Rose, Hoffman, & Char, 1997).

Even though these findings have encouraged us in the continued exploration of 3-D computer modeling technologies to promote learning, a word of caution is necessary. Most of these findings are derived from informal studies and there is little rigor in the methods used. There are notable exceptions where researchers used experimental methods to compare 3-D computer technology-enhanced instruction with non 3-D technology-enhanced instruction. Most relevant here is work by Dede, Salzman, Loftin, and Sprague (1999) who conducted experiments using MaxwellWorld and EM Field, a similar 2-D software. These experiments found that the MaxwellWorld group developed more accurate and causal mental models than the EM group. Specifically, Dede and colleagues found that the MaxwellWorld group was able to understand space as a whole, recognize symmetries in the field, and relate individual visual representations (e.g., test charge traces, field lines, and equipotential surfaces) to the electric field and electric potential.

Research findings also generally have shown that learners both enjoy their virtual reality (VR) educational experiences and see the potential of VR in instruction. There are notable exceptions; one in particular is the Virtual Reality Rover project at the Human Performance and Interactive Technology Laboratory (HITL), which found that enjoyment and sense of presence decrease with age (Winn & Jackson, 1999). Researchers have also found that when learners create a model in VR, they are creating a transitional object (Roschelle, 1992; Roth, 1995, 1996), an object that lies somewhere between a concrete object and a symbolic abstraction. However, where Roth and Roschelle were referring to 2-D animate circles in educational physics tools, in 3-D these transitional objects have a photo-realistic appearance. Near photo-realistic virtual models are clearly identifiable from the real objects. Yet to the learners, they still maintain their mathematical properties. Why? Because the learners have built them and have justified them to their teammates and instructors. Such models facilitate reflection on their limitations (Stratford, Krajcik, & Soloway, 1998). This is dramatically different in a pre-developed virtual world where, although the underlying mathematics are present, they can be easily overlooked or misunderstood by the learners. The advantage for learners is that they are "doing math" on what appears to be "real" objects (Wilson, Foreman, & Tlauka, 1997).

Astronomy Learning

During the past decade, a number of research studies have reported on the difficulty that students have in developing understandings of astronomical phenomena (Sneider & Ohadi, 1998; Stahly et al., 1999). In general, these studies have reported that students typically have poor understandings of scientific phenomena--understandings that are in conflict with the explanations currently accepted by the scientific community. These understandings have been referred to by many terms in the literature, including misconceptions, pre-conceptions, alternative conceptions, and alternative frameworks (Wandersee et al., 1994). In this paper, we refer to these understandings as alternative frameworks because to develop an understanding of many astronomy concepts requires an understanding of the relationships between multiple objects that are frequently embedded within a larger conceptual structure (Smith, diSessa, & Roschelle, 1992).

The fact that students typically hold alternative frameworks regarding astronomical concepts has been well documented in the literature (Pfundt & Duit, 2004). For example, over the course of three semesters, Comins (1993) identified 553 separate alternative frameworks in his introductory undergraduate astronomy courses. In another example, in the film, A Private Universe (1988), only 2 of 23 recent Harvard graduates and alumni selected at random were able to provide a satisfactory scientific explanation for the causes of the Earth's seasons. Similarly, a study conducted by Atwood and Atwood (1996) found that 39 of 42 pre-service elementary teachers held alternative frameworks in regard to the causes of the Earth's seasons.

This difficulty in developing an understanding of astronomical concepts arises, in part, because the science of astronomy requires students to develop an understanding of the complex relationships and dynamics between objects in 3-D space, as well as to examine objects and events from different perspectives (Parker & Heywood, 1998b). However, despite the 3-D nature of astronomy, most resources available to students are in the form of 2-D charts and images in textbooks, which attempt to emulate astronomical phenomena from different 3-D perspectives. Additionally, students have only one perspective from which to develop their understanding of astronomy concepts--namely from the Earth's perspective. As a result, developing learning activities that afford students opportunities to examine astronomical phenomena from different perspectives has traditionally been difficult because students simply cannot visit the Moon and look back at the Earth to observe the effects of the change in their perspective (i.e., does the Earth have phases when viewed from the Moon?). Yet to date, there have been only limited studies that examine how technology-based tools can impact students' abilities to shift their frame of reference or perspective and how that skill improves their conceptual understanding of science (Dede et al., 1999; Windschitel, Winn, & Headley, 2001). In fact, most research studies have focused on examining how students interpret and navigate spatial information depending on their particular frame of reference rather than examining conceptual learning outcomes (Furness, Winn, & Yu, 1997; McLellan, 1996; Thomas & Wickens, 2001; Windschitel et al., 2001). To address this gap in the research, we set out to develop a course in which students could engage in the process of designing models using 3-D modeling tools to develop their understanding of the complex dynamics associated with the Solar System.

ENABLING TECHNOLOGY

The creation of 3-D computational models has traditionally required advanced computer hardware and advanced programming skills. However, recent advances in 3-D modeling wysiwig (what you see is what you get) editors, coupled with the declining cost and increasing power of personal computers, has opened up opportunities for students to build complex 3-D models. Students constructed their models using Virtual Reality Markup Language (VRML). VRML is similar to HTML in that it is a language used for viewing virtual reality worlds on the World Wide Web (WWW). Additionally, VRML is platform-independent and is easily viewed over the Web using a free plug-in and a Web browser.

Rather than programming their models by hand, students use a wysiwig VRML editor. This reduces the tedious coding of VRML to a few mouse clicks. Instead of typing in abstract commands to create an earthlike object, a student can simply drag a sphere from the object toolbox into the workspace and resize, reorient, change its lighting, and texture the sphere to look like the Earth--and do it all within a short period of time using a few mouse clicks (see Figure 1). This procedure takes only a few seconds, freeing the student to concentrate on learning astronomy instead of struggling to learn the syntax and structure of programming. In fact, in our experience in working with our students, we have found that students who have very little computer experience easily learn how to manipulate the software to construct their models.

One of the strengths of VRML is that it allows for the use of viewpoints (see Figure 2). Viewpoints refer to perspectives or "camera positions" that can be placed within a VRML model, allowing viewers of the model to immediately shift to various locations and examine the model from the new perspective. This functionality provided by VRML opens many learning opportunities that are not normally available to many beginning astronomy students. For example, if a student wishes to view the Earth from the Moon, he or she would traditionally be relegated to viewing static pictures that show the Earth from the Moon. However, by using viewpoints in their 3-D models, a student can place a viewpoint on the Moon, observe the Earth, and determine whether the Earth has phases, as well as visualize other astronomical concepts that can be better understood by gaining a different perspective (i.e., does the Sun rise on the Moon?).

[FIGURE 1 OMITTED]

THE COURSE CONTEXT

The course is a learner-centered, project-based (Blumenfeld et al., 1991), undergraduate astronomy course in which students work in dyads and triads to build models of different aspects of the Solar System. The course that we are reporting on in this paper occurred during an 8-week summer term that met for 2 hours a day, 4 days a week. Given the pilot status of this course, only 15 students were enrolled. The curriculum for the course is comprised of three modeling projects of increasing sophistication designed to engage students in modeling various astronomical phenomena that are typically covered in a traditional lecture-based class. For the purposes of this study, we focus on course activities surrounding the second project, but to provide a context for the second project, we also describe the other two projects below.

[FIGURE 2 OMITTED]

In the first project, the students design and build a 3-D static model of the celestial sphere. The celestial sphere is a useful concept, first envisioned by ancient astronomers, to represent the location of the visible stars and important positions of the Sun throughout the year (i.e., solstices and equinoxes). The primary student goals of this project are to construct a geocentric model of the Earth-Sun system through which the students become familiar with some essential astronomical terminology (e.g., right ascension, declination, ecliptic), learn the causes for the seasons, and begin to build a conceptual base that they will use to understand future astronomy concepts. Due to the static nature of this project, it serves as a useful mechanism to increase the students' comfort level with the modeling software and provides time for students to become familiar with typical astronomical terminology and jargon.

The second project involves the construction of a 3-D dynamic model of the Earth-Moon-Sun system. The expectations of this project require students to investigate the relationships between the orbital paths and periods, orbital inclinations, and the distances between these bodies and their rotational rates of the bodies. This project extends the conceptual richness of the first project because students concern themselves with the scale of the system, orbital motions of the three bodies, and conditions for lunar and solar eclipses. The students are also asked to compare their model with the real Earth-Moon-Sun system and report on any discrepancies (e.g., scale, orbital speeds) between the two.

In the third project, students construct a 3-D dynamic model of the entire Solar System. They are expected to build a model of the Solar System that takes into account the rotational and revolution rates of the planets, and the relative size and distance between the planets. In constructing their models, students have to grapple with the difficult concept of the vast scale of the Solar System. Lastly, they are expected to investigate the similarities and differences between the planets' orbital motions, spins, interiors, moon systems, and atmospheres.

At the conclusion of each project, the students are given a range of opportunities to reflect on the model-construction process and the astronomy concepts demonstrated through and embodied in their computer models. First, students present their model to the entire class and discuss the astronomy concepts that are represented in their model. Second, each team creates a joint paper describing the astronomy concepts that the team modeled, and how the simplifications made affect the model's behavior (e.g., using circular instead of elliptical orbits).

Ensuring Scientifically Accurate Models

The software that was used in the course did not provide students with contextual feedback regarding the scientific accuracy of their model. Therefore, to support the students in critically examining the scientific accuracy of their models, we used two instructional strategies. First, each student wrote a report that documented how his or her model was similar and dissimilar to the actual Solar System in terms of behavior. In these reports, students were asked to explain their design decisions in creating the model and how those decisions impacted the dynamics of their model, particularly as compared to the real Solar System. In addition, at the end of each project each group of students presented its model to the class with the explicit purpose of explaining how the model was similar and dissimilar to the actual Solar System in terms of celestial dynamics. During this presentation, students were expected to describe how their model simulated various astronomical phenomena (such as phases or eclipses of the moon) or why their model did not adequately simulate particular phenomena. This presentation provided a chance for the students to articulate their model design and the limitations of their model. This last point is important because most students, when constructing models, do not realize that a model is a simplified and often idealized version of the actual system under study (Penner, Lehrer, & Schauble, 1998).

Students and Teams

During the first week of the course, the students divided themselves into teams. The teams were based upon proximity--students seated closest to each other formed a team. For the purposes of this paper, we focus on two of the teams that we refer to as Team Green and Team Yellow. We are focusing on these two teams of students because their responses to our interview questions and the challenges that they experienced in learning astronomy were representative of the class as a whole.

Team Green

Team Green consisted of three males: Taro, Todd, and Roger. They entered the course with large disparities in their computer experience and science background. Roger, a computer science major, entered the course with good computer skills, but minimal astronomy knowledge. Roger reported that he tends to dislike science courses and had tried to take as little science as possible throughout his academic career. He enrolled in the class only to satisfy a science requirement. Todd, a history major, entered the course with fair computer skills, but very little astronomy knowledge. Unlike Roger, Todd had a genuine interest in learning about astronomy. Todd also enrolled in the course to satisfy a science requirement. Taro was a unique member of the class due to his status as an upperclassman and his major, which was physics. He entered the course with considerable astronomy knowledge and good computer skills. Taro enrolled in the course for the opportunity to learn more about the VR software in an astronomy context.

Team Yellow

Team Yellow consisted of three students: Steve, Jessica, and Erica. All three students were telecommunications majors. Steve began the course with modest astronomy knowledge and fair computer skills. Steve reported that the primary reason he enrolled in the course was to learn about the VR software, although he also had a strong interest in learning more about astronomy. Jessica began the course with minimal astronomy knowledge and very little computer knowledge in comparison to Steve. Both Steve and Jessica enrolled in the course to satisfy a science requirement. In contrast, Erica had an interest in science and found the course description to be more interesting that the other courses that she could have taken.

METHODOLOGY

In this study, we examined the conceptual understanding of eight college students enrolled in a 1-semester, 3-D computer-modeling, enhanced astronomy course. Seven of the eight students were typical introductory astronomy students. That is, they varied in technological and scientific expertise and interest level in astronomy and science in general. Six were non-science majors. In this paper, we focus on the six non-science majors because non-science majors usually constitute the majority of the students in most undergraduate astronomy courses.

Data Collection

All of the students were interviewed twice, once at the beginning of the course, and once immediately following the completion of the course. The interview questions were semi-structured and consisted of nine questions covering a wide range of astronomy concepts typically found in traditional introductory astronomy courses. The questions were derived from alternative conception research (Comins, 1993; Schoon, 1993; Treagust & Smith, 1989), and through consultation with faculty from the astronomy department at the local university. In this paper, we focus on the questions that explore astronomical concepts related to changing frames of reference (perspectives). The question related to perspectives had three parts as follows:

1. Suppose you are standing on the Moon and you can see the Earth. When you are on the Moon, does the Earth set?

2. When you are on the Moon, does the Earth have phases? If so, what would they look like?

3. How long is a day on the Moon?

The pre-interviews, which were videotaped, were conducted during the first 2 days of the class to capture students' conceptual understanding prior to their construction of 3-D models. They examined the ability of students to articulate and explain their understanding of astronomy and to identify the prevalence of alternative frameworks. Students were provided with a set of spheres for manipulation and a white board for drawing to allow them to demonstrate their explanations. The post-interviews, which were also videotaped, were conducted during the last week of the course. We also examined student models, tests, and written reports to gain additional insight into students' use of their models in explaining their understandings. The pre-interviews typically lasted between 20 - 30 minutes and the post-interviews lasted for 30 - 45 minutes.

Data Analysis

We assessed student conceptual growth by extensive viewing of the videotapes and analysis of the transcribed interviews. We scored the student responses with a rubric (see Table 1) based upon the categorization scheme used by Simpson and Marek (1988) and Muthukrishna, Carnine, Grossen, and Miller (1993); this was modified slightly to reflect a hierarchical conceptual sequence. Both the course instructor and the researcher/interviewer scored every pre- and post-interview response with our rubric and obtained an inter-rater reliability of [r.sub.pre] = 0.90, and [r.sub.post] = 0.85.

RESULTS

During the course of the pre-interviews, it became evident that students, not surprisingly, were constructing their responses during the interview, because they would frequently state that they were simply guessing or just simply did not know. However, in the post-interviews students demonstrated a significantly improved understanding of many astronomical concepts that required a change in frame of reference. Averaging scores across all students in the class, according to our rubric, scores increased from M = 1.65 (SD = 1.15) on the pre-interview assessments to M = 3.15 (SD = 1.05) on the post-interview assessments. The students that serve as the focus of this study increased from M = 1.70 (SD = 1.01) on the pre-test to M = 3.25 (SD = .97) on the post-test interview. Scores for each individual student are summarized in Table 2.

In this section, we will examine the students' conceptual understanding from Team Yellow and Team Green.

Team Yellow

After finishing the course, both Jessica and Steve developed a better understanding of astronomical phenomena that required a change in their frames of reference. Erica had a good understanding of many astronomy concepts as noted by her responses on the pre-interview assessment (M = 3.0). However, both Steve and Jessica provided rather confused responses in their pre-interviews (M = 1.0). In their post-interviews, both Steve (M = 3.5) and Jessica (M = 3.0) improved their understanding dramatically. It is interesting to note that Erica also improved her score to M = 4.0. For example, during their pre-interviews, both Jessica and Steve conjectured that the Earth would set when viewed from the perspective of someone standing on the Moon. This is demonstrated in Steve's pre-interview response:

        Interviewer:  Does the Earth rise and set when viewed
                      from the Moon?
        Steve:        I have seen pictures titled Earth setting
                      on the Moon. I am sure it does happen.
                      Yes, it does.

Likewise, Steve also developed a more scientifically sophisticated view as shown in the following dialogue:

         Interviewer:  Let's pretend we are on the Moon and we
                       are looking back at the Earth. Can you
                       tell me will we see the Earth set? That
                       is, do we ever see the Earth go below
                       the horizon on the Moon?
         Steve:        No, since the same side of the moon
                       always faces the Earth, you would never
                       see the Earth set. The Earth would
                       always be sitting there.

When asked, in the pre-interviews, how long a day on the Moon is, both Jessica and Steve struggled to articulate a response. This is evident in Steve's response:

         Interviewer:  How long is a day on the Moon?
         Steve:        I would imagine the Moon is spinning
                       on its axis. I don't know.

During his post-interview response, Steve shows that he has developed an understanding that the Moon is locked into a synchronous orbit around the Moon, but more importantly recognizes the implications of this fact as shown in the dialogue below:

          Interviewer:  So how long would the day on the Moon
                        be?
          Steve:        It would be 27.3 days, I think.
          Interviewer:  What makes a day on the Earth?
          Steve:        The Earth's rotation. One day is the time
                        it takes for the Earth to rotate around
                        its axis. So that means that a day on the
                        Moon is 27.3 days, because that is how
                        long it takes for the Moon to rotate on
                        its axis.

Similar to Steve, Jessica could not provide a cogent response to the above question in the pre-interview assessment. However, in her post-interview, she could explain why the Earth does not rise or set when viewed from the Moon:

         Interviewer:  Let's pretend we are on the Moon and we
                       are looking back at the Earth. Can you
                       tell me if we will see the Earth set?
                       That is, do we see ever see the Earth go
                       below the horizon on the Moon?
         Jessica:      No.
         Interviewer:  Why?
         Jessica:      The same side of the Moon always faces
                       the Earth. For example, if you were on
                       the dark side, you would never see the
                       Earth.

Steve and Jessica both conjectured that the Earth had phases during their pre-interviews, but could not explain the dynamic relationships between the Sun, Moon, and Earth that produce the phases of the Earth. However, through the designing of their model, Steve and Jessica engaged in several discussions concerning where to place viewpoints in their model to examine whether the Earth does have phases when viewed from the Moon. Through these discussions and struggles with constructing their model, Steve and Jessica were in a position to accurately answer the question as shown below:

         Interviewer:  Does the Earth have phases when viewed
                       from the Moon?
         Steve:        Yes [Using spheres to demonstrate his
                       thinking]. As the Moon moves around
                       the Earth, we go through full Earth,
                       quarter Earth, new Earth, the same set
                       of phases as the Moon has.

Jessica also developed an understanding of whether the Earth has phases when viewed from the Moon, but unlike Steve, she referenced their team's model that they had designed during the course.

        Interviewer:  So does the Earth have phases when viewed
                      from the Moon?
        Jessica:      Yes. It does.
        Interviewer:  Want to use some spheres to show me?
        Jessica:      As the moon goes around the Sun, sunlight
                      hits the Earth. We definitely see phases
                      of the Earth because our model showed it.

Team Green

At the conclusion of the course, Todd, Taro, and Roger improved their understanding of astronomical phenomena that required a frame of reference change. However, their conceptual change was not as great as Steve's and Jessica's (see Table 2). Both Roger (M = 0.75) and Todd (M = 1.25) either could not articulate a response or provided rather confused responses in their pre-interviews, and still struggled in their post-interviews with both Todd and Roger improving to a final score of M = 2.50. Taro, on the other hand, began the course with a near complete understanding (M = 3.25) and improved his conceptual understanding to M = 4.0, according to our rubric.

Similar to Steve and Jessica, Todd and Roger conjectured that the Earth had phases during the pre-interviews, but could not explain the dynamical relationships between the Sun, Moon, and Earth that produce the phases of the Earth. For example, Todd attributed the phases as due to the rotation of the Earth and the interaction of the Earth and Moon as shown in Todd's pre-interview below:

         Interviewer:  Does the Earth have phases?
         Todd:         Trying to think why the Earth does that. I am
                       going to say, yes. This is a complete guess.
                       I think it is all a matter of perspective. It
                       might have to do something with the rotation
                       of the Earth. I would say, yes.

In Todd's post-interview, he was still struggling with the notion that the Earth could have phases when viewed from the Moon, but does eventually state that when one is standing on the Moon during its full phase and looking at the Earth, one will see a new Earth below:

          Todd:  [goes up to board] I am going to try to
                 explain the phases to see if that helps
                 me answer the first question. [drawing
                 the Earth-Sun-Moon system on the
                 available white board] Draws the new
                 Moon [in correct position]. Thus, when
                 the Moon is between the Earth and the
                 Sun, we see a new Moon because we can
                 see the Moon because it casts a shadow.
                 When the Moon is on the other side of
                 the Earth, we are not looking at both
                 the Sun and the Moon, thus we can see
                 the Moon. It is a similar situation if
                 you are standing on the Moon. Yes, the
                 same thing. When the Moon is behind the
                 Earth, we would see a new Earth.

Taro demonstrated his strong conceptual understanding of astronomy prior to taking the course and correctly responded during the pre-interview that the Earth does not set when viewed from the Moon because the same side of the Moon always faces the Earth. Whereas Roger, recalling viewing photographs taken from the Moon, responded correctly, but could not elaborate on his response. In his post-interview, Roger's conceptual understanding concerning whether the Earth sets when viewed from the Moon is rather fragile:

        Roger:  No, it is not always visible. If the
                Moon is full then you have a new Earth.
                Wait, is that right? The Sun is blocking
                the Moon. I am going to say no. I wish I
                had a really good response. I am going
                to say no, because the Moon doesn't
                produce its own light so we can only see
                it when the Sun shines on it.

Roger did not refer to his model and his conceptual understanding did not change significantly from the pre-interview. On the other hand, Todd, who could not express a cogent response to the question during the pre-interviews, was very thoughtful and called upon models used during the class to demonstrate his understanding during his post-interview:

        Todd:  No, the reason I think? There are times
               when we can't see the Earth like at new
               Earth, but I don't think we are going to
               see the Earth rise and fall. I want to say
               no, because if you are on the same side,
               the moon's rotation and revolution go at
               the same rate and you put a flag on the
               Moon, you are still going to look at the
               Earth constantly if you are on the other
               side you won't. I don't think that it
               does.

Changing perspective to solve and investigate a problem is a rather difficult concept for most beginning science students, but it is particularly important for learning astronomy (Gazit, Chen, & Yair, 2004). There are many astronomical concepts that are easier to comprehend if a different reference frame can be accessed from which to examine the problem. For example, two of these problems (i.e., When you are on the Moon does the Earth set? and Does the Earth have phases?) would be best understood if one had the means to stand on the Moon and observe the Earth and the Sun. This is difficult to do in a normal classroom. However, using VRML computer models students can easily shift their reference frame by setting viewpoints in their models. From the results discussed above, the students who participated in the course experienced significant gains in their conceptual knowledge of astronomical phenomena that required a change in perspective.

DISCUSSION AND CONCLUSIONS

Computer-based 3-D technologies create exciting opportunities for students to create, manipulate, and interact with their own constructions, which in turn support them in developing understandings through their first-hand experience, especially with respect to learning astronomy. For example, many students are asked to learn 3-D astronomical concepts through the examination and study of 2-D images and graphs, which are difficult to interpret and hard to mentally visualize in three dimensions (Parker & Heywood, 1998b). Unfortunately, it is typically left to the students to somehow abstract in their minds these 2-D diagrams and images into forms that make sense to them in 3-D (Gazit et al., 2004; Yair et al., 2001). To do this requires students to make intuitive leaps that do not have any concrete attachment to previous experience. For example, the causes of eclipses cannot be adequately represented in a 2-D format because the Moon is not in the same orbital plane as the Earth and the Sun, but tilted at an angle and revolving around a rotating Earth within 3-D space. Therefore, if we are to better support students' learning of astronomy, it is necessary that instructional activities reflect this inherent spatial and dynamic nature of astronomy. The results of this study are a first step to the development of an understanding of what particular aspects of astronomy can best be taught using 3-D models and visualizations.

Our findings, which much be tempered by our small sample size, suggest that the incorporation of 3-D modeling activities in science courses has the potential to facilitate students' reevaluation of their alternative frameworks. Conceptual change theory as posited by Posner, Strike, Hewson, and Gertzog (1982), requires that students first be dissatisfied with their existing conceptual understanding before meaningful change can occur. During the interviews, we observed that by being engaged in model-building and evaluation activities, students can quickly compare their existing understanding with their model and then reevaluate their understanding based upon feedback from their interactions with their model (Penner, et al., 1998). This process appears to be facilitated when students are provided with activities that provide them direct experience with the concepts under study (diSessa & Minstrell, 1998). In this study, we found that 3-D computational models allow students to construct a realistic model that they can "step into" and shift their frame of reference from one perspective to another. This affords them multiple opportunities to examine their understanding from multiple perspectives.

This ability to change perspective is particularly important for astronomy education because to understand the reasons for eclipses, a student must, at a minimum, understand the revolution of the Earth and the Moon and their orbital paths. Therefore, if students can view the Earth-Moon-Sun system from different perspectives, the likelihood that they will develop a scientifically accurate explanation is greatly enhanced because they can test if their understanding holds from many different vantage points (e.g., Is the configuration of the Earth-Moon-Sun the same during a lunar eclipse when viewed from the Earth and the Moon?). In closing, we feel that new and emerging technologies have significant potential to improve student learning, but it is critical that as a community we examine which aspects of these tools are best leveraged for improving student understanding of scientific concepts.

Table 1

Rubric Used to Analyze Student Pre- and Post-Interview Responses.

Score  Category       Description

0      No conception  The students is unable to articulate a
                      response to the question

1      Confused       The student guesses at correct responses and
                      may guess one or two correctly, but cannot
                      extend or clarify what the response means.
                      The student shifts answers as he or she is
                      responding and has alternative frameworks.

2      Incomplete     The student indicates that the Earth does not
       understanding  set from the Moon, but has incorrect
                      reasoning concerning why. The student
                      responds that the Earth does not have phases,
                      but suspects that it might have phases. The
                      student does not respond with the correct
                      length of the day on the Moon.

3      Partial        The student indicates that the Earth does not
       understanding  set from the Moon, and states the reason is
                      that the Moon is in synchronous orbit around
                      the Earth. However, the student does not
                      understand what synchronous rotation is. The
                      student also knows that the Earth has phases,
                      and that they are caused by how much the
                      Earth is being struck by sunlight. The
                      student response also indicates knowledge
                      that the length of a day on the Moon is
                      around 28days (a month), but does not connect
                      the time he or she states to the Moon's
                      rotation rate.

4      Complete/      The student indicates that the Earth does not
       Sound          set from the Moon, states the reason is that
       understanding  the Moon is in synchronous rotation about the
                      Earth, and clarifies what is meant by the
                      term synchronous rotation. The student also
                      knows that the Earth has phases, and that
                      they are opposite those of the Moon or caused
                      by how much of the Earth is being struck by
                      sunlight. The student response also indicates
                      that he or she knows that the length of a day
                      on the Moon is around 28 days, and connects
                      that time with the Moon's rotation rate.

Table 2

Summary of Student Scores on the Interview Questions That Require the
Students to Change Their Perspectives

Student  Pre        Post       Change

Todd     0.75       2.50       1.75
Roger    1.25       2.50       1.25
Taro     3.25       4.00       0.75
Steve    1.00       3.50       2.50
Jessica  1.00       3.00       2.00
Erica    3.00       4.00       1.00

Average  1.70       3.25       1.54

         SD = 1.11  SD = 0.69  SD = 0.66

Acknowledgement

We would like to thank the Virtual Reality/Virtual Environments group for its support in using the CAVE Automatic Virtual Environment (CAVE).

REFERENCES

Ainge, D. J. (1996). Introducing primary students to VR with Virtus Walk-Through: Two pilot studies. VR in the Schools, 2(1), 10-15.

Atwood, R. K., & Atwood, V. A. (1996). Preservice elementary teachers' conceptions of the causes of seasons. Journal of Research in Science Teaching, 33, 553-563.

Barnett, M., & Morran, J. (2002). Addressing childrens' understanding of the Moon's phases and eclipses. International Journal of Science Education, 24(8), 859-879.

Bell, J. T., & Fogler, H. S. (1995). The investigation of virtual reality based educational module for safety and hazard evaluation training. Paper presented at the American Society for Engineering Education, Indiana Sectional Meeting, Peoria, IL.

Blumenfeld, P., Soloway, E., Marx, R., Krajcik, J., Guzdial, M., & Palincsar, A. (1991). Motivating project-based learning: Sustaining the doing, supporting the learning. Educational Psychologist, 26(3&4), 369-398.

Byrne, C. M. (1996). Water on tap: The use of virtual reality as an educational tool. University of Washington, Seattle, WA.

Cobb, S., Crosier, H., Wilson, J., & Wilson, J. R. (2002). Development and evaluation of virtual environments for education. In K. M. Stanney (Ed.), Handbook of virtual environments; Design, implementation, and applications (pp. 911-936). Mahwah, NJ: Lawrence Erlbaum Associates.

Comins, N. (1993). Sources of misconceptions in astronomy. Paper presented at the Proceedings of the Third International Seminar on Misconceptions and Educational Strategies in Science and Mathematics, Ithaca, NY.

Dede, C. (2000). Emerging influences of information technology on school curriculum. Journal of Curriculum Studies, 32(2), 281-303.

Dede, C., Salzman, M., Loftin, R. B., & Sprague, D. (1999). Multisensory immersion as a modeling environment for learning complex scientific concepts. In W. Feurzeig & N. Roberts (Eds.), Modeling and simulation in science and mathematics education. New York: Springer Verlag.

diSessa, A., & Minstrell, J. (1998). Cultivating conceptual change with benchmark lessons. In J. Greeno & S. Goldman (Eds.), Thinking practices in mathematics and science learning (pp. 155-187). Mahwah, NJ: Lawrence Erlbaum Associates.

Dixon, J. K. (1997). Computer use and visualization in students' construction of reflection and rotation concepts. School Science and Mathematics, 97(7), 352-358.

Furness, T., Winn, W., & Yu, R. (1997). The impact of three dimensional immersive virtual environments on modern pedagogy (Prepared for the National Science Foundation). Seattle: University of Washington.

Gay, E. (1994). Is virtual reality a good teaching tool? Virtual Reality Special Report, Winter, 1(4), 51-59.

Gazit, E., Chen, D., & Yair, Y. (2004). Developing understanding of basic astronomical concepts by using a virtual Solar System. Paper presented at the Sixth International Conference of the Learning Sciences, Santa Monica, CA.

Gotwals, R. R. (1995). Scientific visualization in chemistry, better living through chemistry, better chemistry through pictures: Scientific visualization for secondary chemistry students. In D. A. Thomas (Ed.), Scientific visualization in mathematics and science teaching (pp. 153-179). Charlottesville, VA: AACE.

Hay, K. E., Marlino, M., & Holschuh, D. R. (2000). The virtual exploratorium: Foundational research and theory on the integration of 5-D modeling and visualization in undergraduate geoscience education. In B. Fishman & S. O'Connor-Divelbiss (Eds.), Proceedings of the 4th International Conference of the Learning Sciences (214-220). Mahwah, NJ: Lawrence Erlbaum Associates.

Keating, T., Barnett, M., Barab, S.A., & Hay, K. E. (2002). Developing conceptual understanding of scientific concepts through building three-dimensional computational models. Journal of Science Education and Technology, 11(3), 261-275.

McLellan, H. (1996). Virtual realities. In D. Jonassen (Ed.), Handbook of research for educational communications and technology (pp. 457-487). Boston: Kluwer-Nijhoff Publishing.

Moshell, J., Michael, J., & Hughes, C. E. (2002). Virtual environments as a tool for academic learning. In K. M. Stanney (Ed.), Handbook of virtual environments; Design, implementation, and applications (pp. 893-910). Mahwah, NJ: Lawrence Erlbaum Associates.

Muthukrishna, N., Carnine, D., Grossen, B., & Miller, S. (1993). Children's alternative frameworks: Should they be directly addressed in science education? Journal of Research in Science Teaching, 30, 233-248.

Osberg, K. M., Winn, W., Rose, H., Hoffman, H., & Char, P. (1997). The effect of having grade seven students construct virtual environments on their comprehension of science. Paper presented at the American Educational Research Association, New Orleans, LA.

Parker, J., & Heywood, D. (1998a). The earth and beyond: Developing primary teachers' understanding of basic astronomical events. International Journal of Science Education, 20(5), 503-520.

Parker, J., & Heywood, D. (1998b). The earth and beyond: Developing primary teachers' understanding of basic astronomical events. International Journal of Research in Science Education, 20(5), 503-520.

Pea, R. D. (1993). Practices of distributed intelligence and designs for education. In G. Salomon (Ed.), Distributed cognitions: Psychological and educational considerations (pp. 47-87). New York: Cambridge University Press.

Penner, D. E., Lehrer, R., & Schauble, L. (1998). From physical models to biomechanics: A design-based modeling approach. The Journal of the Learning Sciences, 7(3&4), 429-449.

Pfundt, H., & Duit, R. (2004). Students' alternative frameworks and science education bibliography. Retrieved May 11, 2004, from http://www.ipn.uni-kiel.de/aktuell/stcse/stcse.html

Posner, G. J., Strike, K. A., Hewson, P. W., & Gertzog, W. A. (1982). Accommodation of a scientific conception: Toward a theory of conceptual change. Science Education, 66, 211-227.

Pyramid Film & Video (1988). A private universe: An insightful lesson on how we learn: Harvard-Smithsonian Center for Astrophysics.

Roschelle, J. (1992). Learning by collaborating: Convergent conceptual change. Journal of the Learning Sciences, 2, 235-276.

Roth, W.-M. (1995). Authentic School Science. Netherlands: Kluwer Publishers.

Roth, W.-M. (1996). Art and artifact of children's designing: A situated cognition perspective. Journal of the Learning Sciences, 5(2), 129-166.

Salzman, M. C., Dede, C., & Bowen-Loftin, R. (1995). Learner-centered design of sensorily immersive microworlds using a virtual reality interface. Paper presented at the 7th annual International Conference on Artificial Intelligence and Education, Pittsburgh, PA.

Schoon, K. J. (1993). The origin of earth and space science misconceptions: A survey of pre-service elementary teachers. Paper presented at the Proceedings of the Third International Seminar on Misconceptions and Educational Strategies in Science and Mathematics, Ithaca, NY.

Simpson, W. D., & Marek, E. A. (1998). Understandings and misconceptions of biology concepts held by students attending small high schools and students attending large high schools. Journal of Research in Science Teaching, 25, 361-374.

Smith, J. P., diSessa, A. A., & Roschelle, J. (1992). Misconceptions reconceived: A constructivist analysis of knowledge in transition. The Journal of the Learning Sciences, 3(2), 115-163.

Sneider, C., & Ohadi, M. (1998). Unraveling students' misconceptions about the Earth's shape and gravity. Science Education, 2, 265-284.

Stahly, L. L., Krockover, G. H., & Shepardson, D. P. (1999). Third grade students' ideas about the lunar phases. Journal of Research in Science Teaching, 36, 159-177.

Stratford, S. J., Krajcik, J., & Soloway, E. (1998). Secondary students' dynamic modeling processes: Analyzing, reasoning about, synthesizing, and testing models of stream ecosystems. Journal of Science Education and Technology, 7(3), 215-234.

Thomas, L. C., & Wickens, C. D. (2001). Visual displays and cognitive tunneling: Frames of reference effects on spatial judgments and change detection. Paper presented at the 45th Annual Meeting of the Human Factors and Ergonomics Society., Santa Monica, CA.

Treagust, D. F., & Smith, C. L. (1989). Secondary students' understanding of gravity and the motion of planets. School Science and Mathematics, 89, 380-391.

Wandersee, J. H., Mintzes, J. J., & Novak, J. D. (1994). Research on alternative conceptions in science. In D. Gabel (Ed.), Handbook of research on science teaching and learning. New York: Macmillan.

Wilson, P. N., Foreman, N., & Tlauka, M. (1997). Transfer of spatial information from a virtual to a real environment. Human Factors, 39, 526-531.

Windschitel, M., Winn, W., & Headley, N. (2001). Using immersive visualizations to promote the understanding of complex natural systems: Learning inside virtual Puget Sound. Paper presented at the National Association for Research on Science Teaching, Seattle, WA.

Windschitl, M., & Winn, W. (2000). A virtual environment designed to help students understand science. Paper presented at the Proceedings of the Fourth International Conference of the Learning Sciences, Ann Arbor, MI.

Windschitl, M., Winn, W., & Headley, N. (2001). Using immersive visualizations to promote the understanding of complex natural systems: Learning inside virtual Puget Sound. Paper presented at the National Association for Research on Science Teaching., Seattle, WA.

Winn, W. D., & Jackson, R. (1999). Fourteen propositions about educational uses of virtual reality. Educational Technology, 39(4), 5-14.

Yair, Y., Mintz, R., & Litvak, S. (2001). 3D-Virtual reality in science education: An implication for astronomy teaching. Journal of Computers in Mathematics and Science Teaching, 20(3), 293-305.

Youngblutt, C. (1998). Educational uses of virtual reality technology (No. IDA Document Report no. D-2128). Alexandria, VA: Institute for Defense Analyses.

MICHAEL BARNETT 

Boston College 

USA

[e-mail]

LISA YAMAGATA-LYNCH

Northern Illinois University 

USA

TOM KEATING 

Tech Museum of Science and Innovation

USA

SASHA A. BARAB

Indiana University 

USA

KENNETH E. HAY

Indiana-Bloomington

USA