Education Spatial Intelligence with Augmented Reality

Spatial abilities present an important component of human intelligence. Spatial ability is a heterogeneous construct that comprises a number of subfactors, such as mental rotation, visualization, and environmental orientation. Many studies have shown that spatial abilities can be improved by well-designed trainings (e.g., Souvignier, 2001; see overview in Section 2.4). Geometry education has proven as orte powerful means of improving these skills; recently, a number of training studies have shown the usefulness of virtual reality (VR) in training spatial ability. However, little to no work has been done towards systematic development of VR applications for practical education purposes in this field. This project aims to introduce Augmented Reality (AR), a technology closely related to VR, to mathematics and geometry education. The simultaneous sharing of real and virtual space possible in AR is an ideal match for computer-assisted collaborative education settings. However, there are a number of requirements for an augmented reality tool with the aim of effectively improving spatial skills. These have never been addressed by existing systems, nor studied in an educational context. No VR/AR application for actual use in high scheel or higher education has ever been developed with the main purpose of improving spatial skills. Hardly any evaluations can be found in literature which give hints to the actual learning transfer from a VR/AR leaming environment to the real world. The proposed project is intended to address these two key issues by developing a prototype AR system for geometry education, thereby building a bridge between human psychology research and AR research in Austria. As a pilot study for this proposal, a three dimensional geometric construction tool called Construct3D (Kaufmann et al., 2000) has been developed that will serve as a platform for the proposed research work. A comprehensive evaluation study will be conducted in order to study the general and differential effects of the training an several components of spatial ability. The project is proposed for an initial duration of 2 years with a potential 1 year extension. The project is not intended to be a "orte-time installation". The mid- to long-term plans are to integrate it in high school and higher education by collaborativg with Austrian schools and external partners such as the Institute of Geometry at Vienna University of Technology.

This project examined the effects of an Augmented-Reality-based training on spatial ability. Spatial ability, a component of human intelligence, concerns the mental representation and manipulation of visual-spatial information. Spatial ability plays an important role in many everyday activities as well as educational and occupational careers. As many studies have shown, spatial abilities are trainable, especially by hands-on experience with three-dimensional objects. Such experience is a typical feature of Augmented Reality (AR) technology, which combines real environments with virtual objects: Through special glasses, users see both the actual environment and three-dimensional objects projected into space. In Construct3D, the AR software used here, three-dimensional objects can be constructed, viewed, and manipulated. In a pre-posttest control design, the hypothesis was tested that active interaction with three-dimensional objects in Construct3D has positive effects on spatial ability. Pre- and posttests measured the three main subcomponents of spatial ability; 317 students (14 to 20 years old) participated. Between pre- and posttest, 47 students took a 6-week training with Construct3D, consisting of geometry tasks in a dyadic setting (two students and a tutor). Three comparison groups were included: 1) training with CAD3D (a construction software for three-dimensional objects in two-dimensional screen presentation), 2) a control group taking geometry classes at school, and 3) a control group taking no geometry classes. Data on predictors of spatial ability such as interest, self-confidence, experience, and strategy use were also collected. Two results are especially interesting and bring up new questions. First, all groups improved between pre- and posttest, but one of the largest positive changes was shown by girls who took no geometry classes or training. Their improvements can only be attributed to practice induced by the pretest. Thus, the spatial abilities of individuals with little prior experience may be underestimated if they are only tested once - even a small amount of practice may result in marked improvements. The second important finding concerns the actual training effects. The Augmented-Reality training had relatively small effects compared to the CAD3D training. One explanation is that both CAD3D and the spatial tests require transformation of two-dimensional pictures into three-dimensional mental representations, mental manipulation of the representations, and back-transformation into a two-dimensional picture. As Construct3D presents objects three-dimensionally, it makes no such requirements - and in our view, the same is true for many real-life activities involving spatial abilities. This may be one explanation why correlations between spatial test performance and everyday spatial skills are often low. In a follow-up FWF project we develop an Augmented-Reality-based three-dimensional test of spatial abilities. In order to improve measurement in individuals with little prior experience (such as the girls with no geometry education), the test will be a dynamic assessment, that is, it consists of a pretest, a training phase, and a posttest. In this way, developmental potential can be assessed in addition to developmental status.