Augmented reality (AR) is an interactive system that allows users to interact with virtual objects and the real world at the same time. The purpose of this dissertation was to explore how AR, as a new visualization tool, that can demonstrate spatial relationships by representing three dimensional objects and animations, facilitates students to learn chemistry, a discipline requiring students' visual-spatial thinking to comprehend abstract concepts. In addition, this dissertation examined the effect of AR in a collaborative learning environment and the impact of visual-spatial ability and cognitive load on student learning performance. The Protein Magic Book, developed by Human Interface Technology Laboratory at the University of Washington and the SCRIPPS Research Institute at La Jolla, was the learning material in this study. The Protein Magic Book introduces basic concepts of protein structures with AR representations. Students were randomly assigned into three settings, including studying with texts only (N = 26), studying with AR alone (N = 26), and studying with AR in dyads (22 pairs). Totally, ninety six students participated in this study. They were required to complete background questionnaire, chemistry self-efficacy scale, and chemistry knowledge test, before they studied the protein structures. After the learning activity, they took the post-test, cognitive load scales, and visualization rotation test. The results of data screening showed that a nested effect existed within dyads. Therefore, the Hierarchical Linear Modeling (HLM) was employed in data analysis. The results indicated that students in the AR alone setting gained greater learning performance than those who studied with texts. However, students in the collaborative setting did not perform better than those who studied alone with AR scaffolding. Students with higher visual-spatial ability performed better, and students with higher spatial ability reported less cognitive load, especially AR load. Cognitive load, on the other hand, did not affect student learning performance in this study and students in the collaborative setting did not report higher cognitive load than other groups. Discussion of findings, limitations of this study, and future research directions are presented. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]