This chapter provides a review, using both qualitative and quantitative approaches, of experimental studies on natural language interfaces. As data utilization is an important aspect of information systems, numerous experimental studies have been conducted on user performance involving database-related tasks with various database models and/or languages. We propose a two-dimensional conceptual framework aimed at classifying and systematically analyzing these studies, in order to provide a bigger picture facilitating systematic understanding of this body of research. The classification exercise shows that studies involving natural language interfaces did not have very consistent findings. Correspondingly, we applied the meta-analytic technique to attempt to gain insight into these differences.