Abstract

The Statistical educational process entails the deployment of diverse methodologies and levels of theoretical frameworks to mitigate understanding of complex statistical concepts effectively. Specifically , the confidence intervals subject is commonly encountered by students as the first time in the field of statistical inference, which is a domain where statistics plays a pivotal role in decision-making processes. This first meeting introduces terminology and procedures that can be perplexing and overwhelming, potentially influencing subsequent attitudes and approaches towards statistics and mathematics. There is a substantial body of instruments, techniques, and considerations that exists, striving to augment the efficacy of confidence interval instruction but these methods are showing mixed results and today’s curricula mostly rely on basic educational paradigms. Additionally, field-of-study disparities come into play, with STEM (Science, Technology, Engineering, and Mathematics) students typically possessing a stronger mathematical foundation, contrasting with non-STEM counterparts who often benefit more from an intuitive approach to statistical concepts. Consequently, effective statistics education necessitates a blend of formal and conceptual teaching strategies, tailored to each group. All of these complexities together with the proliferation of computers and digital tools in education, coupled with evolving student expectations. Necessitates the exploration of innovative and engaging pedagogical methodologies leveraging these technologies.

This research study investigates the effectiveness of computer-based simulation-based learning for statistics education among both STEM and non-STEM undergraduate students. To this end, a Java-based web simulation has been designed to facilitate the interpretation of confidence intervals as the probable range encompassing the expected value. Participants have the ability to manipulate two key parameters, sample size and confidence level, and observe their impact on the range of confidence intervals. Furthermore, statistics anxiety, a prevalent issue among non-STEM students, can pose a significant obstacle to their learning and performance in statistics coursework. This study aims to evaluate whether simulation-based learning can mitigate the negative effects of statistics anxiety. The study's primary focus is to assess the effectiveness of computer simulations on participants from behavioral sciences (non-STEM, n=170) and life sciences (STEM, n=130). Pre- and post-simulation assessments ascertain the immediate impact on comprehension of confidence intervals, while long-term benefits are gauged by examining the simulation's influence on end-of-course grades, particularly in questions pertaining to the simulation material on confidence intervals. Attitudes towards statistics and academic self-efficacy are measured for each participant using the SATS and GASE questionnaires, respectively. Preliminary findings indicate that both STEM and non-STEM students exhibited significant short-term improvements in their understanding of statistical concepts post-simulation, suggesting the value of this approach in enhancing statistics education across various disciplines and potentially alleviating statistics anxiety. These findings underscore the need for developing novel teaching tools and have implications for pedagogical strategies in statistics education, emphasizing the potential of simulation-based learning to optimize student outcomes and reduce anxiety across diverse academic backgrounds. Data collection and analysis are ongoing.