Investigating the Evolution of Student Attitudes toward Science in a General Chemistry Course Using Latent Class and Latent Transition Analysis

• Oluwatobi O. Odeleye ^[1] ; Oluwaseun D. Agunbiade ^[1] ; Adam Garber ^[2] ; Karen Nylund-Gibson ^[2]
  1. [1] West Virginia University

     West Virginia University

     Estados Unidos

     
  2. [2] University of California − Santa Barbara, United States
• Localización: Journal of chemical education, ISSN 0021-9584, Vol. 102, Nº 5, 2025, págs. 1745-1754
• Idioma: inglés
• Enlaces
  • Texto completo
• Resumen

  • As science, technology, engineering, and mathematics (STEM) education researchers continue to explore ways to increase college student persistence in STEM fields, the affective domain (e.g., attitudes, perceptions, and self-efficacy) stands out as an area that can significantly impact these efforts. Latent class analysis (LCA) and latent transition analysis (LTA) are mixture modeling approaches that take a person-centered approach to quantitative research, which can help us to further our efforts to diversify STEM fields. This study seeks to use LCA and LTA to investigate how students’ attitudes toward science in general chemistry evolve over a semester. Using the Modified Attitudes toward Science Inventory (mATSI), we grouped students based on their responses to pre- and postsurvey items from the mATSI. We found three distinct groups (classes) of students at the beginning of the semester: (i) students with strong desires to pursue science fields and high self-belief in their abilities to do well in science courses (high–high), (ii) students with moderate desires and low self-belief (mod-low), and (iii) students with moderate desires to pursue science fields and moderate self-belief (mod-mod). Over the course of the semester, these 3 groups evolved into (a) high desires and high self-belief (high–high), (b) high desires and low self-belief (high-low), and (c) low desires and low self-belief (low-low). At the beginning of the semester, about 80% of the participants were classified in the high–high group with the remaining 20% categorized into the other two groups; however, by the end of the semester, about 70% were in the high–high group, with 30% distributed across the other two groups. Using LTA and exploring the characteristics of the student groups, we found that in groups where female and second-year students were overrepresented, male and first-year students tended to be underrepresented and vice versa. For example, female and second-year students were overrepresented in groups more likely to leave the general chemistry course with lower desires and self-belief, while male and first-year students were overrepresented in groups more likely to leave general chemistry with higher desires and self-belief Using the LCA approach, we were able to explore groups (e.g., “high-low” and “low-low”) that tend to get swallowed up by the noise of the majority (in this case, the “high–high” group). We hope the findings from this study encourage equity-based researchers to continue to think about how they approach quantitative data to give a voice to participant groups that may sometimes be hidden under the guise of not having enough statistical significance/power.