This paper presents the results of a confirmatory factor analysis on two self-efficacy scales designed to probe the self-efficacy of college-level introductory astronomy (Astro-101) students (n ¼ 15181) from 22 institutions across the United States of America and Canada. The students undertook a course based on similar curriculum materials, which involved students using robotic telescopes to support their learning of astronomical concepts covered in the "traditional" Astro-101 courses. Previous research by the authors using these self-efficacy scales within a pre-/post-test approach showed both high reliabilities and very high construct validities. However, the scale purporting to measure students' self-efficacy in relation to their use of the astronomical instrumentation associated with online robotic telescopes was particularly skewed and required further investigation. This current study builds on the previous work and shows how a slight adjustment of the survey items presents an improved and robust scale for measuring self-efficacy.