Table 1: Course codes, their associated entry-level requirements (Grade 12 Mathematics and Physical Sciences scores), FCI deployment period, number of student responses per class, and average scores for five first-year classes involved in the 2022 FCI testing (N = 337). `Ext.’ refers to the extended courses, where students who do not meet entry-level requirements for introductory physics classes can complete a four-year Bachelor of Science in which the traditional first-year physics course is taught over a three-semester period. Physics Course Ent. % Maths Ent. % Phys. Sci. Deployment Period Responses/ Class Mean (%) PHYS1A1 (Majors) 70 60 21/02 - 25/02 28/55 36.7 PHY1EA1(Phys. Ext. Sem. 1) 60 50 21/02 - 25/02 176/250 26.7 PHYE0A1 (Engineering) 60 60 24/02 - 28/02 105/500 33.3 PHYG1A1 (Earth Sci.) 60 50 21/02 - 25/02 9/9 33.3 PHYL1A1 (Life Sci.) 70 50 21/02 - 25/02 19/40 30.3 Our output and innovations: i) Quantifying conceptual understanding by identifying dominant misconceptions From analyses of the dominant responses, we found that the most popular correct answers were observed for questions 1-3 that focused on falling objects, and questions 4, 15, 16, and 28 that tested Newton’s third law. We surmise that these concepts were well understood. Questions 5, 11, 17, 19, 26, and 30 had an incorrect answer as the dominant response. These questions involved an interplay of active forces and objects in motion, which highlight what Martin-Blas et al. call ‘dominant misconceptions’. For example, many of the incorrect answers preferred by students were derived from misconceptions around the relationship between force, velocity, acceleration, and motion. Given that this information was collected at the very beginning of the semester, before lectures got underway, the course instructors could attempt to adjust the teaching and learning to try to elucidate these poorly understood concepts. ii) Correlating Matric results and university performance Table 2: Pearson’s correlation coefficients for different classes. We distinguish between English and non-English home languages using ‘Eng’ and ‘Other’, respectively. Matric / Pretest Corr. Coeff. PHYS1A1 (Majors) PHY1EA1 (Ext. Sem. 1) PHYG1A1 (Earth Sci.) PHYE0A1 (Engineering) Eng. Other Eng. Other Eng. Other Eng. Other Phys. Sci. +0.375 +0.683 +0.125 +0.314 -0.945 -0.445 +0.060 +0.303 Maths +0.147 +0.707 +0.338 +0.220 -0.979 -0.559 +0.562 +0.521 English +0.293 -0.157 +0.214 +0.103 +0.645 -0.614 +0.099 -0.054 In Table 2, we measured the strength of the linear correlation between a Matric subject and the FCI score using Pearson’s correlation coefficient p, where -1 ≤ p ≤ 1. For a perfect correlation, p = 1; for a perfectly inverse relationship, p = -1. We do note that the correlations presented were for fairly small groups of students, reducing the overall reliability e.g., three of the nine students in the PHYG1A1 cohort spoke English at home. More rigorous statistical testing with larger group sizes will be left for future research. However, there does seem to be an overall trend that correlations between English scores and the FCI are more positive for students whose home language is English. This is in agreement with the observations made in Ref. [8], where FCI performance was found to have a dependence on English reading ability. Teaching Innovation for the 21st Century | Showcasing UJ Teaching Innovation Projects 2022 29
RkJQdWJsaXNoZXIy MjU1NDYx