The article by Richard R. Hake discusses the analysis of change/gain scores in educational settings, particularly focusing on the comparison between Interactive Engagement (IE) and Traditional (T) teaching methods. Hake defines "Interactive Engagement" as methods that promote conceptual understanding through interactive activities and immediate feedback, while "Traditional" methods rely on passive lectures and algorithmic problem-solving. The study uses a normalized gain <g> to measure the effectiveness of courses in promoting conceptual understanding, defined as the ratio of the actual average gain to the maximum possible average gain. Key findings include: - All 14 T courses fell in the Low-g region, with an average normalized gain of 0.23. - 85% of 48 IE courses fell in the Medium-g region, with an average normalized gain of 0.48. - No courses fell in the High-g region. - The correlation between normalized gain <g> and initial test scores (<S_i>) is very low (+0.02), suggesting that <g> is a suitable measure for comparing course effectiveness across diverse student populations. - An effect size (d) of 2.78 is calculated from the difference in normalized gains between IE and T courses, indicating a significant improvement in conceptual understanding with IE methods. - The study also compares these results with a meta-analysis of small-group learning, finding that IE methods achieve a higher effect size (d = 2.78) compared to the lower effect size (d = 0.51) reported in the meta-analysis. Hake suggests that normalized gain <g> is a robust method for analyzing pre/post-test results and recommends its use over ANCOVA and other traditional methods.The article by Richard R. Hake discusses the analysis of change/gain scores in educational settings, particularly focusing on the comparison between Interactive Engagement (IE) and Traditional (T) teaching methods. Hake defines "Interactive Engagement" as methods that promote conceptual understanding through interactive activities and immediate feedback, while "Traditional" methods rely on passive lectures and algorithmic problem-solving. The study uses a normalized gain <g> to measure the effectiveness of courses in promoting conceptual understanding, defined as the ratio of the actual average gain to the maximum possible average gain. Key findings include: - All 14 T courses fell in the Low-g region, with an average normalized gain of 0.23. - 85% of 48 IE courses fell in the Medium-g region, with an average normalized gain of 0.48. - No courses fell in the High-g region. - The correlation between normalized gain <g> and initial test scores (<S_i>) is very low (+0.02), suggesting that <g> is a suitable measure for comparing course effectiveness across diverse student populations. - An effect size (d) of 2.78 is calculated from the difference in normalized gains between IE and T courses, indicating a significant improvement in conceptual understanding with IE methods. - The study also compares these results with a meta-analysis of small-group learning, finding that IE methods achieve a higher effect size (d = 2.78) compared to the lower effect size (d = 0.51) reported in the meta-analysis. Hake suggests that normalized gain <g> is a robust method for analyzing pre/post-test results and recommends its use over ANCOVA and other traditional methods.