We divided each semester's students into four groups of approximately the same size based on the students' cumulative GPAs, and we used independent-samples t tests to compare the lowest GPA group in the control section, spring 2012, with the corresponding GPA group in each of the treatment sections, fall 2012 and spring 2013, in terms of mean scores on the ACS exam. The results are displayed in Figs. 3 and 4.
As the figures show, it seems that the impact of the change from a traditional to a hybrid, flipped format affected all aptitude/ability segments of the student population approximately equally. For the spring 2012efall 2012 contrast, we found a 3e4 point advantage for the hybrid-flipped version of the class in the lower three quartiles, and a 1-point advantage in the top quartile. For the spring 2012espring 2013 contrast, we found small advantages (0.5e2 points) for the hybrid-flipped version in the bottom two and top quartiles, along with a 0.35- point nominal advantage for the traditional version in the second-highest quartile.
None of the differences in either contrast were statistically significant, probably due to the lower per-quartile N, and a ceiling effect may explain the finding of somewhat lower gains in the top quartile as compared to the lower quartiles