David Webb

@djphysicswebb.bsky.social

210 Followers 583 Following

Thinks a lot about teaching and learning. Retired physicist. Hunts for demographic inequity within the system. #EduSky 🍎 #Science 🧪

As explained in the text, each in-term exam was z-scored on a class-by-class basis and the gender gap for each group of classes was determined with HLM while controlling for each student's incoming GPA. This in-term exam score gender gap is plotted as a function of time through the 10-week term at the times the exams and retakes were taken. On the right, after the end of the term, are the overall averages for those first-attempt exams and for the recorded scores after retakes were completed. One sees that the female students generally had higher exam scores (i.e. a positive gender gap) after the first attempt and that, overall, retakes favored the male students. For reference the (clearly negative) gender gaps for 179 non-retake classes are also shown on a week-by-week basis through their terms.
Graph with final exam grade gaps between students from historically marginalized racial/ethnic groups and their peers. Two types of class are compared: 1) traditional physics courses where topics were introduced chapter-by-chapter as usual and 2) a class where all of the concepts from the entire quarter are introduced in the first 60% of the term and then computations are practiced (using those concepts) in the final 40% of the term. On the same final exam (graded at the same time) the concepts first class had no demographic grade gap while the usual classes had grade a grade gap of about 0.7 standard deviations.
Graph showing a 0.2 grade point gender gap in physics courses where students were not offered retake exams and no gender gap in the same physics course but with retake exams offered to all students.
Graph showing that active learning physics courses at San Jose State Univ. and at UC Davis had roughly 60% to 80% fewer students dropping the course than dropped the lecture-based courses that they supplanted. Also showing that students from historically marginalized racial/ethnic groups dropped the lecture-based course at higher levels than their peers but dropped the active-learning courses at about the same levels.
A graph comparing the grade gaps between physics students from historically marginalized racial/ethnic groups from classes using percent scale grade and classes using GPA-like (4-point) scale grading. The grade gaps are nonzero even after controlling for physics understanding so they represent a grading bias. Importantly, the percent scale grade gap is more than three times larger than the 4-point scale grade gap showing that the percent scale amplifies an underlying grade bias against historically marginalized students.
Graph showing early life blood levels in US children - showing that if you were 38-50 in 2015 then you had the highest blood levels in your early life. From: https://www.pnas.org/doi/10.1073/pnas.2118631119