Assessing gender differences in response system questions for an introductory physics course
Typical examples of lower level (a) and higher level (b) questions from LB 272. The top question asks the student to “remember” how the net resistance is calculated for resistors in parallel and in series. The bottom question requires the student “analyze” the system by identifying the system's component parts and how the relationships among the components translate graphically.
The fraction of correct clicker responses within a single response period without peer discussion as a function of initial and final response times. The data is further divided into separate cognitive levels. We find that there are no statistically significant differences between men and women, regardless of Bloom's level. This shows that genders are correctly responding on the same timescale. We assume Poisson statistics for each bin. The total number of initial responses totaled 3099 and 4030, and final responses totaled 2742 and 3676, for men and women, respectively.
The fraction of responses for all questions as a function of final response time for a single response period without peer interaction. We find no statistically significant differences in response times by gender, showing that one gender does not respond faster while the clicker station is open. We also find no differences in response time by gender for different cognitive levels.
The fractional difference in response times between the first time a question is posed and after peer interaction. Negative values show that less time is taken to answer the question with peer interaction. The fraction difference in final response times is given in the top panel while the initial response time differences are given in the bottom panel. We find that both genders respond quicker to a question during peer interaction; however, there are no differences between genders. The error in each bin assumes Poisson statistics.
The fraction of changed responses in a single response period without peer discussion. Men are slightly more likely to change their answer in lower-level questions and for the entire question set. We assume Poisson statistics for the error in each case.
The fraction of responses that changed from incorrect to correct, and vice versa, after peer instruction. This represents a change in the response given with and without peer discussion. The benefits of peer instruction are clear: more students switch to a correct answer after they allowed to discuss the question with their neighbor. There are no statistically significant differences between genders. The error assumes Poisson statistics.
Article metrics loading...
Full text loading...