How Gender, Team Size, and Commit History can Predict the Effectiveness of Software Engineering Teamwork in an Educational Setting
This project concerns itself with the relationship between the effectiveness of software engineering teamwork with other relevant factors such as gender, team size, and commit history. A variety of methods are utilized including calculations of summary statistics, chi-square and two-sample proportions tests, logistic regression, cross validation, stepwise selection, and variables recoding.
Overall, results indicate that teams with at least one female member tend to better demonstrate best software engineering practices while the products of all-male teams tend to be rated higher. Additionally, for process grade, this project finds that teams of 5-7 students are more likely to receive an A than teams with 3-4 students. Last but not least, for process grade, an increase in the average number of commits is associated with a decrease in the odds of receiving an F.
- seteamwork_report.pdf: a full report including motivation, data, methods, results, and discussion of the project.
- process-design.xlsx, process-implementation.xlsx, product-design.xlsx, product-implementation.xlsx: data files
- model-3.Rmd: the cleaning and exploratory analysis of data and modeling process
San Francisco State University. (2016). Data for Software Engineering Teamwork Assessment in Education Setting Data Set [Data files]. Retrieved from https://archive.ics.uci.edu/ml/datasets/Data+for+Software+Engineering+Teamwork+Assessment+in+Education+Setting#