Experiences from Teaching PSP for Freshmen

Author(s): Per Runeson
Venue: Proceedings of the 14th Conference on Software Engineering Education and Training
Date: 2001

Sample Size: 162
Class/Experience Level: Undergraduate Student, Graduate Student
Participant Selection: class participation
Data Collection Method: Project Artifact(s)


Link: http://portal.acm.org/citation.cfm?id=794193.794906
Type of Experiment: Experience Report

Results in teaching the PSP to freshmen are compared to teaching it to graduate students. There are 131 graduate students in this study and 31 freshmen. The data collected is from the students’ PSP data logs. The programs that are written are the standard PSP course programs. One interesting finding is that the freshmen tending to write smaller (less lines of code) programs than the graduate students. The authors hypothesize that this might be because the freshmen’s programs are less sophisticated with less error checking. The study also finds that freshmen spend significantly more time (average 46.8% more time) than the graduate students in writing each program. Interestingly the defect count is essentially the same between the two groups. The authors hinted that it’s quite possible that the freshmen didn’t collect defect data correctly. Other significant findings are that freshmen improve productivity between PSP0 and PSP1, while graduate students improve productivity between PSP1 and PSP2. The data quality of the PSP logs was analyzed, and it was found that the freshmen’s data was often twice as bad as the graduate students’ data.

A point of interest is that the freshmen were much more accepting of the PSP process, and were willing to see it as part of the software development process. The graduate students weren’t so willing. The authors hypothesize that the best time to teach the PSP is during a students second year.

I thought that the study’s approach in exploring when the training works best was novel. The data collected seemed to be relevant and the analysis was thoughtful. There were any strong claims made, but any made were backed up.