Personal Software Process - Classroom Experiences from Finland

Author(s): Pekka Abrahamsson and Karlheinz Kautz
Venue: 7th European Conference on Software Quality
Date: 2002

Sample Size: 20
Class/Experience Level: Undergraduate Student
Participant Selection: class participation
Data Collection Method: Project Artifact(s)


Type of Experiment: Experience Report

The study is another of the classic PSP type that occurs in an academic setting. The estimation skills and defect injection rates of the students are measured over the course of the PSP class. The expectation is to see improvement as the class progresses and the subsequent steps of the PSP are introduced. In this case 31 students enrolled in the class but only 20 successfully complete the course. The authors recognize that the collected data may be error prone.
The results show that there wasn’t the expected improvement in students’ ability to either estimate program size or required programming effort (time). One measurement that did correspond with the authors’ expectation was the drop in the number of defects found in the test phase. This drop is expected because the PSP emphasizes early detection of defects.
The students reported that overall they didn’t feel the process would be something that they would take to industry. The study also found that some of the students didn’t have enough knowledge with regards to statistics to be able to appreciate the PSP metrics.
This isn’t a particularly strong empirical paper, similar to many of the academic PSP experience reports. Perhaps most interesting it that the results don’t reflect the successes that other academic based PSP experience reports display.