A Cross-Program Investigation of Students’ Perceptions of Agile Methods

Author(s): G. Melnik and F. Maurer
Venue: Proc. 27th Int’l Conf. on Software Eng. (ICSE)
Date: 2005


Quality
3

Problem
The IEEE and ACM recommend agile concepts and practices as essential topics for schools to teach in Computer Science. Among these topics is Test-Driven Development. Melnik and Maurer wanted to explore three aspects of this. The aspects were to see the perceptions of agile practices from student’s perspectives, the perceptions of agile methods in general and individual practice of these methods, and lastly how these perceptions varied between academic age and industry experience.

Study
The University of Calgary and Southern Alberta Institute of Technology have been teaching agile methods in their curriculum since 2001. In this three-year time frame, they conducted a wide range of surveys from two hundred forty volunteers. The volunteers are spread out between five levels of students in their Computer Science curriculum: Freshman, Sophomore, Junior, Senior, Graduate. After completing coursework students were asked to take surveys and the results were analyzed.

Results
After putting together all the results from the surveys, it turned out that 78% of the volunteers agree or strongly agree that test-driven development (TDD) improved productivity of small-sized teams. Similarly, 76% agreed or strongly agreed that TDD improved the quality of the code. 65% would recommend agile development methods including TDD to businesses. Other agile methods were taken into consideration as well, and there was statistical significance in the fact that students enjoyed pair-programming more. Lastly, there was a weak positive correlation, yet not statistically significant, between the age of students and the attitude towards TDD most likely explained by higher level of discipline and greater maturity of students in the upper division courses of Computer Science.

Conclusions
The studies showed that students preferred agile methods such as pair-programming and test-driven development. Furthermore, there is no correlation between the students age/years of experience and the quality of code/productivity in small-sized teams. Melnik and Maurer summarize that students are very enthusiastic about core development processes and there is no significant differences in perceptions of carious levels of educational programs. Also, agile-methods promote development of professional skills through team-work which is highly valuable in the work-force. One known limitation of this study is that 693 students were asked to fill out the survey and only 240 responded (35%). This percent of feedback could have skewed the results since those who didn’t respond could have not cared for agile-methods, therefore biasing the gathered results.

0