Despite the growing popularity of test-driven development (TDD), there is no empirical confirmation of the benefits that this contemporary practice confers on its users. Prior research findings on its efficacy have largely been inconclusive. We conducted a laboratory experiment to assess the impact of TDD on software quality and task satisfaction. Additionally, we investigated the productivity aspect of TDD as compared to the traditional test-last method of software development. Results indicate that software quality and task satisfaction are significantly improved when TDD is used. Despite the additional requirements of testing, TDD is not more resource intensive than the test-last method. We also examined TDD’s impact on learning post hoc and discuss the implications of our findings and directions for future research.
Bhadauria, Vikram S.; Mahapatra, Radha K.; and Nerur, Sridhar P.
"Performance Outcomes of Test-Driven Development: An Experimental Investigation,"
Journal of the Association for Information Systems: Vol. 21
, Article 2.
Available at: https://aisel.aisnet.org/jais/vol21/iss4/2