Detection or isolation of defects? An experimental comparison of unit testing and code inspection

Code inspections and white-box testing have both been used for unit testing. One is a static analysis technique, the other, a dynamic one, since it is based on executing test cases. Naturally, the question arises whether one is superior to the other, or, whether either technique is better suited to detect or isolate certain types of defects. We investigated this question with an experiment with a focus on detection of the defects (failures) and isolation of the underlying sources of the defects (faults). The results indicate that there exist significant differences for some of the effects of using code inspection versus testing. White-box testing is more effective, i.e. detects significantly more defects while inspection isolates the underlying source of a larger share of the defects detected. Testers spend significantly more time, hence the difference in efficiency is smaller, and is not statistically significant. The two techniques are also shown to detect and identify different defects, hence motivating the use of a combination of methods.