A major goal of analytical quality assurance (QA) activities, e.g., inspection
and testing, is detecting defects in software artifacts to increase product
quality and decrease rework effort and cost. Inspection aims at identifying
defects early and traditional testing focuses on test case generation and execution
late in the development process. Combining inspection and test-case generation
to inspection-based testing (UBT-i) can help identifying defects early,
increasing testability by systematically capturing requirements and quality attributes,
and generating most valuable test cases based on inspection results.
This paper reports on a controlled experiment to investigate the temporal behavior
of UBR inspection and inspection-based testing regarding defect detection
performance, i.e., effectiveness, efficiency, and false positives. Main findings of
the study are that there are no significant advantages of UBR and UBT-i regarding
defect detection performance and the temporal behavior of defect detection
delivered contradictory results in two sessions of the study.