The study investigated how the order of presenting test code versus production code during code reviews influences the effectiveness of the review. It also examined developers' perceptions of test-driven code reviews.
For RQ1, 92 developers participated in a code review experiment comparing reviewing test code first versus production code first. Reviewing test code first led to finding more bugs in the test code but did not influence bugs found in production code.
For RQ2, interviews and a survey of 103 developers found that while some saw advantages to reviewing test code first, issues like low test code quality and the inability to choose review order were perceived problems. Adoption of test-driven reviews varied, with most using them sometimes.