Finding the factors leading to defects in software engineering can provide a competitive advantage for companies, project managers, programmers, and others, enabling them to better manage a project. Relying on findings from prior research in software defects prediction models, this study aims to examine the effects of the size, the complexity, and the number of the prerelease defects of a software module on the postrelease defects. Regression analysis of 327 responses from the PROMISE database reveals that such effects do exist under the Eclipse platform, a Java language development environment. This study confirms the presence of the quadratic effect of module size and the significant influence of the complexity and the number of prerelease defects in predicting postrelease defects.