The Department of Defense (DoD) cost estimating methodology traditionally focuses on parametric estimating using ordinary least squares (OLS) regression. Given the recent advances in acquisition data collection, however, senior leaders have expressed an interest in incorporating "data mining" and "more innovative analyses" within cost estimating. Thus, the goal of this research is to investigate nonparametric data mining techniques and their application to DoD cost estimating. Using a meta-analysis of 14 cost estimating studies containing 32 datasets that predominantly relate to commercial software development, the predictive accuracy of OLS regression is measured against three nonparametric data mining techniques. The meta-analysis results indicate that, on average, the nonparametric techniques outperform OLS regression for cost estimating. Follow-on data mining research that incorporates DoD-specific acquisition cost data is recommended to extend this article's findings.