In genomewide genetic association studies, prior biological knowledge may help distinguish variation that is truly associated with a quantitative trait from the vast majority of unassociated variation that may be significant in hypothesis testing due to chance. However, formal methods for integrating prior biological knowledge into association studies have only been proposed recently, and their potential utility has not been thoroughly evaluated. Herein, gene set methods from genomewide analysis of gene expression data are adapted for application to genomewide genetic analysis of quantitative traits. The proposed gene set method was tested in simulations with gene sets that included up to 500 total variants, among which up to 20 collectively explained 5% of the variance. In a population of 1,000 individuals, the gene set method was largely more efficient at detecting truly associated variants in these gene sets than a comparably calibrated conventional approach relying on P-values alone. While extremely strong associations remain best identified by conventional methods, the gene set approach may provide a complementary mode of analysis for revealing the full spectrum of genes that influence a quantitative trait.