In recent years, genome-wide association studies (GWAS) have proven to be valuable in the detection of genetic variants associated with complex diseases. In 2007, the Wellcome Trust Case Control Consortium reported GWAS for seven common human diseases conducted under multiple inheritance models (dominant, recessive as well as additive). In subsequent GWAS, reversion back to methods of genome wide screening under a single additive test has become the norm because the multiple testing introduced by considering all inheritance models inflates Type 1 error rate, yet simple Bonferroni correction is too stringent since the tests are not independent. Considering only an additive model risks loss of power from misspecification of the inheritance model, particularly for single-nucleotide polymorphisms (SNPs) that follow a recessive mode of action. The true model of inheritance is unknown a priori but is expected to differ between SNPs. Recently, So and Sham (2011) presented a robust association test applicable to large scale studies, that considers the maximum of three test statistics under additive, dominant and recessive models (MAX3), while accounting for multiple testing. The method allows for the fitting of covariates and principal components where previously proposed robust tests have not. Under this alternate analytical approach, the disease inheritance model does not need to be assumed, presenting a more powerful association test to identify genetic variants, and increasing the reliability of findings. We applied this methodology to two GWAS previously published using only additive association analysis.