In the analysis of salivary cortisol data, researchers often perform statistical analysis for hypothesis testing in the absence of data mining procedures. In this article, I demonstrate the utility of screening data from a study investigating the effects of acute stress on salivary cortisol reactivity through the application of procedures recommended by Tabachnick and Fidell (2001). Specifically, an examination for the presence of both univariate and multivariate outliers (Study 1) and methods for correcting skewed distributions (Study 2) were used in order to demonstrate the efficacy of screening data prior to hypothesis testing. The results suggest that there were no outliers present in the data set. Application of algorithms from a family of transformations showed that they were effective in reducing skewness, kurtosis and variability.