Compared with other types of data, reaction time data are distinctive in that they show unique patterns of distribution, central tendency, and variability. Hence, the statistical analysis of these data calls for special techniques and knowledge. We believe that the mixed-effects model using R provides a good solution to the common problems found in RT data analysis, such as positive skewness, strong temporal auto-dependency, and extreme outliers. This article first reviews traditional methods used for the statistical analyses of RT data. Then, using examples from real studies, it introduces some useful principles and practices for employing the mixed-effects model to fit the data and obtain the model of best fit.