This is the process of designing scientific experiments to minimize bias and maximize the statistical power of the results.