I am trying to determine the best way to randomize my patients for a study I am conducting. There are 2 experimental groups and a control group. I work in a mental health setting so being aware of their diagnoses is critical as I dont want too many of 1 type of diagnosis in 1 group and another in another giving skewed results. I am trying to determine if the best way to do this is through stratifing my patients or doing a cluster randomization or strictly doing a single randomization and doing an anova on each group to make sure that they are equally balanced. Something else to consider is that I work in a 17-bed unit so my n for each group on a weekly basis will always be small. I plan on conducting my study over the course of 6mos weekly to increase my overall N but want to make sure that I am able to divide up my patients as evenly as possible. Thank you in advance for your thoughts!