The experimental units are what you apply the treatments to. In this experiment it would be the subjects or the customers in this situation since they will be treatment (either standing at their regular height or squatting to meet the eye level which could affect the amount of tip they give). An example of randomization can be numbering each table in the restaurant and picking them out of a hat and going to that table that you randomly chose and imposing the treatment on the customers there. Because tips depend on the income of the customers and how much they are willing to give for their meal. If not even income some people are willing to give despite how much money they make so their tips may higher than the wealthy crowd. Other factors the experimenters may not be able to control is the mood of the customers as well isn't it more likely if someone is enjoying their time to give more to the server despite the service the waiter.
