An Experiment - to test the hypothesis You have just heard again, in the experimental and survey research sections, that self-reported answers can be quite biased. To remove this bias, you want to gather as much behavioral data as possible. For example, instead of asking customers to rate their satisfaction with service, you want to use the behavioral variable Tip Amount as the variable to measure satisfaction with customer service. To avoid disrupting regular business as much as possible, you conduct the experiment on a Tuesday and a Wednesday over lunch. You instruct your wait staff about this experiment. They are told that on Tuesday, besides the assigned primary server, all staff members are to help out each other. For example, if one server notices a customer's glass empty, s/he should fill it up. Servers are also asked to make as much eye contact with customers as possible. For example, instead of walking by a table looking elsewhere, they are told to make eye contact to see if the customer needed something, even if the table was not assigned to them. On Wednesday, it was back to normal - each server was responsible for her/his table. Data were collected from 30 customers on Tuesday and 30 on Wednesday. The data you collected were