A researcher varies the amount of caffeine given to three groups of participants (0, 100, and 200 mg) and then observes the anxiety level for each group. The most appropriate statistical test used to determine whether the groups differ is