Understanding Statistical Power

In Episode 28 of Med School Question of the Week for USMLE, Alisa Khomutova, MedSchoolCoach expert tutor, answers this medical school question: A group of investigators is evaluating the efficacy of intranasal ketamine in decreasing acute suicidality in teenagers admitted to the ED. They conducted a study in which 25 participants were randomized to receive intranasal ketamine while 25 participants were randomized to receive placebo. The investigators ultimately found no statistically significant difference i suicidality after 1 month of follow up with the level of significance fixed at 5%. They suspect inadequate statistical power. Assuming the investigators intend to keep the level of significance at 5%, which of the following changes would be most effective to increase the statistical power?

• Decrease sample size
• Increase type 1 error
• Increase alpha
• Decrease type 1 error
• Decrease alpha
• Increase beta
• Decrease type 2 error