MATH 148 Chapter Notes - Chapter 1: Cirrhosis, Scientific Control, Treatment And Control Groups
11 Sep 2016
School
Department
Course
Professor
Document Summary
Should assign subjects into treatment or control group at random. Double-blind neither subjects nor doctors know who is in which group. Treatment and control groups should be as similar as possible. Any difference in response between groups is because of treatment, not another factor. If a factor besides the presence or lack of treatment exists between groups, effect of this factor may be confounded with effect of treatment. Better to rely on chance (ie: tossing a coin has no bias) Laws of chance guarantee that with enough subjects, treatment and control groups will resemble each other very closely with respect to most important variables. Randomized controlled when an impartial (unbiased) chance procedure is used to assign subjects to treatment or control. Important basic precaution: give the control group a placebo. Study done to show effectiveness of surgery for cirrhosis of liver; redirect blood flow through portacaval shunt. 32 studies without controls - 75% enthusiastic about shunt.
