The randomized controlled trial (RCT) is generally viewed as the gold standard for health science research.1 However, an implication of random treatment assignment is that there will occasionally be substantial discrepancies in distributions of background characteristics across treatment arms. When these background characteristics are related to research outcomes, imbalances in the distribution of characteristics across treatment arms can mislead investigators and can induce erroneous inferences about connections between treatments and outcomes. Such imbalances can induce spurious association or can mask actual association. The analysis of covariance (ANCOVA) is a technique that can be used to help answer scientific questions that arise in this context. (July 2009)
Psychiatric Annals
2009
Belin TR, Normand S-LT
http://www.healio.com/psychiatry/journals/PsycAnn/%7BCC73E9F6-DF77-418E-8EED-255F1D27369B%7D/The-Role-of-ANCOVA-in-Analyzing-Experimental-Data