ABSTRACT
The major interest in intervention trials is often the estimation of intervention effects for individuals who actually receive the intervention. However, some percentage of noncompliance is usually unavoidable in intervention trials when dealing with human participants. In addition, it is not easy to control compliance behavior of individuals who may decide not to participate even with highly attractive incentives. Noncompliance is a major threat to obtaining power to detect intervention effects (Jo, 2002), and may bias the estimation of intervention effects if not handled carefully in the statistical analysis.
