All research has flaws. Some flaws are so trivial that the research can still stand as the definitive study. Other flaws prevent a study from being definitive, but the study still provides useful guidance in the context of other research. Some flaws are so serious that the research provides no useful information at all. The tricky part is not finding flaws in the research but in deciding to what extent the flaws erode the credibility of the research. In general, the use of RCTs can add substantial credibility to a research study. There are calls for greater use of RCTs in many areas, such as surgery (Baum, 1999) and psychiatry (Andrews, 1999). Of course, nonrandomized trials are an important complement to RCTs when the latter are ethically inappropriate or logistically impossible (Black, 1996). Failure to use randomization or blinding, however, is not a fatal flaw. Furthermore, the artificial nature of RCTs will often restrict their applicability to overly simple interventions. When RCTs focus on narrow patient groups or exclude important segments of the population, there may be difficulty in generalizing their results. So it would be a mistake to label the RCT as a gold standard for all research. A silver standard may be a more appropriate label.