Although not the gold standard of clinical research, observational studies can play a central role as the nation's health care system embraces comparative effectiveness research. Investigators generally prefer randomized trials to observational studies because the former are less subject to bias. Randomized studies, however, often don't represent real-world patient populations, while observational studies can offer quicker results and the opportunity to investigate large numbers of interventions and outcomes among diverse populations--sometimes at lower costs. But some decisions based on observational studies have turned out to be wrong. We recommend that researchers adopt a "body of evidence" approach that includes both randomized and observational evidence.