A new study of randomized clinical trials found significant observer bias toward a more beneficial treatment effect in nonblinded trials when the researcher knew the treatment being given to the participant. The study is published in CMAJ (Canadian Medical Association Journal).
