Double blind: double talk or are there ways to do better research

In an attempt to counteract placebo effects, scientists studying the effectiveness of medication often use the 'double blind' trial. Unfortunately, in less than 5% of the studies reported in major medical and psychiatric journals was there a check to determine whether patients and observers were blind. Between 1972 and 1983 there was increasing reliance on active placebos which attempt to mimic the effects of the medication. It is unlikely however, that patients are not aware of significantly different physiologic changes within them resulting from the different chemicals. Whenever a physician informs a subject that they may be on placebo or active treatment medication, they will spend much time in trying to guess when they are taking which. In these situations Philip's paradox will apply. This states that the more potent a therapeutic variable the less likely its efficacy can be 'proven' in a double-blind study. One can only 'prove' that medication is no more effective than placebo. To have a truly blind procedure, the active placebo must have identical physiological effects to those of the medication being studied. We are now using the patient's and the observer's awareness of changes as useful information to obtain: a guess factor, the percentage of correct guesses whether the patient is on medication or placebo; a direction factor, whether the patient considers the change beneficial or detrimental; an attributional factor, whether the change is considered to result from medication or other intervening variables. This information is usually reliable and can be analyzed with the usual statistical procedures.