Pitfalls in Scientific Research, concl.page 2
Controlled experiments rest on the assumption that each treatment component has only an additive effect, meaning that none of the components interact to result in a second or multiple other effects that are interpreted as all resulting from the single added component. “For instance, psychological and physical effects may interact with one another. An optimistic outlook may enhance the efficacy of a physical effect, and a physical effect may buoy a patient’s optimism that a treatment is in fact working. This multiplicative relationship between treatment components would tend to undercut the ability of a trial to focus on particular components in isolation” (Weijer, p. 69).
The interaction of the doctor or researcher with a sick person or test subject affects how the individual reports his initial condition. The doctor/researcher can lead the patient/subject to provide answers that will yield approving nods, though the answers may not necessarily be true. Doctors attempt to convince their patients to take one treatment over others. Test subjects know that researchers attempt to fool them by asking them to concentrate on one feature in the test while the researcher evaluates some other component of the subject’s behavior.
The self-report post-experiment survey biases and guides a subject’s responses. When a placebo recipient check marks that he experienced body aches, for instance, after administration of “the medicine,” he suspects that the researchers are seeking to determine ache levels, which he rationalizes must be a possible side effect. Thinking that he can contribute to the creation of a safer drug, he affirms he had aches, though he doesn’t know that he was in the placebo control group. This kind of suggestive questioning on post-experiment surveys can lead subjects to report exacerbated disease effects, such as stronger asthma symptoms. Adverse effects of a drug also are determined by self-reports where such effects aren’t easily testable or evident, including mild-to-moderate headaches, chills, nausea, pain and fever, all of which a subject can falsely identify as having experienced.
If subjects in a study suspect, based only on their interpretation of the researcher’s behavior, a sly grin, for instance, that they have been placed in a placebo control group, they will have a preconceived notion of their lousy fortune, and behave correspondingly. Participants who think they are assigned to a group that will be receiving a new treatment will experience, generally, either favorable expectations or increased apprehension of possible, as-yet unascertained negative effects. “Those assigned to standard treatment, however, might feel deprived or relieved. Despite evidence to suggest that new treatments are as likely to be worse as they are to be better than standard treatments, participants probably assume that new treatments will be better than standard treatments—new means improved. In any case, knowledge of the intervention received, and perceptions of that treatment, can affect the psychological or physical responses of the participants” (Schulz and Grimes, p. 696).
Researchers can’t know whether an effect in the placebo control group is the result of the subject’s response to an expectation of a cure and not an effect of natural healing or remission, or lying by the subject, or any of the other things.