When I use a word . . . Defining bias in research

Interpreting the results of an interventional studyConsider three possible outcomes in an interventional study, investigating, say, a novel medicine. The medicine improves the condition, worsens it, or has no effect. How is each of these outcomes to be explained?In all cases the apparent effect could be real. The drug really does what it appears to do. However, leaving aside the possibility that the positive or negative outcome arose merely by chance, or, in the case of the no-effect result, because the study wasn’t adequately powered to detect a difference, there is always the possibility that a bias or biases of some sort played a part. Perhaps bias enhanced or impaired the response, making it appear that a positive or negative effect had occurred, when in fact no such thing had happened