This is a
fundamental flaw inherent in the social and
natural sciences. It is the process of, after having developed a
hypothesis, interpreting every piece of
experimental evidence as support for that hypothesis.
It is best explained using a true example. In the 1980's, a group of investigators led by Dr. David Rosenham, a psychology and law professor at Stanford University, admitted themselves as patients to various mental hospitals. The investigators did not act as though they were mentally ill; they gave honest, "normal" responses when questioned. When interviewed, the investigators would take notes. Psychologists recorded, "patient engages in writing behavior." In another interview, a "patient" told doctors that, although he was close to his mother as a child, he had grown closer to his father as he grew older. This was recorded as "unstable relationships in childhood." The point is, once the doctors "realized" they were looking at a mentally ill person, they interpreted all types of natural behavior as being evidence of an illness. At no point did a doctor stop and think that perhaps he wasn't dealing with a neurotic.
A similar process takes place when one explains the motives of others. Said to be the greatest flaw of such moral theories as ethical egoism, it is also present in the natural sciences as well, though not on such an extreme level. In the laboratory, it is often difficult not to interpret results as evidence supporting one's own
beliefs.
A teacher explained this by comparing this to a Play Doh Fun Factory. Whatever goes into the Factory (or one's mind) will come out exactly as the mold placed in front of it. Similarly, a pitfall many fall into when investigating and experimenting is to take any piece of information and mold it to suit one's view.