Scientists don't count on their expectations and observations matching exactly because so many different factors can affect observations and test results. In fact, observations thatÂ tooclosely and consistently match expectations are sometimes taken as a warning sign of bias or fraud. For example, in the 1800s, Gregor Mendel performed many, now legendary experiments involving crosses of pea plants in order to learn more about inheritance. He was the first to notice the distinct hereditary patterns that were eventually recognized as the foundations of modern genetics. The only problem? Mendel's data seem to beÂ too good for too many of his experiments.
In the 1930s, a prominent statistician pointed out that many of Mendel's results matched his expectations surprisingly closely. For example, when Mendel expected a ratio of 3:1, he got ratios of 2.96:1 and 3.01:1. These observations (and supporting statistical analyses) led many scientists to wonder whether Mendel had "fudged" his data. After 70 years of debate and investigation into Mendel's scientific ethics, modern scientists and historians can find no evidence that Mendel intentionally committed fraud. Nevertheless, because of the unusually close match between his expectations and observations, scientists continue to wonder exactly how his results wound up biased in support of his ideas."
The question of scientific and experimental ethics is one that doesn't get discussed in amateur (or perhaps also professional) circles as often as it might. For an excellent introduction to this subject, we recommend that you visit the Online Ethics Center for Engineering and Research, where you can find a series of excellent case studies that illustrate the problems associated with data fabrication and falsification.