Scientific researchers try to control the conditions surrounding their experiments. These controlled conditions reduce the chance of encountering random variations in their data. Still, even under optimal conditions, experiments can uncover elements that researchers could not predict. These may include fluctuations in wind, temperature differences or unexpected chemical reactions. Researchers perform experiments multiple times and eliminate obviously erroneous findings. For instance, if a researcher reviews 20 results of an identical experiment and notices that in two cases the results were drastically different from the other findings, she can consider those results to be erroneous.
Researchers can review multiple findings and develop a statistical average for their experiment. This serves as a form of experiment confirmation, providing a method for double-checking their results against those of other experiments. A statistical average of numerous results supports a conclusion about the trend in the subject of their results. As an example, if a researcher was testing levels of toxins in a water supply, the results of a single test proves only that the tested water has a certain toxin level; they say nothing about the other water in the supply.
Researchers' work faces scrutiny from various groups, including those who disagree with the findings, those performing alternate tests and other researchers who question the validity of the findings. By performing an experiment multiple times, a researcher can demonstrate data integrity in his findings. Data integrity is the natural defense against those who criticize the researchers' methods. If researchers' scientific techniques are sound, scrutinizers tend to accept their findings.
Researchers collect data from multiple experiments as a way to evaluate additional questions involved with their research. This additional data points to alternate experiments if the cause of data fluctuations cannot be determined. For instance, if a researcher was testing the toxin levels in a water supply and noticed that the results vary significantly in certain geographical areas, she can design a new experiment to determine what is happening in the water supply at those points.