Science as a way of knowing  (Page 6/8)

The other important part of this experimental approach is the concept of control experiments , also known by a shorter term as just controls . A control experiment is a setup where the variable is not introduced, so that it can be directly compared to the experimental situation where the variable (access of dust particles to the broth) is included. So a control experiment for Pasteur's incubation of broth in an open-necked flask would be incubation of broth in the S-necked flasks. If the variable is introduced by tilting the flasks, the control would again be the S-necked flasks. All other conditions (heating temperature, amount of broth, size of flasks, etc.) were the same in the experimental and control situations. The only thing that was different was a single variable  (access of dust particles to the meat broth), because that was the hypothesized source of the living cells that grow in meat broth left out in the open. A single control experiment is usually all that is needed if there is only a single experimental variable being manipulated.

But it is not always possible to simplify a system so that there is only one variable. In those cases, as you will learn in the studio exercises for this class, you might need multiple control experiments. Experiments and controls will also be repeated before the investigator reports the results. It will be described in a way such that other investigators can readily repeat it as well. In some situations the results will be subjected to statistical analysis, although this was not necessary in Pasteur’s experiment. Statistical analysis is critical in many scientific approaches, particularly in studies involving hypotheses about human subjects (e.g., the hypothesis that smoking causes lung cancer), where experimental manipulation of the subjects is difficult or impossible. A scientific experiment, no matter how the results are analyzed, should lead to a conclusion that either supports, or fails to support, the hypothesis. Finally, the experimental results should lead to additional hypotheses, and additional predictions, that can generate further support (or lack of support) for the hypothesis. Try to think of a few additional experiments that you might have suggested to Louis Pasteur if you were alive in 1860, and if you could speak French!

Other aspects of science

The characteristics inherent in the scientific process lead to another property of science, and that is that science is self-correcting . By that we mean that errors can be made, but that continued application of the tools and processes of science will usually lead to elimination of the errors and a more accurate understanding of the natural world. Science never really and finally proves anything to be true; it can, however, prove things to be untrue. To some people, in fact, that characteristic of science, its fluid and changing nature, is maddening. If you require solid ground to stand on, and immutable truth in all aspects of your life, you probably shouldn’t become a scientist. If you find excitement in being part of an enterprise that is constantly changing the extent, and even the nature, of knowledge, then you have some of what it takes to be a scientist. But even if you don’t become a scientist, a bit on scientific knowledge, and a bit of practice with the scientific process, will help you understand the things you need to understand in order to make intelligent decisions about many things, such as health care, climate change, or other concepts that are in your future.