DESIGNING EXPERIMENTS


Unlike a survey, an experiment involves actually doing something to the subjects or objects rather than just soliciting answers to questions or making measurements. For example, instead of asking people whether they think that vitamin C is effective for preventing colds, you might give them vitamin C and observe how many colds they develop. Or you may want to try product A and product B and then compare the results to see which one is better. Sometimes you study the subjects before and after your experimental treatment. Sometimes, instead, you take several groups of subjects, do something different to each of the groups, and then compare the results.

Experimentation on people poses ethical questions that deserve careful thought. Many responsible institutions have committees that regulate experiments involving human subjects. If an experiment exposes a subject to risks, such as possible side effects from a new drug, you must certainly inform the subjects in advance. Usually you must have them sign forms to give their consent . Needless to say, that is not a concern when you test products ” even though the test may be a destructive one.

In experiments as well as surveys, the subjects must come from the population that you are interested in. (As you have probably gathered by now, proper sampling is much easier with animals, processes, or products in a laboratory setting than with people in a survey or products in a real world application.) When you design an experiment, you need to fret about some other things as well. For example, to compare different treatments or techniques, you must make sure that the groups receiving them are as similar as possible. Again, randomness is the key. The best way to make groups similar is to assign subjects or objects to the groups randomly . This procedure does not guarantee that the groups will be exactly the same, but it does increase the likelihood .




Six Sigma and Beyond. Statistics and Probability
Six Sigma and Beyond: Statistics and Probability, Volume III
ISBN: 1574443127
EAN: 2147483647
Year: 2003
Pages: 252

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net