Kirsch, Andrew R. (1993, ASQC) 3M; St. Paul, MN
Use the scientific method to answer questions about the links between causes and effects. Barriers to demonstrating a cause-effect link include: a failure to see causes other than the one being studied; being imprecise in asking questions about the link; and using test methods that are convenient rather than appropriate. Remember also that some effects may be too small to see, and that some have multiple causes. An unintentional barrier is the bias we may have toward the paradigms being tested. Successful answers to cause and effect questions may start with the gathering of anecdotal data. These data are interesting and help us think of new possibilities, but they prove nothing. Test the anecdotal data by designing a controlled study or experiment. Confirmation of the results comes from replicating the study at different times. This allows at least some generalization of the results. Recommendations for making this process work are: be your own worst critic; be sure that the process or product being studied is stable in time; do not fear making mistakes; and be aware of where you are in the question answering process.
Data analysis,Design of experiments (DOE),Reliability