Page 141 - The Six Sigma Project Planner
P. 141

Example of Using Worksheet
                    Refer to the output shown in Figure 22. Now, it would seem a simple matter to
                    conclude that females are shorter than males. And this certainly squares with our
                    perception of the world, making it even easier to jump to that particular conclusion.
                    Indeed, this is the sort of thing that is usually done in the business world. We have a
                    preconceived notion, collect some data that “confirm” our notion, and conclude:
                       1.  We were right all along.
                       2.  The data collection exercise was a waste of time.

                       3.  We can continue “hip-shooting” our way to success.
                    The problem is that this retrospective approach to data analysis is inadequate as a way
                          racial or ethnic groups.  TEAMFLY
                    to confirm hypotheses. Its purpose is to help us develop hypotheses. Here are only a few
                    reasons why we can’t conclude that the data in Figure 22 support the conclusion
                    without further analysis:
                       •  Maybe the males and females were children of different ages.

                       •  Maybe the males and females we’re interested in as customers are from different
                          populations than used in the “happenstance” sample.

                       •  Maybe the males were of one racial or ethnic group and the females were of
                          another.
                       •  Maybe the males and females we’re interested in as customers are from different

                    The list could go on and on. The point is, to confirm the hypotheses suggested by the
                    data, we must conduct another study in which we exercise a degree of control over the
                    sampling and/or the experiment.

                    Figure 23 gives an example of how to document these sorts of results.
                                       Figure 23. Example of Evaluating a Hypothesis

                       Data-Based      Hypothesized      Hypothesis Investigation          Result
                       Observation        Cause
                     Males appear to   Natural difference   Our toys are marketed to   Our samples show that
                     be taller than   between the      American boys and girls aged   there is a statistically
                     females. The data  sexes.         9 to 12. We will randomly   significant difference on
                     were obtained                     sample 200 children in each   average, but the
                     from a public                     year of our target market age   difference is much
                     database.
                                                                                   smaller than the original
                                                       groups (ages 9, 10, 11, and
                                                       12) and determine their     data indicated and
                                                       heights.                    varies by age. See “Age
                                                                                   Ht study.htm” for
                                                                                   details.



                                                                  ®
                                                         Team-Fly
                                                             124
   136   137   138   139   140   141   142   143   144   145   146