Page 32 - The Real Work Of Data Science Turning Data Into Information, Better Decisions, And Stronger Organizations by Ron S. Kenett, Thomas C. Redman (z-lib.org)_Neat
P. 32

4







             Understand the Real Problem






             A Telling Example

             One day, one of us (Redman) was approached by a middle manager seeking a sample size
             calculation. Something about performance in his group bothered him and he wanted to do a
             study to learn more. He needed extra budget to do the study, and he needed the sample size to
             justify that request. He confessed that he had visited another statistician earlier. The meeting
             had gone reasonably well, but that statistician had given him a formula, which he didn’t really
             understand. He just wanted “the number.”
               Seem straightforward enough?  The manager has explained the context (i.e. his budget
             request), his specific question (i.e. sample size), and his level of sophistication (i.e. can’t use
             the formula). Plenty on which to proceed.
               But Redman probed further:

                • What was bothering him?
                • What sort of study did he plan to conduct?
                • How would he analyze the data once he collected it?
               A wide‐open exchange followed, covering these topics and more. It emerged that the man-
             ager did not know enough to design a full‐blown study. Instead, he should “turn over some
             rocks to see what crawls out.”  The two discussed a semistructured method for doing so. They
                                     1
             also agreed that this method might not work, so they should talk frequently.
               The discussion on how many rocks to turn over went like this:


               Manager:  “So back to our original question. What’s my sample size?”
               Redman:    “If I told you 50, could you do that without additional resources in the
                          next two weeks?
               Manager:  “Hmmm. Probably not.”
               Redman:    “OK, what if I told you 25?”

             1   George Box (1980) noted “that if you don’t know anything, no experiment could be designed.” We interpret that
             observation as “turn over some rocks to see what crawls out” as the first step in remedying the situation.

             The Real Work of Data Science: Turning Data into Information, Better Decisions, and Stronger Organizations,
             First Edition. Ron S. Kenett and Thomas C. Redman.
             © 2019 Ron S. Kenett and Thomas C. Redman. Published 2019 by John Wiley & Sons Ltd.
             Companion website: www.wiley.com/go/kenett-redman/datascience
   27   28   29   30   31   32   33   34   35   36   37