|
Abstract:
|
We as consulting statisticians are often asked to review material and give our realistic view of the data we see. But many times, the clients wants to hear that the experiment they have done has shown the results they expected; the analyses they have run have nice small p-values where they should be; the study they have designed is flawless; the paper they have written unambiguously demonstrates the thesis they have developed; the company they want to buy is likely to be hugely profitable. We look at the evidence and see, as Gertrude Stein said about Oakland, "no there there". A "kind" assessment is intellectually dishonest but may make the client happy. A brutally honest assessment can dash their dreams in a way that is destructive to their self-image and their future. Some statisticians avoid expressing their own negative views by blaming others. In my world, people who see weak or biased data may say, "The regulators won't like this" instead of, "Your data are not convincing." At this round table we will discuss case histories of giving negative news in a style makes the clients really understand you are on their side.
|
