|
Most academic consulting clients already know some statistics and many understand statistical thinking well. These clients need our technical skill. But others have unfortunate habits of statistical thought. We can enhance collaboration by being alert to those habits.
One unfortunate habit is omitting to plot the data. A client says, "My Y is continuous. Therefore I fit a linear regression."
We can enhance collaborations by emphasizing plots.
A second unfortunate habit is treating statistics as a bag of procedures. A client says, "I want to estimate the treatment effect. Should I calculate t or z?" Or, "I want to learn the relationship between X and Y so I fit a regression. Should I adjust for multiple comparisons and should I use the coefficient from the regression with or without X2?"
We can enhance collaborations by deemphasizing formality.
A third unfortunate habit is the quest for a test. A client says, "Tell me whether the treatment is significant."
We can enhance collaborations by avoiding tests and encouraging estimation.
Many statistics students have these same habits. Our statistics curriculum inculcates these habits, so we should modify our curriculum.
|