Northern Prairie Wildlife Research Center
We actually are more concerned about wildlifers who fall in the mid-range of statistical education, not those at the low end. For these individuals, knowing what they do not know may well be more important than knowing what they do know. They may not feel the need to consult a statistician even if they should. These individuals know how to perform certain statistical procedures but may not know enough to distinguish situations in which those procedures are appropriate from those in which they are not. Also, powerful computer software allows virtually anyone to conduct very sophisticated analyses. The analyses may be totally wrong, but they are sophisticated. We are reminded of one research biologist who had a fair bit of statistical coursework, but all he seemed to remember from the training was the paired t test. Accordingly, every time he wanted to conduct an analysis, he twisted and contorted the data so that he could run a paired t test on them. A little learnin' can indeed be a dangerous thing. This line of reasoning leads to a "fitness" curve for wildlifers, based on their statistical education (Figure 1).
|Figure 1. "Fitness" of a research wildlife biologist as related to statistical knowledge. Fitness is interpreted as 1 minus the probability of making a serious statistical blunder.|