I'll be honest. I'm a bit scared to be writing a blog about statistics that the whole world can see. The typical image of a statistician is of an egghead that spouts off numbers, equations, and conclusions in an erudite and authoritarian way that easily intimidates the audience. I don't think I can do that facilely...yet. I understand the equations in my textbooks but explaining them does not role off my tongue easily.
But more than that, I don't want to walk in the land of intimidation. Statistics and the insights that statistics offer shouldn't intimidate. They should inform and they should inform in a helpful manner. Certainly, detailed explanations using Greek letters and equations are needed to establish the legitimacy of any given statistical method. Yet, the vast majority of people who could be interested in what statistical insight can offer are more interested in the fact that two groups of people show different interests or activity rather than knowing that the discriminant analysis or logistic regression has produced a significant Mahalanobis distance or log odds ratio.
What is more important is that interested parties know that you looked at the differences between two groups and that you can be reasonably certain those two groups respond differently to the advertisement or treatment. Knowing that the difference is "real" enables more confident and effective decisions. Such reassurance goes a lot further than intimidation ever will.
No comments:
Post a Comment