Jeff,
I agree with several of the previous posts regarding the integration of stats and
methods.In many ways, I think it is a more important question than the order of the
two courses. I've had the opportunity to teach stats and methods independently as well
as integrated. My conclusion is that the latter is far superior. I am often puzzled
why more departments don't adopt this model. In fact, there are very few integrated
books.
Here are some thoughts to ponder...
1. The separation between stats and methods is, in many cases, an artificial one. In
practice, we generally use them together. To cover stats separately requires some
discussion of methods and vice versa. Why not make it more efficient and do it that
way from the start?
2. The separation may make an already difficult topic even more difficult. By this I
mean, in an integrated course, you can more easily follow the entire progression from
research question to conclusion. This is likely to give the student context and help
them see the relevance of the design and analysis tools. Anyone who has taught stats
knows of the "relevance" problem.
3. I also think that the integration promotes transfer to future research
experiences.
Well, I guess I've said enough. I'm curious to hear what others think.
Mark