I taught biostatistics (in public health) for almost 20 years. The important element is that the students had widely varying backgrounds and training. Many of them were clinicians: this means they were trained and practiced to deal with individual patients; they were often completely naive concept of a population or of sampling. Statistical ideas were totally foreign to many as was the notion of a formula. 

During that time, I substantially changed my expectations with respect to statistical computing, changing in ways that might not suit everyone.

I decided that statistical computing should have a small but necessary and useful role in such a course for such an audience. No one was going to become an expert; the computer had to serve a need to analyze a problem with data; the focus should be on problems with conceptual importance, not calculation per se.

I quit asking for calculations on any evaluations: RATs, exams, etc. I had a two hour session on computing in the laboratory where I gave some specific problems using data sets that I had prepared. After that, it was up to teams to do computing as part of doing problems in or out of class. Many of the problems had small data sets that were in the text or an associated CD.

The calculations were always adjuncts to answering a series of related, substantive questions. The teams were assigned 2-3 problems to work on in class and report on. We were usually able to have 5-6 team presentations each week.

Sometimes I had one 3 hour session a week, sometimes two 1 1/2 hour sessions. In the latter case, teams started problems before the end of the first session and reported their answers at the beginning of the second session. (A very few topics used one session or three sessions.)

I decided a long time ago that testing or examining computational skills was a big mistake. All the statistical issues are substantive, not arithmetical. I knew this from the start, but I was also swimming upstream against culture. This is a good way to be declared insane, of course. If some students are highly skilled at computing, or wish to become so, this is a good thing, but this is difficult to teach to everyone and something that people will learn when it becomes necessary.

We have to decide what we think is important. I don't think arithmetic is.

Regards,

David Smith






On Thu, Jan 16, 2014 at 1:36 PM, Van Patten, Isaac T <[log in to unmask]> wrote:

I am teaching a course in quantitative methods to first year graduate students and introduced team-based learning to this course for the first time last spring.  For the most part it went well with one exception, Application Exercises.

 

Here is my dilemma.  The subject matter certainly provides ample material for application exercises.  Working with real-world data, students conduct real-world analyses.  The problem has to do with the team-based aspect of the activity.  The applications readily lend themselves to individual effort.  However, figuring out how to best meet the single answer, simultaneous reporting criteria is difficult.

 

What I observed was usually the one student on the team who had best statistical knowledge would sit at their workstation and do all the work.  The rest of the team watched over their shoulder.  Obviously this much social loafing is counter-productive.  As a team, they were more focused on "getting the right answer" than on learning the statistical concepts behind the analyses.  I encouraged everyone to work the problem and to help one another out.  Still, the students would quickly gravitate to the one workstation where the student already "got it."

 

My question is how to use team-based applications in this situation to greater effect.  I am striving for that mystical "cognitive

 

__________________________________________

Isaac T. Van Patten, Ph.D.

Professor of Criminal Justice

307 Adams Street, Office 1A

Box 6934, Radford University

Radford, VA 24142

(540)831-6737

[log in to unmask]

http://ivanpatt.asp.radford.edu

It doesn't matter how beautiful your theory is;

It doesn't matter how smart you are;

If it doesn't agree with the [data], its wrong.

-Richard Feynman

 

 




--
David W. Smith, Ph.D., MPH
Chartered Statistician