A hands-on, interactive method of teaching statistics to agricultural researchers

Abstract

In this paper, I focus on two topics. The first topic is several day-long workshops, which I run annually for agricultural researchers. These cover the “linear model” methods, based upon the normal distribution, which are very commonly used by agricultural researchers. These methods are analysis of variance, regression and analysis of covariance. I describe the manner in which these workshops are run, cover the content of the workshops, describe the course evaluations, speculate upon why the workshops have proven to be popular, and attempt to draw conclusions. The second topic is a method of teaching linear model theory using N-dimensional geometry. This method has been successfully used for both second-year and graduate- level university statistics courses.

Learning from experience

“…Even on in-house courses those attending could be a very heterogeneous group as regards their needs and backgrounds. Successful on-site training programmes, which have been developed over several years, are described by Horgan et al (1999) and by Saville (2001) together with the thinking behind the programmes and an assessment of their effectiveness. Although the programmes described are for biological researchers (Horgan, 1999) and agricultural researchers (Saville, 2001), much of what is said is of relevance for courses for researchers in other disciplines.…”

“…Perhaps the most important condition for a consulting session to be successful is that the statistician and the researcher can communicate with one another (Belli, 2001;Saville, 2001). Notation, abbreviations, vocabulary and jargon can be a barrier in communication, something which both the statistician and the researcher will need to bear in mind, and to remind one another that they do not always speak the same language.…”

“…By consulting with statisticians, researchers learn when to consult and what questions to ask, as well as learning about statistical concepts, design and analysis (Belli, 2001;Saville, 2001;Svensson, 2001). It is sometimes easier for the consultant to suggest a design for a study and to analyse the data than to tell the researcher how to do this, but by helping the researcher to do these things him or herself the consultant is giving the researcher statistical training.…”

How much can be taught about stochastic processes and to whom?

“…Another way is to seek help from experienced researchers. It will be nice if it is possible to come up with materials that are like Saville's (2001) workshops for agricultural researchers. In the mean time, we have to make do with textbooks and the researchers who want to learn stochastics processes as our guides.…”