The first objective of the course is to give the participants an overview of the basic experimental designs available. During the course, the reasoning behind the designs will be emphasized, in order to provide the participants a sound base to choose the design for their own research problem. The designs will be demonstrated by computer software.
Experimental designs are typically described by statistical terms only. Eventually, the results of a statistical analysis have to be interpreted in non-statistical terms. The second objective of the course is to give the students an understanding of the logic of argumentation in empirical research. This takes place by an conceptual analysis of their research problems from general methodological point of view.
The third objective is to create an intellectual environment where specialists in statistical experimental design, general methodology and specific sub-disciplines in agricultural field meet. In this context, the students will have an opportunity to outline their own research problems both to statistical and methodological expertise.
Basic statistical courses, including basic ANOVA. Some familiarity with statistical software, preferably SAS. At least a preliminary formulated research problem for the thesis.
Within the fields of agricultural research, a variety of experimental designs have been developed. They provide a vast sortiment of means to test hypotheses and handle sources of variation. Many of these designs are well known within the agronomical research, but less familiar in other fields.
With the help of easy-to-use software, extensive use of statistical analysis of data takes place. Publishing an empirical research without statistical analysis is practically impossible. Statistical analysis, however, often seems to be more a ritual than severe attempt of argumentation. To understand the reasoning behind the designs is very important in increasing the level of empirical research.
Experimental designs are typically described by statistical terms only. This neglects the other relevant part: the design must be flexible with respect to the requirements of the research object under study. Ideally, each component in the analysis has an interpretation in terms of the substance under study. It has been said that an essential amount of statistical consulting work is to make the researchers clarify their research problem that statistical advice can be given.
Successful experimentation is not possible without clearly formulated research problem. Only it renders it possible to use the existing statistical arsenal in a efficient way.
The course will consist of two parts, the statistical theory of experimental design and the methodological analysis of empirical research. The parts will be interwoven by the research problems of the participants in the project works.
In the projects the participants will tackle their research problem from these both points of view. After clarifying the question its empirical solution is outlined. Various experimental designs will be evaluated with respect to their applicability in solving the problem. The aim is to the matching of the research problem and the experimental design used.
Each day will consist of lectures and seminar-type small group work. In addition, statistical software exercises take place.
Refer to the appendices for presentation of the teachers.
Author: phd@dina.kvl.dk. Updated:
23 November 2001