Colloquium Biometricum

Vol: 36 Page: 87

Authors:

Title:
The usefulness of confounded designs in experiments with

Language: Polish

Keywords:

Summary:
In the study it was attempted to estimate the usefulness of the design confounding an interaction of higher order in the experiment with edible potato. To achieve the assumed aims a field experiment was set up at the Zawady Experimental Station owned by the University of Podlasie in Siedlce in the years 1999-2001. The experiment was established in accordance with the 3³ type design, completely confounding an interaction of higher order (three factors at three levels each). The confounded interaction was A¹B²C². According to the requirements, the construction of confounded designs makes it possible to eliminate the soil variability between sub-blocks so as to obtain more accurate estimates of main effects and interactions of lower orders. It is the interactions of the third, fourth and higher orders and not main effects that are confounded with sub-blocks in this design as they are, in most cases, less important to a researcher. The more significant the soil variability between sub-blocks is, the more effective the method of complete confounding proves. Confounded designs, when set up on the fields with considerable soil variability, eliminate the effect of soil differences occurring within a block which result from an increase of the number of combinations included in the block and thus increasing its area.
The highest effectiveness of the 3³ type design in the three-year period was found in the case of the characteristics directly connected with soil variability, that is the fresh matter yield of tubers, starch yield and the dry matter yield of tubers, which are the consequence of the effectiveness expressed in terms of yield. In respect to other quality characteristics, that is the starch content, the dry matter content and the vitamin C content in potato tubers, the effectiveness of the applied design was low and showed little variation.