Testing Hierarchical Treatment Components in Analysis of Covariance

Rayner, A. A.; Bingham, C. Raymond; Fienberg, Stephen E.

doi:10.2307/2532671

Cited by 2 publications

(1 citation statement)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most important uses of the covariance analysis are: (a) to control the error and to increase the precision, (b) to adjust treatment measures from the dependent variable to the different groups of values of corresponding independent variables, (c) to interpret the nature of the effects of the treatments, and (d) to divide a total covariance or sum of crossed products in components (Finney, 1946;Federer and Schlottfeldt, 1954;Outhwaite and Rutherford, 1955;Fairfield, 1957;Rayner et al, 1991 andBadii et al, 2008).…”

Section: Analysis Of Covariance In the Agricultural Experimentationmentioning

confidence: 99%

Covariance analysis of sugarcane variety experiments (Saccharum spp.) in contrasting environments

Rodríguez¹,

Berres

Antunez³

et al. 2015

Weber, T., McPhee, M.J. And Anderssen, R.S. (Eds) MODSIM2015, 21st International Congress on Modelling and Simulation

View full text Add to dashboard Cite

This contribution presents the results of a covariance analysis of experiments with sugarcane varieties (Saccharum spp.) in contrasting environments on a Vertisol soil type, both under normal soil conditions and in flooding conditions. The varieties of the used sugarcane were commercial varieties of the cuban sugarcane breeding program, C87-51, C1051-72, C120-78, C8612, C86-456 and C90-469. Plant cane and first ratoon cane, both at 12 months of age, were evaluated in a Randomized Complete Block Design (RCBD) with five replications. The first ratoon cane corresponds to a plant in its second season. It is formed during the first harvest of a plant, when the lower part of the plant is left uncut, such that ratoons can grow from the buds. First ratoon cane provides a faster maturing but delivers a less amount of yield. The dependent variable was the cane yield measured in tonnes (metric tons) per hectare (tcane. ha-1). As a covariate variable of the model the millable stalk number plot was chosen. The two variables were evaluated at the age of 12 months. The results demonstrate the usefulness of a covariance analysis in the presence of a quantitative variable, which affects the experimental units or plots. In particular, adjusted average values of agricultural performance and a correct ordering of the merit of these variables could be obtained. Thereby, a bias in the results is avoided, which might occur if the variety experiments are evaluated in contrasting environments. The covariance analysis minimizes the variability among experimental units by adjusting their values to a common value of the covariate. The use of the covariate reduced the mean square error in the analysis and increased the reliability of the experimental results. If the millable stalk number per plot were not measured, then the variation of the cane yield in dependence on the impact of the specific crop-plot combinations could not be determined and would be included in the error term (mean square error). The use of a covariate reduced the mean square error in the analysis and increased the significance of the experimental results in plant cane and in the first ratoon cycle. Under the effects of water excesses in the soil a marked reduction of the cane yield could be detected for all varieties. Thereby, the varieties C1051-73, C87-51 and C120-78 were affected most, and the other group integrated by C90-469, C86-12 and C86-456 in a significantly smaller proportion. This evidences that the latter group of varieties counts with a morphophysiology and biochemical mechanisms that make them adapt better to flooded soils. No marked difference between the first and second sugarcane cutting could be detected, except a decrease in the cane yield when the number of cuttings is increased. Therefore, similar results were obtained for the plant cane and in the first ratoon cycle, with only slight differences in the magnitudes.

show abstract