Comparison of Different Methods for the Analysis of Long Term Fertilizer Experiments

Abstract

With a view to compare the various methods employed for the analysis of long-term fertilizer experiments on crop sequence, yield data were collected from a long-term fertilizer experiment conducted at Gujarat Agricultural University, Anand for seventeen years. The methods of analysis (i) groups of experiments, (ii) split plot analysis, (iii) principal component analysis and (iv) rank analysis were compared using four different criteria viz. F test, coefficient of variation, presence of year × treatment interaction and consistency in the results and agreement of the ranking of treatment responses by Duncan's New Multiple Range Test and rank correlation coefficient. On the basis of results obtained, it was inferred that principal component method proved slightly more efficient than the usual split plot analysis and the analysis of groups of experiments. However, it did not provide any information on year × treatment interaction.