ON ANALYSIS OF LONG TERM EXPERIMENTS WITH CHANGE IN INPUT
Loading...
Date
1996
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
AAU, Anand
Abstract
Among various types of agricultural experiments, long term experiments are continued on the same site with the same set of treatments and package of practice for many years to study the effect of treatments on soil productivity. For such experiments, statistical procedures are available for pooling the results. Due to advancement in agrotechnology, scientists are sometimes changing the package of practices (input factor) but not the basic treatments in the ongoing long term experiments. This situation affects/alters the basic concept of the long term experiment. Such an experiment is in progress on bidi tobacco crop since 1960-61 at the Bidi Tobacco Research Station, Gujarat Agricultural University, Anand, wherein variaties were changed as and when they were released by the station. The statistical method of data analysis of such experiment is not available in the literature (having direct application). Therefore, present study was undertaken with a view to (i) compare analysis of variance and rank analysis techniques for combined analysis of data, (ii) explore the feasibility of applying principal component analysis in combined analysis of data, (iii) apply regression technique in predicting long term effect of each treatment, and (iv) study suitability of transformation-sustainability index as a method of analysis of long term experiment involving changed input factor(s). Five methods viz.. Analysis of variance (ANOVA), Rank analysis, Principal component analysis. Regression analysis and transformation were employed for analysing the data. On the basis of results obtained through statistical analysis of the long term experiment on bidi tobacco, it can be generalized that the Rank analysis was the best method for analysing such complicated experiments. Other methods such as ANOVA, Regression analysis and Principal component analysis could not prove effective due to disturbance in basic assumptions like common errors, additivity of effects etc., due to varietal variation which was confounded with year effect. A new type of transformation termed as sustainable transformation is proposed for analysing the data of long term experiment with changed input factor (practices).
Description
Keywords
AGRICULTURAL STATISTICS, Analysis