A note on the analysis of 3³ and 3⁴ designs (with three factor interactions confounded) in field experiments in agriculture

Abstract

In this paper the authors deal with two modern designs of field experiments described by Yates in Technical Communication No. 35 of the Imperial Bureau of Soil Science. The object of their note is to "explain fully the computations involved in the analysis of such designs so that they may be easily understood by the agricultural experimenter in India". Data from an experiment on sugar cane carried out at Shahjahanpur during 1938-9 are used to demonstrate the analysis of a partially confounded 3³ design in which three different methods of interculture were tested at three different levels of nitrogen manure with the number of irrigations varying between 2 and 6. To demonstrate the analysis of a confounded or partially confounded 3⁴ design data from a real experiment were not available and the results of a uniformity trial with dummy treatments superimposed had to be used for the purpose.