Statistical package for augmented designs (pamphlet)

Abstract

It is is useful for designing agricultural experiments conducted for comparing existing practices / check varieties, called controls, with new practices / varieties / germplasm collections, called tests, where the experimental material for the tests is limited and it is not possible to replicate them in the design. The package generates a randomized layout of an augmented randomized complete block (RCB) design and augmented complete block design with equal or unequal block sizes. The optimal replication number of the control treatments in every block is obtained by maximizing the efficiency per observation for making tests vs controls comparisons. User has a flexibility to choose the replication number of the control(s) in each of the blocks. The package generates randomized layout of the design as per the procedure of Federer (1956), which is generally overlooked while conducting such experiments. The package also performs the analysis of data generated from augmented block designs (complete or incomplete). The treatment sum of squares is partitioned into different components of interest viz. (i) among test treatments, (ii) among control treatments and (iii) among test treatments and control treatments. Multiple comparison procedures for making all possible pairwise treatment comparisons can also be employed through this package. A null hypothesis on any other contrast of interest can also be tested.