Abstract
TT HAS BEEN pointed out by R. A. Fisher (1) that randomized blocks I have the advantage that it is possible to isolate the appropriate components of error applicable to any specified comparison of the treatments. This is useful when there is any reason to question the use of the pooled error term. The Latin Square effects a double elimination of block differences and there appears to be no reference to separating the residual interaction of rows and columns into components which could be associated with the comparisons among the letters which designate the treatments in a Latin Square. It is possible to achieve this segregation for the Second Transformation Set of 4 x 4 Latin Squares (2). The standard square of this set is
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
