Premium
Efficiency of Type 2 Modified Augmented Designs in Soybean Variety Trials
Author(s) -
Lin C. S.,
Voldeng H. D.
Publication year - 1989
Publication title -
agronomy journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.752
H-Index - 131
eISSN - 1435-0645
pISSN - 0002-1962
DOI - 10.2134/agronj1989.00021962008100030024x
Subject(s) - randomized block design , mathematics , statistics , efficiency , test weight , coefficient of variation , replication (statistics) , yield (engineering) , zoology , agronomy , biology , grain yield , estimator , materials science , metallurgy
Replicated designs are not practical when large numbers of test lines with limited seed supplies must be evaluated. Therefore modified augmented design (Type 2) was used to assess yield and other agronomic variables in a large number of soybean [ Glycine max. (L). Merr.) strains in Ottawa. Thirty‐three experiments, each involving 100 to 500 test lines, were studied to determine the relative efficiency (RE) of adjusted versus unadjusted observations. There were seven groups of experiments in which the same set of test lines were used, and these were further investigated to determine RE in replicated conditions. Thus, RE was defined in two different ways; for nonreplicated (single) experiments the RE was measured by the ratio of variances between the unadjusted and the adjusted values of the subplot controls, and for the replicated experiments the RE was measured by the ratio of the mean squares of the “Replication ✕ Test line” (based on a randomized block design) between the unadjusted and the adjusted values. Adjustment was made for fertility variation by using the row and the column effects of control plots (Method 1), and the regression coefficient of the test plots on the control (Method 3). The results show that, for yield, the average RE by Method 3 was 220% for the nonreplicated and 167% for the replicated experiments, and 160 and 147% for Method 1 for the corresponding experiments. The adjustment was most effective for yield and plant height, somewhat less effective for protein and oil, and not very effective for maturity and 100‐seed weight.