COMPARISON OF TWO PROCEDURES TO CORRECT THE STANDARD
ERROR OF THE QUOTIENT OF RANDOM VARIABLES IN LER INDEX, USED IN INTERCROPPING EXPREMENTS

Euclides Braga MALHEIROS[1]

§ ABSTRACT:
The precision of the analysis of variance of the LER index (Land Equivalent
Ratio) has been studied by many researchers. The conclusion that can be drawn,
when the index is standardized by unique divisors in the same analysis, is that
the precision of the analysis of variance is satisfactory. According to Oyejola
& Mead (1989), index defined lije this do not present distributional
problems since the divisors are constants in the same analysis and the indices
are simply sums of the random variables. The problems is that the divisors,
even being constants, are random variables and their variability must be
considered. In the present work, the precision of the LER index analysis of
variance, using computer simulated data, was compared considering two
procedures to correct the standard error of the quotient of random variables in
LER index. The LER index was standardized by the yield mean of the single crop,
when a correction of the standard error was made using the maximum of the
standard error of the parts of the sum, according to Oyejola & Mead
(1989),and using the average of them. The results show large advantage,
referring to the analysis of variance precision, when the average is used.

§ KEYWORDS:
Intercropping; correction of the standard error; LER index; analysis of
variance precision.

[1] Departamento de Ciências Exatas da
Faculdade de Ciências Agrárias e Veterinárias - UNESP- 14870-000 - Jaboticabal
- SP - Brasil. Bolsista do CNPq.