Abstract

RAMIRO VASQUEZ, Edison  and  CABALLERO, Alberto. Inconsistency of the coefficient of variation for expressing the variability of an experiment in a model of analysis of variance. cultrop [online]. 2011, vol.32, n.3, pp. 42-45. ISSN 0258-5936.

From a variable distributed normally with mean zero and variance one, arose a succession of p treatments with constant variance equal to one and whose mean increased k units, valid to form two separate experiments; one with the t first treatments of the succession and the other with the remaining p; It was found that despite the fact that both experiments maintained equals its estimate of experimental error variance, its variation coefficients are different. As a practical illustration were analyzed three experiments through an analysis of variance, which had the same variance and different structure of half of its four treatments. The three experiments presented different coefficients of variation in spite that their estimates of variance of the experimental error were the same, which shows that this statistician does not accurately reflect the variability existing in an experiment when it is parsed through an analysis of variance. Also the coefficient of variation was associated with the level of significance or capability of the F test to detect significant differences between treatments, or the range between the values of the averages of treatments.

Keywords : coefficient of variation; variation of the experimental error; analysis of variance.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )