Application of covariance analysis to feed/ ration experimental data

Abstract

The use Analysis of Covariance (ANOCOVA) to feed/ration experimental data for birds was examined. Correlation and Regression analyses were used to adjust for the covariate – initial weight of the experimental birds. The Fisher's F statistic for the straight forward Analysis of Variance (ANOVA) showed significant differences among the rations. With the ANOVA, the calculated F statistic was 4.025, with a probability of 0.0149. On the other hand the F calculated using the ANOCOVA was 2.25 with a probability greater than 0.05. This was not significant (as f_0.05 = 2.93). This means that by using ANOVA, the calculated F will be over estimated and a researcher may arrive at a false conclusion indicating that there were treatment differences when actually there may be none.