Different statistical methods of data analysis were performed on the effect of different dosages of Diflubenzurom in the consumption of velvetbean caterpillar, Anticarsia gemmatalis (Hübner). The best descriptives statistics to visualize the lack of normality were skewness and kurtosis suggesting the necessity of data transformation. When data without transformation were analyzed no statistical differences were observed among consumption of larvae maintained on leaves treated with Diflubenzurom. However, analysis of transformed data showed that larval feeding was lower on leaves treated with 15 g a.i. of Diflubenzurom, differing from 7.5 g a.i. treatment. It was observed a reduction coefficient of variation from 47.8 % to 2.7 % of foliar consumption one day after treatment, reducing the experimental variability.
Insecta; Lepidoptera; skewness; kurtosis; maximum likelihood