The objective was to utilize exploratory data analysis techniques for evaluating dry matter production (DMP) obtained from 16 cuts of 92 accessions of alfalfa (Medicago sativa L.). A significant effect on DMP was observed both within and among cuts; the coefficient of variation ranged from 17.7% (cut 2) to 51.7% (cut 9). The DMP data obtained from 16 cuts of 92 accessions of alfalfa did not fit to the normal distribution. The skewness coefficients were negative for dry matter production data in cuts 1 to 4, 10 to 12, and 14 to 16, while kurtosis coefficients were negative in cuts 1, 9, 10, 12, and 16. Robust estimates of standard deviation obtained from four different methods were different from the traditional standard (S) in all cuts, showing that estimate S should be replaced by these robust estimators in statistical analysis.
dry mater production; location tests; robust measures of scale; tests for normality
