Routine statistical analyses of field experiments by using an Excel extension

This paper wishes to bring to the attention of readers a new EXCEL® VBA macro add-in, which may be useful as a support for teaching statistics or to perform statistically sound analyses of routine agriculture experiments. The code was progressively developed over the years to support the intensive field research activity that is carried out at the author's institution and has been specifically thought to address all the peculiarities of routine field experiments, with particular reference to agronomy and plant protection. A very user-friendly interface has been developed to be used by people (mainly students and technicians) with a limited background in statistics and computer programming. Starting from experimental data, users may easily perform 23 different types of Analyses of Variance (ANOVA), including one-way balanced or unbalanced designs and two to four way balanced designs. Fully randomised, randomised block, latin square, factorial and split-plot designs may be analysed, as well as multi-location and/or multi-year experiments with one or two experimental factors, considering annual and perennial crops. A series of diagnostic tools have been implemented for one-way experiments, to verify whether basic assumptions for ANOVA are met (i.e. Tukey's test for non-additivity, Bartlett's and Levene's tests for homogeneity of variances), to seek for possible outliers and to select the most appropriate transformation of data. Several multiple comparison tests have been as well implemented, to cover all the most common types of routine analysis for field experiments. Users can also calculate correlation matrices, perform simple and multiple linear regression analyses and compare regression lines. DSAASTAT has been extensively tested and validated; it is freeware and can be easily downloaded from the author's web-site.