The AgroR Shiny App allows the entry of data from files in .csv (Separated by comma, period or semi-colon), .xls or .xlsx extension.
The data input parameters change according to the experimental design and the arrangement of factors.
Note: The tabulation of the data must be done in columns, each one being represented by a factor or response variable according to the experimental design
The application performs fixed effects analysis of variance for all experimental designs. The analysis output is basically divided into: Anova, boxplot, assumptions, residual plot, supplementary tests (Averages or regression test) and one or more plots.
The application returns in the 'Anova' window, the analysis of variance frame.
In the case of corruption, one of the assumptions is returned a window warning of the invalidity of the model, as follows:
It is also in this window that the user provides information on the level of significance of the Anova and the complementary tests, as well as the need for data transformation, as shown in the following table:
Assumptions of analysis of variance
The assumption of normality of errors and homogeneity of variances is returned in the 'Assumptions' window. In this case, you can change the desired tests. The exception is for the error independence test, which only the Durbin-Watson test is implemented.
Standardized residuals chart
In the graphic window 'Graph of residuals', the graphic of standardized residuals is returned. In this topic, it is possible to identify the presence of outliers. In the graphic, limits -3 and 3 are used as a way of identifying these outliers.
Averages test or polynomial regression
In the graphical window 'Complementary tests', the average or regression tests are returned. In this topic, you must inform whether the factor or factors have qualitative (default) or quantitative levels. It is also in this topic that you define which average test you want to use and/or what degree of polynomial (in case of quantitative factor).
For the tests of averages, use 'tukey' (Tukey's test), 'sk' (Scott-Knott), 'duncan' (Duncan) and 'lsd' (Fischer's LSD)
In the case of more than one factor studied, such as a factorial scheme or subdivided plot, and when there is significant interaction, use comma separation to define the degree of the polynomial.
To choose the degree of the polynomial, use 0 (Not significant), 1 (linear), 2 (quadratic) and 3 (cubic).
Finally, on the 'Graph' tab, one or more graphs are displayed according to the experimental project. It returns column, box, segment or regression graphs. It is in this tab that it is possible to change various graph parameters, such as color, font size, font, dimension, axis name, etc...