State-of-the-art certified and supported ANOVA facilities in an R environment

msanova is an R package that provides a clear and comprehensive analysis of variance for any balanced multi-stratum design. It uses the tried and tested ANOVA algorithm from the Genstat software system that was originally written at Rothamsted, and which is the direct descendant of the original analysis-of-variance algorithms of Fisher and Yates.

A particular strength is that msanova can handle designs with several sources of random error (i.e. multi-stratum designs).

For example:

  • Split-plot designs
  • Latin squares
  • Row-and-column designs
  • Strip-plot designs
  • Split-split-plot designs
  • Nested Latin Squares
  • Latin squares with split plots

The treatments can have a factorial structure or a nested structure, or any mixture of the two. The condition of balance covers all orthogonal designs, designs with balanced confounding (e.g. balanced incomplete blocks, balanced lattices, etc.) and various partially balanced designs (e.g. square lattices).

msanova also identifies when a design that it is given is unbalanced, and allows you to generate the necessary arguments to produce an alternative REML linear mixed model analysis using the asreml-r package.

VSNi’s msanova facility can be integrated seamlessly with asreml-R, providing the most comprehensive R package available for the analysis of variance in balanced designs.

You can find out more about msanova in the Reference Manual:


Installation guides:


Email info if you would like further information or you can purchase in the webstore.