This project has moved and is read-only. For the latest updates, please go here.


Feb 15, 2014 at 4:41 PM
I would like to make an anova on several factors
Is there a simple example available?

best regards

Feb 17, 2014 at 7:20 AM
We currently support only one-factor ANOVA. We have plans to support two-factor ANOVA ( but as you can read in those bug notes, the correct analysis for unbalanced designs is debatable. I do intend to eventually go back to this question, and I will count your request as a vote for that feature.

I'd like to ask a few follow-up questions if I may. (1) How many factors do you intend to simultaneously analyze? (2) Is your design balanced, i.e. equal numbers of entries in each variant of each factor? (3) Are you sure ANOVA is appropriate to your question, i.e. factors variants really are discrete and unordered (e.g. Continent of origin = {America, Europe, Asia, Africa} not Dose = {High, Medium, Low})? For ordered (or binary) factor variants a multi-linear model is probably more appropriate.
Feb 18, 2014 at 6:10 AM
Thank for your answer. Yes the factors are dicrete (Mean 1, 2..) method (1,2,3..) Sometimes I have 5 or 6 factors. if the factors have 2 or 3 terms, it is still possible to use a quadratic linear model, but it is more difficult when the number of terms exceeds 3.
Meta numerics is a very usefull tool... thank