MInitab - basic two-way ANOVA using a general linear model

Minitab - basic two-way analysis of variance using the General Linear Model (I use this option instead of the two-way so that if you do have a third factor or categorical variable it is simple to add it into the model). I look at main effects plots and interaction plots. I don't cover the assumptions again but of course as we violated the normality and equal variance assumptions in the one-way anova we wouldn't want to draw any strong conclusions here either.

Category:

Education

Tags:

• minitab
• two-way
• glm
• anova
• interaction
• main
• effects

[louiseisloopy]

Hi thanks so much for this post, its incredably useful. I was wondering if you could tell me if there are any scienarios when we cannot use the GLM?

Thanks

[st8tistics]

@louiseisloopy Hi, do you mean that if you know two-way anova is approriate can you always use a GLM to do it? Hmmm, I would suggest yes - there are other options for two way anova if you know something special about the data (eg it is perfectly balanced). These take mathematical advantage of it, which was handy decades ago when analysis was done by hand or by clunky computers! You can always try running it through both methods and see if it makes much of a difference. (Or ask a mathematician!)

[st8tistics]

@louiseisloopy If you mean more generally can you throw any data at a GLM - unfortunately not. It is a very useful family of models but it cannot deal with every possible dataset - you always have to check basic assumptions and if they don't hold you have to go searching for a 'fix' (like a transformation) if there is one, or perhaps use a different type of model altogether.

[DrJadeBalfour]

Thanks! You explained it very clearly, and it helped a lot.

[hiepsihon]

Many thanks indeed. This is really helpful for me.

[tr0ynt]

great accent :)

[ajwolftill]

You saved my bacon! Thank you so much :)

[kkh226]

Thank you so much!! This is exactly what I've been looking for! Great explanation! :D

[luichiification]

Quite helpfull. Thanks a lot!

[KapteinSabeltan]

Thanks aussie! from Norway