I have a dataset in which I have identified a three-way interaction. Factor A has 3 levels, factor B has 4 levels, and factor C has 3 levels. I would like to produce the following: 1) A manova in which I can test the effect of factor C for each combination of factor A * factor B. I know how to do this using a series of "do" commands, but I wonder whether there is a more significant solution. For other converts from SAS, I am specifically looking for the R equivalent of the "by" command in PROC GLM. 2) Create a table of means that, for each level of factor B, shows the interaction of factors A and C. Thus, I need to manipulate summary.formula or another table-generating program to produce subtotals. If that is not possible, is there another way to avoid generating and binding multiple subsets into one table? I have searched the R-help archives and the internet, but to no avail. The closest posting that I have found is: http://tolstoy.newcastle.edu.au/R/help/05/04/3239.html. Would someone please help me? Thank you. Matthew M. Harbur University of Minnesota Southwest Research and Outreach Center 23669 130th Street Lamberton, MN 56152 (507) 752-5091 (office) (507) 752-5097 (fax) [[alternative HTML version deleted]]
I am resubmitting my question from earlier today -- this time with the proper plain text formatting. I have a dataset in which I have identified a three-way interaction. Factor A has 3 levels, factor B has 4 levels, and factor C has 3 levels.? I would like to produce the following: 1) A manova in which I can test the effect of factor C for each combination of factor A * factor B.? I know how to do this using a series of ?do? commands, but I wonder whether there is a more significant solution. For other converts from SAS, I am specifically looking for the R equivalent of the ?by? command in PROC GLM. 2) Create a table of means that, for each level of factor B, shows the interaction of factors A and C. Thus, I need to manipulate summary.formula or another table-generating program to produce subtotals. If that is not possible, is there another way to avoid generating and binding multiple subsets into one table? I have searched the R-help archives and the internet, but to no avail.? The closest posting that I have found is: http://tolstoy.newcastle.edu.au/R/help/05/04/3239.html. Would someone please help me? Thank you. Matthew M. Harbur University of Minnesota Southwest Research and Outreach Center 23669 130th Street Lamberton, MN? 56152 (507) 752-5091 (office) (507) 752-5097 (fax)