5

I would like to be able to take something like a two-way table of cell means with standard errors and sample sizes from a journal article and compute an ANOVA table, create interaction plots, and other things that are possible.

I can do this manually, I am just wondering if there are already some ways (e.g., a package) to do this automatically in R.

Even just some better search terms might help, because searching for ANOVA, cell means, and standard error, is obviously not very selective.

Jdub
  • 589
  • 5
  • 14
  • If you can do this manually already, you should be able to code it up yourself, at least as a last resort. Have you tried writing your own function in R? Is that off limits for some reason? Do you need resources about coding in R? – gung - Reinstate Monica May 21 '12 at 03:18
  • @gung, I will just write the code myself if I can't find anything, which I haven't so far. I just wanted to make sure I wasn't missing something buried somewhere not obvious. I was just hoping there was something that resulted in an object that could be used by anova(), etc, similar to the result of using lm() on the data. Thanks. – Jdub May 21 '12 at 13:35
  • There may well be something, w/ R there usually is--I just don't know of it. Hopefully, we'll find you something here. Alternatively, I wonder if this might do better on Stack Overflow, since maybe it's more about the use of R than about stats. – gung - Reinstate Monica May 21 '12 at 15:09
  • If you could provide some example input it might encourage more thorough answers/be more useful to future readers. – Ben Bolker Jun 01 '12 at 09:44

1 Answers1

0

Have a look at package HH. It seems to provide, what you are looking for.

Roland
  • 127,288
  • 10
  • 191
  • 288
  • Thanks. This is the kind of thing I was looking for. For example, according to the documentation the function anovaMean() produces "Oneway ANOVA table from the summary information consisting of group sample sizes, means, and standard deviations. The full dataset is not needed." – Jdub Jun 01 '12 at 13:52
  • If you don't only want one-way ANOVA, you should check out function aovSufficient. – Roland Jun 01 '12 at 13:58