[R] Calculate Mean for each Treatment/Trial Interaction in DF

[David Winsemius]

On Nov 5, 2009, at 7:53 PM, jimdare wrote:
>
> I am create a new DF that summarizes the mean angle per treatment/ 
> trial, of
> the original DF (see below).  I have had some success using:
> (tapply(df$Angle, INDEX=interaction(df$State,
> df$Trial),  FUN=mean));

That's rather difficult to accept, since State is not a column name of  
df.

> however, this gives the answer as a list, which
> means I would have to split the name to get the categories back.  Does
> anyone know a simple way to transform the Original DF into Summary DF?
>
> Thanks in advance,
>
> James
>
>
> Original DF
>
>> df
>       Angle Trial    Treatment
> 1    43.1297     1   C
> 2    62.3613     1   C
> 3    88.2767     2   C
> 4    75.2431     2   C
> 5    91.3668     3   C
> 6    61.2800     3   C
> 7    55.5575     1   U
> 8    69.4661     1   U
> 9    67.5512     2   U
> 10   95.5528     2   U
> 11   75.8689     3   U
> 12   66.9070     3   U
>
> Summary DF
>
>> sdf
>
>       Angle Trial    Treatment
> 1    52.7455   1   C
> 2    ...           2   C

> snip
>
Try:

  aggregate(df$Angle, list(df$Treatment,
   df$Trial), sum)
   Group.1 Group.2        x
1       C       1 105.4910
2       U       1 125.0236
3       C       2 163.5198
4       U       2 163.1040
5       C       3 152.6468
6       U       3 142.7759

> -- 

David Winsemius, MD
Heritage Laboratories
West Hartford, CT