R help - Feb 2003 - how to do regression analysis for multiple dependent variables at once

Dear R-users

I have, say, 5 dependent variables, d1 to d5.
And I also have 2 independent variable x1, x2.
Suppose I need to do regression analyses for all
the dependent variables, using the same set 
of independent variables, x1 and x2.

How can I do the analyses without writing five lines of
lm(d1~x1*x2)
lm(d2~x1*x2)
lm(d3~x1*x2)
lm(d4~x1*x2)
lm(d5~x1*x2)

I like to write the R codes something like
l <-
c("d1","d2","d3","d4","d5")
for( i in l ) lm(i~x1*x2).

However, this does not work.

I hope you get the idea about what I want to do.
So, Could you help me.
Sincerely

---------------------------
Hiroto Miyoshi???????
h_m_ at po.harenet.ne.jp

On 02/15/03 14:08, Hiroto Miyoshi wrote:
>Dear R-users
>
>I have, say, 5 dependent variables, d1 to d5.
>And I also have 2 independent variable x1, x2.
>Suppose I need to do regression analyses for all
>the dependent variables, using the same set 
>of independent variables, x1 and x2.
>
>How can I do the analyses without writing five lines of
>lm(d1~x1*x2)
>lm(d2~x1*x2)
>lm(d3~x1*x2)
>lm(d4~x1*x2)
>lm(d5~x1*x2)
lm(cbind(d1,d2,d3,d4,d5)~x1*x2)

This is because the dependent variable in lm() can be a matrix
instead of a vector.

It says this in "In introduction to R" under "Formula for
statistical models."  But I could not find this in the help page
for lm() or formula().

-- 
Jonathan Baron, Professor of Psychology, University of Pennsylvania
R page:               http://finzi.psych.upenn.edu/