R help - Apr 2006 - arrange data for simple regression analysis

Hello, I want to arrange data from a table to perform a simple
regression. All the examples I saw deal with paired data, e.g. 'x' and
'y' have the same dimensions (e.g. 5 values for x and 5 for y).

But I have more than one 'y' for each 'x' value, e.g. the data
file
has a x = 0, 30, 60, and 120 columns. And for each of them I have
several replicate responses (e.g. individuals), not allways the same
number. After I read the data with read.table(), ending with 4
columns, what is next? how can I regress this against c(0, 30, 60,
120)?

0       -->   n1 y values
30     -->  n2 y values
60     --> n3 y values
120  --> n4 y values

Thanks,

Tomas

Hi

not sure what do you want to do but what about

y<-colMeans(your.data)
x<-c(0,30,60,120)
fit<-lm(y~x)

Is this what you want?

Better to use advice suggested in posting guide and to show some 
reproducible example.

HTH
Petr


On 24 Apr 2006 at 5:15, Tom?s Revilla wrote:

Date sent:      	Mon, 24 Apr 2006 05:15:58 -0400
From:           	"Tom?s Revilla" <tomlists at gmail.com>
To:             	r-help at stat.math.ethz.ch
Subject:        	[R] arrange data for simple regression analysis
> Hello, I want to arrange data from a table to perform a simple
> regression. All the examples I saw deal with paired data, e.g. 'x'
and
> 'y' have the same dimensions (e.g. 5 values for x and 5 for y).
> 
> But I have more than one 'y' for each 'x' value, e.g. the
data file
> has a x = 0, 30, 60, and 120 columns. And for each of them I have
> several replicate responses (e.g. individuals), not allways the same
> number. After I read the data with read.table(), ending with 4
> columns, what is next? how can I regress this against c(0, 30, 60,
> 120)?
> 
> 0       -->   n1 y values
> 30     -->  n2 y values
> 60     --> n3 y values
> 120  --> n4 y values
> 
> Thanks,
> 
> Tomas
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
> http://www.R-project.org/posting-guide.html
Petr Pikal
petr.pikal at precheza.cz

Here are a couple of possibilities using the builtin iris data set.  Note
that although the coefficients come out the same, the degrees of
freedom, etc., would differ:
> n <- rep(1:3, 50)
> lm(Petal.Length ~ Petal.Width, iris, weight = n)
Call:
lm(formula = Petal.Length ~ Petal.Width, data = iris, weights = n)

Coefficients:
(Intercept)  Petal.Width
      1.057        2.262
> lm(Petal.Length ~ Petal.Width, iris[rep(1:nrow(iris), n),])
Call:
lm(formula = Petal.Length ~ Petal.Width, data = iris[rep(1:nrow(iris),
    n), ])

Coefficients:
(Intercept)  Petal.Width
      1.057        2.262

On 4/24/06, Tom?s Revilla <tomlists at gmail.com>
wrote:
> Hello, I want to arrange data from a table to perform a simple
> regression. All the examples I saw deal with paired data, e.g. 'x'
and
> 'y' have the same dimensions (e.g. 5 values for x and 5 for y).
>
> But I have more than one 'y' for each 'x' value, e.g. the
data file
> has a x = 0, 30, 60, and 120 columns. And for each of them I have
> several replicate responses (e.g. individuals), not allways the same
> number. After I read the data with read.table(), ending with 4
> columns, what is next? how can I regress this against c(0, 30, 60,
> 120)?
>
> 0       -->   n1 y values
> 30     -->  n2 y values
> 60     --> n3 y values
> 120  --> n4 y values
>
> Thanks,
>
> Tomas
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
>