🚀 go-pugleaf

I think you may pool the samples togather because you assume the contant terms
are the same for men as for women.

So the model looks like,
y=a + b*x + b'*z

where [y] =yx if from men
           yz if from women
Where [x] =x if from men
           0 if from women

Where [z] =z if from women
           0 if from men

Here is the simulation of your problem

data t1;
  do x = 1 to 30;
    y=10 + 3*x + rannor(0);
    z=0;
    output;
  end;

  do z = 1 to 40;
    y=10 +4*z + rannor(0);
    x=0;
    output;
  end;

run;

/* the data is generated by diff. model, */
/*so you may expect to reject you H0     */

proc reg data=t1;
 model y=  x z;
 xz: test x=z;

quit;
run;

Hope this helps.
In article <3649a3cc.23439818@news.ucm.es>,
  jportela@eucmax.sim.ucm.es (Javier) wrote:
>
> Hi,
>
> I want to know how to perform a test for the hypotheses:
>
> H0: The slope b in the regression curve Y=a+bX (for population A=men) is equal
> to the slope b' in the regression curve Y=a+b' X (for population B=women)
>
> versus H1: they are different
>
> That is, I have two independent samples (men and women)
> and realise two independent least squares regression models.
>
> For men, it estimates b by the OLS estimator
> For women, it estimates b' by the OLS estimator
>
> I want to test H0:b=b'
>
> Is there a simple way to do it in SAS/STAT? (I mean using REG, GLM or other
> procedures)
>
> (Suppose I have data with variables height, weight, and sex and want to
compare
> the slopes of regression lines between height and weight for men and women)
>
> Thanks in advance,
>
> Javier
>
>

-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/       Search, Read, Discuss, or Start Your Own