RE: st: difference in lfit slop

[Nick Cox <n.j.cox@durham.ac.uk>]

Austin's answer leaves some details. _cons is not the slope; it is the intercept. Look at p values to get significance levels. 

Nick 
n.j.cox@durham.ac.uk 


-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of Austin Nichols
Sent: 04 October 2011 15:07
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: difference in lfit slop

Lars Folkestad <lfolkestad@health.sdu.dk>:

regress metode1 age
est sto m1
regress metode2 age
est sto m2
suest m1 m2
test [m1_mean]_b[age]=[m2_mean]_b[age]

On Tue, Oct 4, 2011 at 9:30 AM, Lars Folkestad <lfolkestad@health.sdu.dk> wrote:
> dear list
> i am doing a twoway scatte:
>
> twoway (scatter methode1 age) (scatter metode2 age) (lfit methode1 age) (lfit methode2 age)
>
> this gives two different lfit slopes, one for metode1 and one for metode2 by age. i would like investigate if this slope is different. i understand that i could investigate the coef from the two liniar regressions:
>
> regress metode1 age
> regress metode2 age
>
> using the methode described by clogg et al:
>
>  z= (b1-b2)/squ(SEb1e2-SEb2e2)
>
> where b1 is the estimated slope1 (metode1) and SEb1 is the standard error for metode1.
>
> my questions:
> how do i do this in stata?
> is the slope1 (b1) the same as _cons dfrom the metode1 regression model
> in the litterature, i've seen that z-values above 2 in large samples are signifikant. my sample is petit.
> so what z value is enough to discard rhe null hypothisis= there is no difference between the groups.
>
> thank you for your answer
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/