Prev: how to make Matlab quit on an error ("quit if error"?)
Next: Wind speed Map Classification
From: Rick on 6 May 2010 00:15
Hi,

I'm searching for how I can easily test if two regression coefficients are the same. I can't seem to find any data on how to do this easily and can only find information pertaining to how to do it in stats.

I.e. y = alpha1 + alpha2*x1 + alpha3*x2 + alpha4+x3 + e

I want to test if alpha2 == alpha3

Thanks
From: Peter Perkins on 6 May 2010 08:49
On 5/6/2010 12:15 AM, Rick wrote:

> I.e. y = alpha1 + alpha2*x1 + alpha3*x2 + alpha4+x3 + e
>
> I want to test if alpha2 == alpha3

The standard thing would be to fit the equivalent model

y = alpha1 + alpha2*(x1+x2) + alpha3*x2 + alpha4+x3 + e

and then test if alpha3 is significantly different from zero. You could
use a t test on the coef (e.g., REGSTATS if you have access to the
Statistics Toolbox), or F test against the simpler model

y = alpha1 + alpha2*(x1+x2) + alpha4+x3 + e

The two tests are equivalent. Hope this helps.
 | 
Pages: 1
Prev: how to make Matlab quit on an error ("quit if error"?)
Next: Wind speed Map Classification