Re: Multiple regression coefficients
- From: Bruce Weaver <bweaver@xxxxxxxxxxxx>
- Date: Thu, 17 Nov 2005 21:00:16 -0500
sangdonlee@xxxxxxxxx wrote:
The order of variables does MATTER if the independent variables are collinear, not independent, near singular, or not full rank. This is called multicollinearity.
Suppose I have 3 explanatory variables, A, B, and C. I perform hierarchical regression, and add one variable on each of 3 steps. There are six possible orders in which I can enter the variables:
ABC ACB BAC BCA CAB CBA
I understood the OP to be asking about the regression coefficients in the FINAL model that contains all 3 explanatory variables. In THAT model, as noted by *** Startz, the coefficients for A, B, and C do NOT depend on the order in which the variables were entered.
-- Bruce Weaver bweaver@xxxxxxxxxxxx www.angelfire.com/wv/bwhomedir .
- References:
- Multiple regression coefficients
- From: statukey@xxxxxxxxx
- Re: Multiple regression coefficients
- From: *** Startz
- Re: Multiple regression coefficients
- From: sangdonlee
- Multiple regression coefficients
- Prev by Date: Re: Multiple regression coefficients
- Next by Date: Circle stats
- Previous by thread: Re: Multiple regression coefficients
- Next by thread: time series with binary data
- Index(es):