Negative Lambda in Spatial-error model

[ashio]

All,

I ran a spatial-error hedonic model in GeoDa and get a lambda
coefficient of -0.96. I believe that negative component implies that
we are therefore less likely to observe similar values for homes
proximate to one another, but does anyone know what the size of the
coefficient implies? What does -0.96 mean?

Thank you.

[Julia Koschinsky]

Make sure that the 0.96 value is not an artifact of your weights.
We've seen values like this e.g., with distance bands that include a
large number of points, so you're over-smoothing. If you are working
with such weights try decreasing your distance band to test how that
affects your lambda. Julia

2009/1/27 ashio <ashi...@hotmail.com>:

--
************************
Julia Koschinsky, Ph.D.
Research Director
Arizona State University
GeoDa Center for Geospatial Analysis and Computation
URL: http://geodacenter.asu.edu
Email: julia.ko...@asu.edu

[Ash Morgan]

Thanks Julia. That did drop the coefficient to -0.47. I'm still unsure how to interpret the size of the coefficient though? As it's a spatial-error model, it's a component of the error term, and so, I'm not sure what it means beyond being less likely to observe similar housing values.



> Date: Tue, 27 Jan 2009 15:50:23 -0700
> Subject: Re: [Openspace] Negative Lambda in Spatial-error model
> From: jkos...@asu.edu
> To: openspa...@googlegroups.com
> CC: ashi...@hotmail.com

[Pedro Amaral]

Dear Ash,

I am not sure about what type of modelling you are using, but one rule of thumb is that if the result makes no straight sense, it is probably wrong. Anyway, are you running an AR process or MA? In a MA process a negative lambda would me expected.

Good luck!

Pedro.

2009/1/28 Ash Morgan <ashi...@hotmail.com>

[Anura Amarasinghe]

Hi Ash

 

Is that lambda significant ? Check your moran-I scatter plots and statistics. Perhaps they may scattered in south west quadrants. Does this imply Low-high value cluster?

 

A

---

From: openspa...@googlegroups.com [mailto:openspa...@googlegroups.com] On Behalf Of Pedro Amaral
Sent: Thursday, 29 January 2009 3:18 AM
To: openspa...@googlegroups.com
Subject: [Openspace] Re: Negative Lambda in Spatial-error model