[R-sig-Geo] Why do I get different P value of LISA from R and GeoDa ?

[Roger Bivand]

On Fri, 27 Jul 2012, ria arinda wrote:

> I wonder why localmoran[,5] or P value which I get from R 
> localmoran(variableX, weight.nb, alternative='two.sided') is different 
> with GeoDa's (LISA_P). I have used alternative 'two.sided', 'greater', 
> and 'less' but the they're still different. Both LISA indices from R and 
> GeoDa are already equal. What should I do to get similar p value from R 
> localmoran and GeoDa localmoran? Thanks

If I remember correctly, GeoDa uses a permutation test (with random 
numbers), running only 99 permutations, while localmoran() in spdep uses 
the analytical expectations and variances. Repeated permutation tests 
and/or increasing the number of permutations should lead to similar 
results as the analytical formulae, possibly unless the data do not meet 
the distributional assumptions of the formulae.

Try rather to use localmoran.sad() for a Saddlepoint approximation.

Hope this clarifies,

Roger

>
>
>
> -----
> Ria Arinda
> Jakarta Institute of Statistics, Indonesia
> Tel:  +62 856 1675312
> --
> View this message in context: http://r-sig-geo.2731867.n2.nabble.com/Why-do-I-get-different-P-value-of-LISA-from-R-and-GeoDa-tp7580621.html
> Sent from the R-sig-geo mailing list archive at Nabble.com.
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>

-- 
Roger Bivand
Department of Economics, NHH Norwegian School of Economics,
Helleveien 30, N-5045 Bergen, Norway.
voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no