[R-sig-eco] negative F value in adonis()

Peter Solymos solymos at ualberta.ca
Tue Aug 10 14:52:13 CEST 2010


Adriano,
what you have reported is strange, however, (1) it is not clear what
the strata 'bloco' is and if it can cause the problem (i.e. too
restrictive permutation scheme, or the strata has something to do with
your independent variable 'trat'), and (2) you are not using Sorensen
index "sor" = 2*a/(2*a+b+c) but  "sim" = pmin(b,c)/(pmin(b,c)+a).
After the presence-absence transformation the default Bray-Curtis
setting of 'adonis' should give Sorensen also.
Cheers,
Peter

Péter Sólymos
Alberta Biodiversity Monitoring Institute
Department of Biological Sciences
CW 405, Biological Sciences Bldg
University of Alberta
Edmonton, Alberta, T6G 2E9, Canada
Phone: 780.492.8534
Fax: 780.492.7635
http://sites.google.com/site/psolymos



On Mon, Aug 9, 2010 at 4:21 PM, Adriano Melo <grumicha at yahoo.com.br> wrote:
> Dear colleagues,
> I and a graduate student (Fabiana Schneck) are using the adonis() function in vegan to test for the effect of two treatments (homogenous and heterogenous substrate) on diatom species composition. Data resulted from a experiment carried out in 11 stream reaches. Data were obtained in six visits after colonization ceased and we opted to use the 66 combination of reaches and time as block (strata in the argument list of the function). We were able to run the function using the default 'bray' method. However, we get negative F-values when using a dissimilarity matrix obtained from simpson formulae. This is our code and result:
> bio.quali<-ifelse(bio>0, 1, 0)bio.quali.simp<-betadiver(bio.quali, "sim")adonis(bio.quali.simp~trat, strata=bloco, permutations = 9999)
>
> Call:
> adonis(formula
> = bio.quali.simp ~ trat, permutations = 9999,      strata = bloco)
>
>                               Df       SumsOfSqs    MeanSqs   F.Model   R2     Pr(>F)
> trat         1.000000  -0.587663  -0.587663  -8.821903 -0.0728      1
> Residuals  130.000000  8.659838  0.066614  1.0728
>
> Total      131.000000  8.072174  1.0000
>
>
>
>
> The reasoning to use the simpson dissimilarity was to assess whether differences among treatments observed with Sorensen (Bray-Curtis for p/a data) was due to differences in species richness (the simpson index is supposed to measure turnover and be robust to differences in species richness: Lennon et al. 2001: J.An.ECol 70:966-79).
> It is hard to imagine a negative F-value in a traditional univariate analysis, but I am not sure whether this is possible in adonis().
> Adriano Sanches Melo
>
> Departamento de Ecologia - ICB
>
> Universidade Federal de GoiásBrazil
>
>
>
>        [[alternative HTML version deleted]]
>
>
> _______________________________________________
> R-sig-ecology mailing list
> R-sig-ecology at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-ecology
>
>



More information about the R-sig-ecology mailing list