[R-sig-eco] Differences between total constrained inertia (rda) and variance explained (varpart()) in vegan

[Jari Oksanen]

Probaby the difference is the adjustment: varpart() uses adjusted R-squared, but rda() output reports unadjusted proportions. RsqureAdj() function gives both.

Cheers, jari Oksanen
________________________________________
From: R-sig-ecology <r-sig-ecology-bounces at r-project.org> on behalf of Tim Richter-Heitmann <trichter at uni-bremen.de>
Sent: 02 November 2015 15:58
To: r-sig-ecology at r-project.org
Subject: [R-sig-eco] Differences between total constrained inertia (rda) and variance explained (varpart()) in vegan

Dear list,

when i perform RDA on a species set constrained with a set of
predictors, which i have tested for their relevance by forward
selection, i get values for inertia:

Inertia Proportion Rank
Total         0.12626    1.00000
Constrained   0.05493    0.43507   11
Unconstrained 0.07133    0.56493   48

So, the amount of explained inertia is
0.05493/0.1262

or roughly 43%.

I am now grouping the same predictors into three groups (soil,
vegetation, space) and perform varpart() on them, i
get a total variance (sum of all explained variance) explained of only
30%. This has been consistent for me in a variety of datasets.
Can somebody explain to me the difference?

Thank you!


--
Tim Richter-Heitmann (M.Sc.)
PhD Candidate



International Max-Planck Research School for Marine Microbiology
University of Bremen
Microbial Ecophysiology Group (AG Friedrich)
FB02 - Biologie/Chemie
Leobener Straße (NW2 A2130)
D-28359 Bremen
Tel.: 0049(0)421 218-63062
Fax: 0049(0)421 218-63069

_______________________________________________
R-sig-ecology mailing list
R-sig-ecology at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-sig-ecology