Interpreting interaction and model choice for trajectory analysis

[alej...@gmail.com]

Dear list,

I have fitted the following model:

efit <- extended.pgls(f1 = Y ~ SP * SEX * CS, data = Data, species = "SP", phy = tree)

which produced the following ANOVA table:

             Df     SS      MS      Rsq      F      Z   Pr(>F)
SP           23  67.18   2.9210  0.07577  1.7611  4.3857  0.001 ***
SEX           1  11.43  11.4303  0.01289  6.8916  5.2829  0.001 ***
CS            1  10.23  10.2329  0.01154  6.1697  4.3949  0.001 ***
SP:SEX       23  73.29   3.1864  0.08266  1.9212  4.5325  0.001 ***
SP:CS        23  92.82   4.0357  0.10469  2.4332  6.7338  0.001 ***
SEX:CS        1   3.36   3.3615  0.00379  2.0267  1.7792  0.039 *  
SP:SEX:CS    23  54.35   2.3629  0.06130  1.4246  2.6777  0.007 **
Residuals   372 617.00   1.6586  0.69590
Total       467 886.61

My interpretation of the significant SP:SEX:CS interaction is that sexual dimorphism in shape varies with size, and that this shape-size relationship differs across species.

My first question is: can I use the output of extended.pgls() directly as input for trajectory.analysis()?

If not, I understand that I should fit a full model using something like Y ~ SP * SEX * CS for the trajectory analysis. However, I’m unsure how to interpret the resulting trajectories in this context. Will they reflect shape differences between sexes, within species, after accounting for size (CS)? Is CS treated as a covariate in this scenario?

Thank you in advance!

Best regards,

Ale

[Adams, Dean [EEOB]]

Ale,

To your first question: Yes, you can use the output of E-PGLS in trajectory analysis. We did exactly this in the empirical example in the original paper (see that empirical example).

However, a three-way interaction becomes more complicated, as trajectory analysis is built for 2 way interactions only.  

Dean

-- 
Dr. Dean C. Adams
Distinguished Professor
Department of Ecology, Evolution, and Organismal Biology
Iowa State University
www.public.iastate.edu/~dcadams/
phone: 515-294-3834

---

From: geomorph-...@googlegroups.com <geomorph-...@googlegroups.com> on behalf of alej...@gmail.com <alej...@gmail.com>
Sent: Friday, April 25, 2025 4:40 PM
To: geomorph R package <geomorph-...@googlegroups.com>
Subject: [geomorph-r-package] Interpreting interaction and model choice for trajectory analysis

 

--
You received this message because you are subscribed to the Google Groups "geomorph R package" group.
To unsubscribe from this group and stop receiving emails from it, send an email to geomorph-r-pack...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/geomorph-r-package/42f57696-51da-4cd4-93eb-38d687e1fd67n%40googlegroups.com.